Software developers aren't losing skills because of AI, but using time saved to skill up, according to research from O’Reilly.

Analysis from the learning platform found that rather than the boom in AI coding tools having a negative effect on developer capabilities, many view the technology as a way to streamline everyday tasks to help them focus on learning.

O'Reilly said that use of generative AI content on its own learning platform was up 89% in the UK over the last year, while machine learning grew by 51% and natural language processing was up 117%.

At the same time, O'Reilly noted that more traditional programming topics saw decline over the last year, with "programming fundamentals" down 74%, Agile falling 31%, and Git down 20%.

While that could be seen as software engineers handing off such work to AI and not bothering to study key subjects, the study noted this suggests many are learning such skills on the job and using AI to support learning.

Formal training time, meanwhile, is spent learning about more advanced AI-related topics.

Alexia Pedersen, SVP International at O’Reilly, said the study shows developers are now making AI a “core part of their learning priorities”.

"The UK tech workforce isn’t turning its back on programming," Pedersen said. "It’s building on years of engineering expertise to take advantage of the opportunities AI creates. That’s exactly what we’d expect from experienced developers."

"Many already have years of programming experience, which puts them in a strong position to adopt AI tools confidently and apply them in meaningful ways," she added.

AI isn’t a replacement for core skills

That said, O'Reilly warned that AI should complement, not replace, foundational programming knowledge – especially for newer developers.

As ITPro reported last year, concerns are rising that those entering the workforce are becoming too reliant on the technology. The result is that many could miss out on valuable learning curves in the early stages of their careers.

"The challenge now is making sure the next generation develops those same foundations before relying too heavily on AI," Pedersen commented. "AI will make great developers even better, but it can’t replace the technical judgement that comes from understanding how software really works."

The company suggested that while developers are choosing to learn in different ways, it's worth noting that employers still require specific programming skills.

React, Node.js and Java skills are still highly sought after while cloud-related knowledge is prevalent in many UK job ads.

Indeed, demand for the Clean Code courses grew 19% over the last year, O'Reilly said, with C# up 17% – showing practical coding capabilities remains important.

"What this data captures is a workforce making active choices about where their time goes," Pedersen said.

"UK tech professionals aren't casually browsing AI content – they're incorporating technical learning into new ways of working, building AI agents, testing different approaches, and exploring how these tools can improve productivity and outcomes.

FOLLOW US ON SOCIAL MEDIA

AI-augmented developer roles have increased nearly six-fold over the last five years, as enterprises move from AI experimentation to implementation, according to new research from Randstad Digital.

While there's been an increase of just 28% for traditional developers, the figure for developers with AI expertise has grown by 597%, with nearly one-in-four developer roles now requiring these skillsets.

Analysis of more than 35 million job postings shows technical professionals who acquire specialized credentials are leapfrogging traditional seniority tiers, with AWS Solutions Architect (Pro) and LangGraph/RAG Architect certifications driving estimated salary increases of 54% and 31%, respectively.

"Enterprise AI is no longer a future investment; it is today's operational reality. Yet the biggest barrier to growth is not access to technology, it is access to the right people," said Michael Morris, global head of platform and talent at Randstad Digital.

"Buying AI is easy. Integrating it safely and securely across a complex enterprise is where the true challenge lies. The specialists who can integrate, govern and scale AI inside complex organizations are in critically short supply."

While foundational roles like prompt engineers are still growing at 174%, demand has rapidly escalated up the skills ladder, with AI trainers now the fastest-growing role globally, up 281%.

As ITPro reported in February, demand for AI trainers has skyrocketed over the last 18 months. Researchers said this reflects a broader market pivot toward roles that turn AI’s potential into real support for business growth.

Demand for AI solutions leads is up 226%, process automation specialists up 196%, and AI architects up 152%, Randstad noted.

Finding talent is harder than it looks

While demand for AI-related skills is there, actually filling these positions is far harder than it appears, Randstad noted. Indeed, enterprises are facing acute challenges in sourcing talent.

AI solutions leads are currently the hardest role to fill globally, for example, with time-to-fill timelines hitting 54 days in key markets and vacancy rates of nearly 27% in the US and 18% in the UK.

Despite having talent pools of roughly 100,000 professionals, machine learning engineers face vacancy rates of 8.2% in the US and 11.2% in India. Japan, meanwhile, is facing some of the sharpest shortages globally, with a 46.8% vacancy rate for AI engineers and 25% for generative AI engineers.

All this is reflected by hiring timelines, according to Randstad. While a standard IT role typically takes 38 days to fill, the recruitment window for advanced AI infrastructure roles has expanded to an average of 54 days in the UK and 53 days in the US.

This stretches to a high of 90 days for Process Automation Specialists in Italy.

The result of this is that salary offers are rising sharply, particularly in the US. Across the Atlantic, large language model (LLM) architects have a vacancy rate of 19%, commanding average salaries of $240,000.

Brazil and Argentina have rapidly emerged as a high-growth corridor for specialized AI services, now representing over 15% of global postings combined. In Europe, the UK, Poland, Spain, and Germany show steady demand, with individual national shares between 1.8% and 2.8%, while China accounts for 7.5% of the global job volume.

"AI talent concentrated in the US and India but fast-growing corridors emerging in Brazil, Argentina and beyond, cross-border hiring is becoming a core enterprise strategy," said Morris.

"Organizations that combine global talent sourcing with deliberate investment in upskilling their existing workforce are best placed to close the gap."

FOLLOW US ON SOCIAL MEDIA

Windows may be a timeless OS, and practically synonymous with modern PCs, but that doesn't mean it's well-loved by its users. In fact, the opposite might be true in some instances, according to anecdotal evidence as well as reports by publications like Windows Central.

While the development world is far from a homogeneous blob, anecdotal consensus suggests that Windows 11 is the least enjoyed mainstream platform for developers, with systems like macOS and Linux overwhelmingly preferred. But why does Windows 11 have such a bad reputation among developers, anecdotally at least?

Whether it's on internet forums, in blog posts, or in our conversations with experts, Windows 11 is often snubbed in favour of macOS or Linux, depending on the context. In fact, one of the best advantages of Windows, developers say, is the capacity to run Linux on mainstream hardware. There are, of course, some advantages to using Windows 11 for development – but what could Microsoft do to get more developers on board?

Which OS do coders and programmers really prefer?

ITPro consulted with more than a dozen programmers, engineers, and developers to understand why they use the platforms they do, and whether Microsoft has the potential to transform Windows into an attractive destination for coding.

Little research exists into which OS developers truly prefer, although Stack Overflow findings suggest that Windows is the dominant platform in which they work; nearly half (49.5%) use it over alternatives, including macOS (32.9%) and Ubuntu (27.7%), among others. But just because Windows is the "primary [OS] in which you work" — that doesn't mean it's the OS of choice, with organizations and administrators usually calling the shots more often than not.

The developers we spoke with were divided evenly between the three major systems – Windows, macOS, and Linux — with some even using all three depending on the task. However, when we asked which OS they prefer to code on, none answered Windows.

The vast majority say they preferred using macOS when coding, including Mike McQuaid, project lead of Homebrew, the open source package manager. "Apple and the wider ecosystem seem to care more about design and building high-quality applications for free or a reasonable one-off price."

What's the trouble with Windows 11?

McQuaid tells ITPro that "Microsoft’s developer experience on their platforms has never been great", with critique among those we spoke with ranging from concerns with the tools available to the notion that Microsoft is simply less easy on the eye. Where developers use Windows, it's largely because "that's the corporate choice", as Egor Danilkov, global development director at Innowise, neatly summarized it.

The OS is, however, "cumbersome", according to Nikita Kuznetsov, founder and developer of GameAP, an open-source game server management panel, and "ships with tools I don't need"; "the ones I do need have to be installed and configured manually". Some tooling is built for POSIX systems (Linux and macOS) first and is then ported to Windows, so these versions don't always behave as expected.

Many complaints center on inconsistency, with the notion that different developers on identical Windows hardware may end up with environments that behave differently. Windows is also a massive target and attracts far more malicious attention.

There's also the question of how "managed" company machines are. That's something that Jessica Kerr, engineering manager of developer relations at Honeycomb, an observability platform, can attest to.

"Companies can put more restrictions on Windows machines, and these restrictions hamper developers. As a software engineer, I install software. I run unverified, unapproved apps (for instance, when I’m writing them)," she says. "I access the file system from the command line and write scripts and run ad-hoc programs. Windows computers can be locked down, and Macs… if they can be locked down like that, I’ve never seen a company do it."

Much of the feedback around Apple's flagship OS is around the Unix-based terminal, which feels more natural to use and can be very powerful, as well as the fact that shell scripting, SSH, and command-line tooling all behave as you'd expect. Some of the positive feedback also centers around the user interface, which is universally considered intuitive, clean, and well-designed – not that Windows 11 doesn't have its fans. An additional benefit for Danilkov is the way that Apple devices work so well together.

Windows has its problems – but is the criticism unfair?

Danilkov describes some of the criticism as "outdated", suggesting Windows 11 is "leaps and bounds better" than its predecessors and that "many reactions to Windows are based on a version of the platform used years ago". This is an idea that Jitesh Keswani, managing director at e intelligence, a digital marketing agency, expands on.

"In the past, developers found it difficult to get their development environments up and running," he explains. "This typically involved a plethora of command-line tools. In addition, many of the development environments set up on a developer’s Windows desktops would not work in the same way on the Linux-based production servers that they were deployed to. This is a legacy problem that Windows still has to overcome."

Windows does have its advantages – and that primarily comes in the form of Windows Subsystem for Linux (WSL), which lets developers run full Linux environments directly on Windows 11 platforms without the need for a dual-boot setup or virtual machine (VM). Dwellwell Analytics' data science engineer, Olga Kuzmich, humorously quips: "The best feature for developers on Windows is the one that lets you run Linux."

This system repeatedly came up in our conversations. Many of the compatibility issues simply aren't relevant anymore thanks to this platform. It also means developers can run Windows applications, Linux tools, Docker containers, and more, all on the same machine. The next iteration, WSL3, is also getting a major upgrade, with new features such as the ability to take full advantage of your machine's GPU and NPU without a "performance tax".

The other advantages come in the form of Visual Studio, which Danilkov and Taimur Ijlal, information security leader at Proxy Coupons, UK, both describe as one of the strongest Integrated development environments (IDEs) available. Ijlal also describes Windows Terminal as "top shelf". The advantages of coding on Windows also extend to certain development platforms. As Jiří Fabián, co-CEO & founder of TopMonks, put it: "I would choose Windows only if I were a .NET developer working on Windows desktop applications or a game developer."

Giving Windows a new lease of life for coding and programming

With so much hesitation and opposition to using Windows as a system for coding, barring the ability to run WSL, is there anything that Microsoft can do to tempt the developers that ITPro spoke with?

There isn't a clean fix, says Omari Wright, web development manager at digital marketing agency seoplus+. "A lot of what makes macOS work well for development teams is structural, and changing it would mean Microsoft moving away from the openness that Windows users value," he says, adding that the tech giant could push harder on WSL and "make it feel less like a workaround and more like a first-class development path".

Nothing, however, would pull Kuzmich back. "Linux is just too natural now, and I think a lot of developers feel the same. That's exactly why, in my opinion, Microsoft had to ship WSL in the first place, and this is very telling."

Doubling down on WSL and improving this as a feature is where the majority of developers we spoke with centered their feedback around. Microsoft could make the remaining gaps between Windows and Linux file systems less noticeable, for example, and improve performance when working with larger repositories, containers, and some tools.

Keswani also suggests that Microsoft should push Windows as close as possible to a native Linux environment without sacrificing the user-friendly elements. This focus on user-friendliness is great "for end-users who don't need to know much about what's going on under the hood," Fabián adds, but to win more developers over, "Microsoft would need to base the core of Windows on Unix, which will never happen".

He continues: "With the rise of AI and heavier terminal usage, I need all core Unix utilities bundled with the system, without having to install and maintain them myself — tools for file manipulation, SSH, search, and so on."

McQuaid, meanwhile, centers his criticism on the direction that Microsoft has been taking Windows in the last few years, lamenting: "Stop trying to overrule my preferences on how to use my computer". He tells ITPro: "I use Edge and Bing there now just because they keep overruling or pressing me to change my defaults.

"There are ads (or what feels like ads) and upselling all over the OS. It’s gross and feels bargain-basement compared to Apple. I'd love to see Microsoft make Windows feel like “spammy” and more of a cohesive, overall pleasant developer experience."

There are various reasons why developers prefer other systems over Windows 11, macOS especially, with the main benefit of Microsoft's flagship OS, according to the programmers that we interviewed, being the fact that it lets you run Linux and bypass Windows entirely.

Nobody, Microsoft included, would describe Windows 11 as perfect, hence why there's such an active Insider program with updates shipped frequently and continuously. There are some unrealistic demands on developers' wish list, like making Windows Unix-based, but more realistic seems to be giving WSL more love – making it streamlined, more functional, and more powerful so that developers feel more supported by their devices at work.

Software developers are reporting marked productivity gains with AI, yet new research from Harness shows the use of the technology is creating greater downstream challenges.

Findings from the company’s 2026 State of Engineering Excellence report suggest areas such as code quality and validation are being overlooked since the mass-integration of AI, and it’s creating bigger workloads and leading to higher levels of burnout.

“AI adoption is now the default in engineering organizations, and self-reported impact is overwhelmingly positive — but the cost is accumulating in places organizations aren't watching,” the company said in a blog post.

Indeed, Harness found that many developers are now spending more of their day on manual remediation tasks. Around 81% said they spend more time in code reviews since before the adoption of AI tools, for example.

More than one-quarter (28%) reported spending 30% longer on these tasks on average, according to Harness.

Worse still, nearly one-third of that activity isn’t tracked by teams, creating a trend of “invisible work”.

Organizations estimate that roughly 31% of developer time is now consumed by invisible work, which typically involves reviewing AI-generated code, fixing bugs, and switching between disparate tools.

Trevor Stuart, SVP and General Manager at Harness, said the findings show that while AI is not only changing how developers build, but how they spend their working day.

This, Stuart noted, means organizations need to overhaul how they measure developer productivity to compensate for changing workflows.

"Cloud and the internet were infrastructure revolutions layered underneath the developer,” he said.

“AI is reshaping the developer's job entirely, and the measurement frameworks that the industry has relied on for the past decade weren't built for this new unit of work."

Changing developer metrics

Harness noted that previous frameworks for measuring productivity and efficiency simply aren’t able to keep up with current AI-fueled working patterns.

Around 89% of tech leaders still trust metrics that don’t accurately reflect AI’s impact on individual developers or teams at large.

Moreover, 94% said key considerations such as tech debt and developer burnout rates are missing from metrics, painting a convoluted picture of overall performance which fails to scratch the surface.

“The biggest AI challenge is measurement itself,” the company noted in a blog post detailing the findings.

“When asked to name the single biggest challenge, the top answers are all visibility problems: measuring true productivity impact (26%), maintaining code quality with AI (24%), and proving ROI to leadership (18%).”

Measuring developer performance

According to Harness, changes to how performance is tracked needs to take AI into account.

The company recommends that enterprises approach this by considering the impact of tasks such as code validation, as well as the overall quality of code produced by AI tools.

This, the company noted, will paint a clearer picture on how workloads are changing due to new follow-up tasks created by AI.

Organizations should also “treat AI performance as its own discipline”, the study noted. This includes tracking AI agent accuracy, acceptance rates of AI outputs, and costs associated with the tools separately from human developer output.

Developers should also be given a say in how performance is measured due to the influx of AI tools, according to Harness. Nearly half (49%) of developers said they want to be involved in defining metrics themselves.

This is crucial, the report noted, largely because there’s a growing perception gap on how AI is actually impacting teams.

Managers, for example, are nearly four-time more likely to report no concerns on how productivity metrics are measured compared to frontline practitioners.

FOLLOW US ON SOCIAL MEDIA

Anthropic believes Claude Code can play a key role in modernizing legacy code like COBOL, but IBM appears to disagree.

In a recent blog post, Anthropic said its flagship coding tool can automate much of the legwork required in COBOL modernization, which is a notoriously difficult process for enterprises.

"With AI, teams can modernize their COBOL codebase in quarters instead of years," the company said.

If as useful as promised, Claude could be the tool that finally breaks the ongoing dependence on COBOL, or Common Business-Oriented Language, a programming language first developed back in 1959.

While it's still widely used, that's largely for legacy operations, and efforts are being made to shift at least some of those workloads to more modern platforms. Anthropic noted COBOL runs 95% of ATM transactions in the US and is used to power critical systems including finance, airlines, and government.

"Despite that, the number of people who understand it shrinks every year," Anthropic noted in its blog post. "The developers who built these systems retired years ago, and the institutional knowledge they carried left with them."

Waning expertise in COBOL is exacerbated by the fact universities no longer teach the language. Indeed, it’s becoming increasingly difficult to find qualified engineers in this domain.

That could have serious implications for major industry players such as IBM, largely because COBOL is a major selling point for IBM systems as the company continues to sell mainframes to customers reliant on the language.

After Anthropic's announcement, shares in IBM fell 13.2% – the biggest drop seen by the company in more than 25 years, according to Reuters.

Claude Code vs COBOL

Anthropic said COBOL modernization is different from standard code, adding "you're reverse engineering business logic from systems built when Nixon was president."

With that in mind, it’s not just about coding, but understanding long lost institutional knowledge that's only represented in the code itself.

That's one reason why "armies of consultants" are now required to update these systems – at high cost and with lengthy timelines.

"Tools like Claude Code can automate the exploration and analysis phases that consume most of the effort in COBOL modernization," the post noted.

Using Claude Code, Anthropic said researchers were able to map dependencies across thousands of lines of code, document long-lost workflows, spot risks more quickly than human analysts, and offer "deep insights" to make key decisions.

"AI excels at streamlining the tasks that once made COBOL modernization cost-prohibitive," the company said. "With it, your team can focus on strategy, risk assessment, and business logic while AI automates the code analysis and implementation."

By going through a step-by-step process that includes validation, the AI can translate COBOL logic into modern languages, creating API wrappers around legacy components and building a system to run old and new code together.

IBM isn't convinced

In an apparent response, IBM published its own blog post about AI coding tools and COBOL – without specifically mentioning Anthropic – that argues there's a clear difference between translating code and modernizing a platform.

"Translation captures almost none of the actual complexity," said Rob Thomas, Senior Vice President, IBM Software and Chief Commercial Officer, later adding: "Decades of hardware-software integration cannot be replicated by moving code."

IBM's mainframe business isn't just about COBOL, Thomas said. Indeed, nearly half (40%) of COBOL doesn't even run on mainframes.

"COBOL on IBM Z is code optimized over decades of tight coupling between software and hardware," Thomas added.

"An analogy is the iOS and iPhone: someone could build an alternative, but it is unlikely to displace a billion iPhones. The performance derives from tight coupling of software and hardware, processor-level acceleration, I/O subsystem optimization, and decades of performance tuning."

Whether that convinces the markets to stop battering IBM's shares remains to be seen, but IBM isn't the first company to see its price fall over an Anthropic post or product launch in recent weeks.

Software and services shares dipped over the launch of Claude Cowork, while another reveal about security tools tanked share in a host of major cybersecurity companies.

FOLLOW US ON SOCIAL MEDIA

Microsoft plans to replace C and C++ with Rust within the next five years as part of a major engineering modernization project – but a senior company figure insists AI won’t be doing the heavy lifting after a social media backlash.

In a post on LinkedIn advertising a job opening last week, Galen Hunt, Distinguished Engineer at Microsoft, said the tech giant plans to “eliminate every line of C and C++ from Microsoft by 2030”.

The scale of the project here is significant, with Hunt suggesting that developers will be working toward refactoring upwards of one million lines of code on a monthly basis.

“Our North Star is ‘1 engineer, 1 month, 1 million lines of code’,” Hunt wrote. “To accomplish this previously unimaginable task, we’ve built a powerful code processing infrastructure. Our algorithmic infrastructure creates a scalable graph over source code at scale.”

“Our AI processing infrastructure then enables us to apply AI agents, guided by algorithms, to make code modifications at scale,” he added. “The core of this infrastructure is already operating at scale on problems such as code understanding.”

The move from Microsoft marks the latest in its shift to the Rust programming language in recent years. In 2022, David Weston, corporate VP for OS security and enterprise at Microsoft, revealed the company was investing around $10 million in Rust to make it the company's "1st class language" for engineering systems.

As ITPro reported back in 2023, the company made a significant leap on this front with the introduction of Rust-based kernel features on Windows as part of the Windows 11 Insider Preview Build 25905.

Progress has continued at pace in the intervening years. During a keynote presentation at RustCon 2025 in September, Azure CTO Mark Russinovich detailed the company’s efforts to embrace Rust, per reports from The New Stack.

Microsoft's Rust project courts controversy

Initially, Hunt hinted at the use of AI in helping Microsoft deliver on the goal of rewriting some of the company’s largest codebases.

The suggestion tracks with how the tech giant is using the technology in software development over the last two years. In May this year, CEO Satya Nadella revealed up to 30% of the company’s code is now written by AI.

However, Hunt’s comments appear to have courted controversy, prompting an edit on the original post to clarify that “Windows is *NOT* being rewritten in Rust with AI”.

Hunt insisted this is a research project aimed at “building tech to make migration from language to language possible”.

“The intent of my post was to find like-minded engineers to join us on the next stage of this multi-year endeavor—not to set a new strategy for Windows 11+ or to imply that Rust is an endpoint,” he wrote.

Rust is in vogue

Microsoft isn’t the only company making the shift to Rust. Last year, Lars Bergstrom, director of engineering at Google’s Android team, said developer teams embracing the programming language have unlocked marked productivity and efficiency benefits.

A 2024 project from DARPA also saw the US government agency announce plans on this front, this time actually using AI tools to translate C code to Rust.

A key factor behind the shift to Rust lies in security. Popular languages such as C and C++ have suffered from recurring security-related issues, making them vulnerable to bugs due to how they use memory.

These ‘memory-safety vulnerabilities’ have become a frequent pain point for enterprise security teams and developers alike, and have caught the attention of security agencies on both sides of the Atlantic.

In 2022, the US National Security Agency (NSA) urged organisations to shift to memory-safe programming languages, such as Rust, Python, or Java. More recent guidance from the NSA and CISA, meanwhile, once again urged devs to make the shift.

FOLLOW US ON SOCIAL MEDIA

AWS has unveiled a set of powerful new “frontier agents” , which the hyperscaler said marks a step change in the potential of agentic AI technology and the automation of far more complex tasks.

Frontier agents aren’t your typical AI agent working away in the background, automating tasks and dealing with customer service queries. As journalists were told at a pre-briefing ahead of the annual re:Invent conference, they’re a “new class” of agent capable of working fully autonomously for hours or days without intervention.

These agents are highly scalable, self-learning, and represent the next iteration of the technology, according to the hyperscaler.

To mark its first foray on this front, AWS has launched three new frontier agents, which it said are primed to fundamentally transform software development. The company’s Kiro coding service, for example, is getting a big update with an agentic twist, dubbed Kiro Autonomous Agent.

Speaking to ITPro ahead of the event, David Yanacek, senior principal engineer for agentic AI at AWS, said these new capabilities will prove vital in helping developer teams speed up production in a safe and secure manner.

“These act as an extension to your team, more so than a tool for developers,” he said. “So we've started this category of what we call frontier agents. We've started by releasing three software development lifecycle related agents.”

According to AWS, the Kiro Autonomous Agent “maintains persistent context across a range of sessions,” meaning it’s lingering in the background learning from human processes.

While this appears similar to what one would consider a ‘traditional’ AI agent, this new class tracks human activity to improve continually, learning pull requests and providing users with feedback.

The agent comes with a range of capabilities, such as triaging bugs and improving code coverage, the company told assembled media.

“You can ask it questions, describe a task, and assign tasks in your backlog directly from GitHub,” AWS explained in a statement. Thereafter, the agent will “independently figure out” how to complete the task and share recommendations for changes, potential edits or pull requests.“

Security agents target safer software

Elsewhere, Yanacek told ITPro the new AWS Security Agent aims to drive the development of safer, more secure software.

According to AWS, the agent “embeds deep security expertise throughout the development lifecycle” by proactively reviewing design documents and scanning pull requests, comparing them to common vulnerabilities and in-house security requirements.

The launch of this frontier agent represents a step change in software development, according to Yanecek, addressing long-standing security-related issues and alignment between developer and security teams. The agent can perform penetration testing to probe for security vulnerabilities, for example, turning what was previously a laborious manual process into an “on-demand capability”.

“If you're a security team, you want to make sure that all of the teams in your organization are doing things in a secure way, ultimately, and so you want them to be penetration testing code, the changes that they make, the systems that they're building, that's an expensive thing to do,” he explained.

“It takes a lot of time and a lot of setup and execution,” Yanacek added. “Sometimes teams only get to practice only their most critical systems, because it just takes so much time.”

One AWS customer that’s already used the agent, photo-sharing firm SmugMug, was able to conduct penetration testing on multiple applications within the space of hours, rather than days, Yanecek told ITPro.

In simplifying and streamlining this aspect of the development lifecycle, agents are reducing manual toil and freeing up developers to focus on more critical tasks, he added.

“Now we've taken pen test time from weeks to hours, we can just cover more ground, so we can test more changes as they're going through than you could in practice before.”

DevOps gets the agentic treatment

The third and final frontier agent unveiled by AWS today, the AWS DevOps Agent, completes the agentic treatment for the SDLC. This tool aims to streamline DevOps practices, again with a focus on reducing toil for developers.

Similar to the security agent, this has a deep knowledge of user IT estates and applications, identifying relationships between areas such as observability tools, code repositories, and CI/CD pipelines.

It achieves this through mapping application resources, according to AWS, which enables the agent to rapidly respond to failures or bugs that often cause developers headaches.

“The DevOps agent works autonomously when there are alerts or things that will engage a human either in real time or later,” he explained. “It will do root cause incident management, come up with a safe mitigation plan, and shows all of its work along the way.”

Yanacek added that the agent will be proactive and continuously improve, studying previous incidents and identifying potential future issues.

“So it will find things around, maybe cost optimization of infrastructure, saying ‘hey, these are running underutilized’, or, ‘hey, your time to detect has been poor over time’,” he explained.

AWS has tracked an estimated root cause identification rate of around 86% using the agent to date, Yanacek told ITPro. Customers such as the Commonwealth Bank of Australia found that the agent was able to identify the root cause of software failure in just 15 minutes.

“It takes hours for this type of incident [to be identified] that they threw at it to try it out,” he noted. “So we’re seeing success with that.”

MORE FROM ITPRO

• AWS re:Invent 2025: All the news, updates, and announcements live from Las Vegas
• AWS targets IT modernization gains with new agentic AI features in Transform
• Microsoft's new Agent 365 platform is a one-stop shop for deploying, securing, and keeping tabs on AI agents

Google CEO Sundar Pichai claims software development has become “exciting again” since the rise of vibe coding, but some developers might disagree while bogged down in flawed code.

Speaking during a recent podcast interview, Pichai suggested the trend that’s taken the industry by storm will have positive long-term implications for both devs and non-technical individuals alike.

“It’s making coding so much more enjoyable,” he told host Logan Kilpatrick. “I feel things are getting more approachable, it’s getting exciting again, and the amazing thing is it’s only going to get better.”

Pichai noted that non-technical staff internally at the company are using AI to support workflows and finding them to be a valuable addition to the toolkit.

“There’s a sharp increase in the set of people who have submitted their first CL’s (changelists) and it’s because these tools are making it more accessible,” he said.

“Maybe you’re a product marketing person, you have an idea… in the past you would have described it. Now maybe you’re kind of vibe coding it a little bit and showing it to people, so you can tangible see that come to work.”

Pichai is the latest in a string of big tech execs to hail the potential of vibe coding over the last year. Indeed, Google Brain founder Andrew Ng specifically highlighted the democratizing effect of AI in coding earlier this month, with workers outside the software profession flocking to the tools.

“The bar to coding is now lower than it ever has been,” Ng said. “People that code, be it CEOs and marketers, recruiters, not just software engineers, will really get more done than ones that don’t.”

Developers aren’t quite sold on vibe coding

Software developers don’t appear fully sold on the use of AI for coding, however, with research showing that although there’s a growing appetite for these tools, they often create more problems than they solve.

Findings from Google’s own State of DevOps report last year, for example, revealed more than one-third (39%) of developers had little trust in AI-generated code even though they use them on a daily basis.

Meanwhile, Stack Overflow’s 2025 Developer Survey found nearly half (46%) of developers “don’t trust the accuracy” of AI outputs when used for coding tasks. This marked an increase compared to 31% of devs who aired concerns on this front in 2024.

A key factor behind this skepticism is that the “productivity gains” delivered by the technology are negated by developers having to debug faulty code.

More than two-thirds (67%) of respondents to a survey from Harness said they now spend more time debugging flawed AI-generated code compared to when they were operating in a manual capacity.

Keep security teams in the loop

Pichai did acknowledge that there are risks associated with vibe coding, particularly in terms of security – and he’s not wrong.

A study from Veracode in July this year found nearly half of code generated with AI is fully secure, while a separate survey from Cloudsmith pointed to a growing trend of complacency among developers using these tools.

The survey found only around one-third of devs review AI-generated code before deployment, leaving enterprises open to an array of potential vulnerabilities.

In the meantime, Pichai said vibe coding should typically be used in an experimental capacity, and security teams should always be involved in the process.

“I’m not working on large codebases where you need to really get it right,” he said. “The security has to be there, so those people should weigh in.”

AI tools are the worst they'll ever be

Notably, the Google chief believes this is the worst vibe coding and AI tools will ever be, pointing to examples like autonomous taxi company Waymo. Initial skepticism and concerns about safety have given way to acceptance among commuters.

With AI coding tools, the same rules apply. Continuous improvements mean the technology will become more reliable moving forward, helping cement its place in the mainstream not only for non-technical individuals but for devs working on the frontline.

“There's something I always used to tell about Waymo, and whenever people talked to me about it is, ‘remember, this is the worst Waymo will ever drive’,” he said.

“This version of all these tools…vibe coding with Gemini 3 in AI Studio, it’s both amazing to see and it’s the worst it’ll ever be. Both are simultaneously true. You’re going to see a lot of progress ahead.”

MORE FROM ITPRO

• Want to supercharge your vibe coding skills? Here are the best AI models developers can use to generate secure code
• Vibe coding security risks and how to mitigate them
• ‘Slopsquatting’ is a new risk for vibe coding developers – but it can be solved by focusing on the fundamentals

Google Brain founder Andrew Ng believes everyone should dab their hand at “vibe coding” as a way of beefing up their programming skills, but some industry experts aren’t quite so keen on the idea.

Speaking at Snowflake’s annual ‘Build’ conference, Ng suggested those looking to build applications shouldn’t do it manually and instead turn to AI tools to speed up the process.

"Don't code by hand. Don't do the old way," he told attendees, per reports from Business Insider. "Get AI to help you to code.”

“That will make people in all job functions much more productive and have more fun."

Vibe coding is the word of the year, literally. In October, Collins Dictionary granted it the annual award following its explosive rise to prominence over the last nine months.

The practice, which involves using natural language prompts to generate code and speed up development processes, has been hailed as a great breakthrough in software development and has gained significant traction across the tech industry.

Indeed, it’s become a burgeoning industry. Some of the most-watched startups globally – such as Cursor, Replit, and Lovable – specialize in ‘vibe coding’ while big AI providers like Anthropic and OpenAI have also jumped on the bandwagon.

Proponents of the practice, like Ng, argue vibe coding helps take the drudge work out of software development and lowers the barrier to entry for non-technical individuals.

“The bar to coding is now lower than it ever has been,” Ng told attendees. “People that code, be it CEOs and marketers, recruiters, not just software engineers, will really get more done than ones that don’t.”

Nigel Douglas, head of developer relations at Cloudsmith, agreed that this aspect of vibe coding has marked benefits. It's democratizing what was once an exclusive domain of expertise.

"AI-assisted vibe coding can be incredibly helpful for beginners as it takes away a lot of the initial intimidation,” he told ITPro. “

"Users can experiment and ask questions in natural language and immediately see how the code behaves,” Douglas added. “This kind of low-friction environment often helps people grasp core concepts faster because they can learn through doing rather than memorizing syntax or other programming languages.”

Looking at it from this perspective, proponents of the trend aren’t wrong. Since OpenAI co-founder Andrej Karpathy coined the term in February this year, countless applications worldwide have been built using AI tools, often by those with no prior experience in development.

Is vibe coding risky?

But is the trend all it’s cracked up to be? Some critics suggest that relying on AI tools to build software comes with significant risks, both in terms of security and the long-term impact on developers themselves.

Others, meanwhile, say using the technology in software development often isn't worth the hassle.

Analysis from Fastly earlier this year, for example, found that while developers are using AI tools to speed up production, the gains delivered by the technology are offset by the fact they spend a big chunk of their time fixing flawed or faulty code.

More than two-thirds (67%) of developers reported spending more time debugging AI-generated code compared to before they started using these tools.

Moreover, relying too heavily on AI tools and essentially giving the technology the keys to the kingdom can have dire consequences. As ITPro reported in August, an incident involving one of the aforementioned vibe coding darlings, Replit, proved disastrous for one startup founder.

While disaster was ultimately averted on this occasion, experts told ITPro at the time the incident highlighted the risks associated with AI coding tools.

Don’t overlook the fundamentals

Security-related issues surrounding vibe coding mean this isn’t a recommended practice in the enterprise, industry experts told ITPro. Tal Zarfati, lead architect at JFrog, agreed with Ng and Douglas on the accessibility front.

However, he warned it “introduces new layers of risk if not approached carefully”.

“While enhancing developer productivity, [vibe coding] significantly amplifies software supply chain risks, especially when third-party models are used without scrutiny,” he said.

Zarfati added that developers relying too heavily on AI could leave their organization exposed to an array of AI-related threats or vulnerabilities such as slopsquatting, where attackers “exploit naming gaps in AI-generated code suggestions”.

“The risks go deeper as model-based training attacks, prompt injections, and guardrail bypasses in generative AI tools introduce vulnerabilities that traditional security practices often overlook,” he continued. “These issues are amplified when vibe-coded apps are embedded into a secure software supply chain.”

Chrissa Constantine, a senior cybersecurity solutions architect at Black Duck, told ITPro that users diving headlong into vibe coding also risk skipping vital learning curves.

“Accessibility doesn’t erase fundamentals,” Constantine said. “Overreliance on AI can lead to shallow learning, fragile codebases, and security risks.”

“Beginners who skip core development and engineering concepts, such as data structures, debugging, and architecture, may struggle when AI-generated code breaks or scales poorly.”

Ultimately, vibe coding should be viewed as a way to “complement, not replace” foundational skills, Constantine added.

MORE FROM ITPRO

• Hackers are using AI to dissect threat intelligence reports and ‘vibe code’ malware
• Vibe coding security risks and how to mitigate them
• Want to supercharge your vibe coding skills? Here are the best AI models developers can use

Anthropic has released a new browser-integrated interface for its AI coding tool Claude Code, as part of a closed beta aimed at power users.

Claude Code for web runs entirely in a user’s browser of choice rather than in a command-line interface (CLI) and can be connected directly to chosen GitHub repositories.

Anthropic is promoting the tool as a potentially quicker and easier way for developers to issue routine bug fixes, catch up on quality assurance lists, and repetitive development work such as testing.

In a demo reel, Anthropic showed the web-based tool being used to action a pull request within an enterprise GitHub repository, based on automatic, test-driven development.

Users can use the interface to run multiple tests or code generation tasks, working in parallel across a single tab.

Initially, Claude Code on the web will be accessible for Pro and Max users in research preview, direct in browser or via a dedicated iOS app. The company has not set out a timeline for wider release.

Claude Code eyes stronger security controls

Anthropic stressed that Claude Code runs in isolated sandbox environments in the cloud, with a secure proxy service used to connect the service to enterprise Git repositories.

This is part of a new sandbox runtime, released alongside Claude Code on the web, which allows admins to set out precise directories and network hosts that are accessible to the tool.

Claude Code will only run code within these limits and automatically applies them to any programs, scripts, or subprocesses it produces.

Anthropic added that its new sandboxing controls are intended to prevent risks such as malicious system file modification enabled via prompt injection attacks, as well as abuse of the model allowing hackers to download malware into sensitive environments.

The firm said that Claude Code running on the web uses “Anthropic-managed cloud infrastructure” to run, a reference to AWS which serves as its main cloud provider.

MORE FROM ITPRO

• ]Anthropic’s new AI model could be a game changer for devs
• There's no single route to AI adoption, and Anthropic is charting all available paths
• Anthropic admits hackers have 'weaponized' its tools

The UK’s tech sector could face a perilous software developer talent shortage in the coming years as the industry faces an aging workforce.

A recent study from Stack Overflow found the average age of UK developers stands at 39 years old, with 74% of professionals boasting more than a decade of coding experience.

This, the study noted, highlights the strong technical capabilities of the developer workforce across the country, outpacing peers in the US and across EMEA.

However, while it shows the workforce is “technically mature”, researchers noted this does raise concerns about future talent and long-term international competitiveness.

Combined this with rising concerns about the impact of AI on entry-level roles, and the talent-related challenges faced by the UK tech industry could be exacerbated looking ahead.

“This depth of expertise is matched by a willingness to explore new technologies, especially AI,” the report noted. “But this Millennial workforce needs new, entry-level talent to establish a steady pipeline of future developers – a challenge at a time when entry-level roles are under threat from AI.”

So how can the sector tackle this simmering problem? Speaking to ITPro, Jody Bailey, Chief Product & Technology Officer at Stack Overflow, said addressing current skills gaps will require stronger cross-sector collaboration spanning industry, government, and academia.

Moreover, the industry must adopt an approach that goes beyond requiring the fundamental technical expertise previously expected of a developer.

“In addition to coding skills, developers need exposure to secure development practices, AI literacy, and real-world problem solving,” Bailey said.

“Employers can offer apprenticeships and entry-level pathways, while universities bring emerging tech into their courses. Community-led platforms, coding challenges, and open source projects help learners get hands-on experience and build confidence.”

Notably, Bailey said cultivating an “ecosystem of continuous learning” will be crucial to attracting younger talent in software development.

Continuous learning was an area specifically highlighted in research from Gartner last year, which noted developers and engineers should upskill and build new skills in anticipation of an influx of AI in the industry.

The consultancy projected that around 80% of those in the software profession will be required to upskill in the coming years, and those who fail to keep up could fall foul of a changing industry landscape and skills requirements.

How can young software developers thrive?

Industry-led efforts to encourage young people into the profession are only one part of the equation, however. The onus still often remains on the individual themselves.

On this front, there are a number of ways young software developers can differentiate themselves - especially in the age of generative AI. Central to this is a concerted focus on “pairing technical fluency with critical thinking and adaptability,” according to Bailey.

“While AI tools can accelerate routine coding tasks, they often require careful oversight,” he said. “Skills like code review, testing, and secure design are increasingly important, along with the ability to explain technical choices clearly.”

“Developers must commit to understanding code generated by AI, not solely learning how to prompt AI tools to write code,” Bailey added.

These skills will be crucial for developers of all capabilities moving forward, especially given that research from Stack Overflow shows 42% of developers spend a significant amount of time debugging AI-generated code.

This has become a recurring talking point amidst the explosion of ‘vibe coding’ in recent months, therefore it’s crucial that developers entering the workforce are both confident using these tools and have the skills required to identify and remediate bad code.

Simply put, sticking to the basics will still be vital despite bold claims from industry providers about AI tools “reducing manual toil” or “streamlining efficiency”.

“Our survey demonstrated 69% of developers using AI agents reported productivity gains, though they emphasized the need for human judgement,” Bailey told ITPro.

“AI can handle boilerplate coding, documentation, and troubleshooting, while human expertise ensures accuracy, ethical use, and trustworthiness,” he added.

“The biggest benefits come from combining AI capabilities with human creativity and peer-reviewed knowledge, letting teams experiment and iterate faster without compromising quality or control.”

There’s still a place for young developers

Bailey said building these foundational skills will be crucial for young developers finding a place in the workforce. But will there be enough room for them? Some industry stakeholders aren’t convinced.

In recent months, repeated concerns have been raised about the impact of AI on entry-level workers across a range of sectors, and developers have been firmly in the crosshairs.

Bailey appears to have an optimistic view on this topic, however. With so many young developers having ‘come up’ with AI in the fold, this could work in their favor.

“Younger developers have a particular advantage as seasoned developers are more skeptical of AI and are slower to adopt,” he said. “Developers who are curious, collaborative, and committed to learning these new tools will have a head start as the types of junior roles evolve.”

Bailey’s comments echo those from several industry leaders on the impact of AI on the workforce. Earlier this year, outgoing GitHub CEO Thomas Dohmke noted that many younger developers are now essentially “AI native”.

This means they have an acute and in-depth understanding of the technology on account of having used it in school and higher education settings.

In August, JetBrains CEO Kirill Skrygan told ITPro he also believes there will still be a place for junior developers despite the scaremongering. Crucially, however, this will hinge on their ability to adapt to changing skills expectations, which are moving toward AI-related proficiency.

MORE FROM ITPRO

• The best free coding courses in the UK
• How to become a software developer: A beginner’s guide
• How to choose a laptop for programming

The benefits of AI coding have been “unremarkable” so far, according to new research, with developers struggling to unlock clear-cut gains from the technology.

That's the assessment from Bain & Company after analyzing how generative AI has impacted software development – and the lack of big payoffs from implementing the technology to date.

The report found that AI coding assistants may be able to take on 40% of a developer's work, including bug fixing, maintenance, code development, review and validation.

However, that hasn't led to a serious transformation in terms of broader operational efficiency.

"Software coding was one of the first areas to deploy generative AI, but the savings have been unremarkable," the consultancy said.

One reason for that is that coding itself makes up less than 40% of a software engineer's work day, limiting the impact of generative AI if they're unable to speed up in other aspects of their role.

Bain added that two-in-three software firms have rolled out generative AI tools, but developer adoption remains low. Those that are using AI assistants are reporting up to 15% increases in productivity, but that time isn't being used in higher-value work.

"So even those modest gains don’t translate into positive returns," the consultancy added.

The Bain report fits with previous research that showed coding tools may actually slow down software engineers, in part because they had to clean up much of the code generated by AI.

Similar research from Atlassian also pointed to an "unexpected paradox" in AI software development. Developers were found to save around 10 hours a week using these tools, but "organizational inefficiencies" were hampering broader productivity gains.

How to win with AI coding

Bain argues that more AI is the answer. The study noted that companies seeing the most gains are applying AI across the software lifecycle and beyond simple coding use-cases.

"If AI speeds up coding, then code review, integration, and release must speed up as well to avoid bottlenecks," the company says, noting that Netflix has successfully "shifted left" to ensure AI code isn't held back by other parts of the process.

The best way to boost returns on AI investment was to redesign processes to use the technology more efficiently, Bain insisted, as well as taking advantage of time saved with AI coding to higher-value work.

"Leading adopters treat generative AI as a fundamental transformation of their software development life cycle rather than a one-off project," the report noted.

"They take a future-back approach to rearchitect their end-to-end software development life cycle around generative AI, embedding it deeply into workflows and scaling it enterprise-wide. They weave it into development workflows and scale it across use cases."

That should all become easier with the rise of agentic AI, the company said, as agents will be able to better manage multiple stages of development.

Leaders find success with AI

Despite that, Bain still advised that the use of generative AI across wider processes, rather than just with coding, was boosting income for those who have led in deployments.

Some companies are delivering gains of between 10% to 25% by scaling the technology across core workflows, the company added.

"If you’re still piloting, you’re dangerously behind," noted David Crawford, Chairman of Bain’s Global Technology, Media, and Telecommunications practice, in the report.

For those companies that are struggling to keep up, Bain advised moving out of "experimental mode" by following the lead of companies that have already found success, as "they established a repeatable playbook that others can now follow – grounded in robust methodology, analytic tools, and clear benchmarks.

MORE FROM ITPRO

• Big tech promised developers productivity gains with AI tools – now they’re being rendered obsolete
• AI in software engineering: Six ways the profession is changing
• Microsoft says AI is finally having a 'meaningful impact' on developer productivity

In March this year, Anthropic CEO Dario Amodei claimed AI would be writing 90% of code previously churned out by software developers within just six months.

“I think we’ll be there in three to six months – where AI is writing 90% of the code,” he said, per reports from Business Insider. “And then in 12 months, we may be in a world where AI is writing essentially all of the code.”

Amodei is by no means the first – or likely last – industry figure to tout the long-term potential of AI in software development. Indeed, the technology is taking the industry by storm, with developers and engineers adopting AI coding tools en masse.

However, having passed the six-month mark, Amodei’s bold prediction hasn’t quite come to fruition – but there are clear signs the technology is gaining traction on this front.

Naturally, gauging exactly how much code is written by AI across a range of industries would be extremely challenging. Big tech companies have been keen to highlight their own gains on this front, though.

In November 2024, Google CEO Sundar Pichai revealed more than 25% of the company’s internal source code was AI-generated. This has likely grown since then.

Microsoft CEO Satya Nadella one-upped his Google counterpart in April when he confirmed around 30% of the tech giant’s code was written by AI. Nadella’s comments came during a fireside chat with Meta CEO Mark Zuckerberg at LlamaCon.

“I’d say maybe 20%, 30% of the code that is inside of our repos today and some of our projects are probably all written by software,” Nadella told Zuckerberg at the time.

Big tech execs have developers sweating

It goes without saying that comments from big industry figures like Amodei et al have had software developers sweating in recent months.

Late last year, Salesforce CEO Mark Benioff suggested the company might not hire any software engineers due to the rise of agentic AI, while Zuckerberg told Joe Rogan he also believes AI will replace engineers.

Comments like these have prompted concerns over the future of the profession and sparked discussions about whether developers will one day be rendered obsolete by the technology they’ve helped build.

For entry-level developers fresh out of college, the situation appears even more concerning, according to Amodei. The Anthropic chief exec told Axios earlier this year the technology will severely impact entry-level roles across a variety of professions.

Don’t believe the hype on AI coding

There are two problems with these types of predictions, however. First and foremost, developers of these solutions have a vested interest in hyping up their products – an issue we’ve highlighted repeatedly.

Secondly, much of the data out there at present doesn’t paint quite as dire a picture for developers as many would have you believe. Stack Overflow’s 2025 Developer Survey shows that practitioners are flocking to AI solutions and intend to ramp up their use of these tools in the year ahead.

Crucially, however, a significant number (46%) of developers “don’t trust the accuracy” of AI outputs and nearly two thirds (61.7%) revealed they have ethical and security-related concerns about AI-generated code.

Research from Fastly also shows that developers frequently find themselves manually fixing AI-generated code. The time spent carrying out fixes essentially offsets the time saved by using the tools, Fastly noted.

A similar study from Google itself even found many enterprises are concerned about AI-generated code quality, despite the company lauding its own gains on this front.

With this in mind, Amodei’s prediction on the scale of AI-generated code doesn’t quite add up yet. The aforementioned Google study did note that concerns over AI-generated code were dissipating, and alternative research does point to an influx of AI code – albeit on a scale far lower than what Amodei expected by this point.

Around 42% of developers admitted to having AI-filled codebases, according to recent analysis from Cloudsmith. Once again, trust and security concerns were major talking points.

Only 20% of developers said they trust AI-generated code “completely”, the study found.

MORE FROM ITPRO

• AI company Anthropic had to tell candidates to stop using AI when applying for jobs
• Anthropic’s new AI model could be a game changer for developers
• Anthropic CEO Dario Amodei lays out his vision of powerful AI — and explains why he sounds a tad pessimistic

The UK government has revealed its push to encourage AI use is delivering marked benefits for developers.

As part of an AI trial across government, more than 1,000 tech workers across 50 different departments trialed AI coding assistants from Microsoft, GitHub Copilot, and Google Gemini Code Assist between November 2024 and February this year.

Figures published following a review of the scheme shows developers are saving around one hour each day, equivalent to around 28 working days a year.

Technology minister Kanishka Narayan said the trial scheme highlights the benefits of rolling the technology out across government.

"For too long, essential public services have been slow to use new technology – we have a lot of catching up to do," said Narayan.

"These results show that our engineers are hungry to use AI to get that work done more quickly, and know how to use it safely."

Most of the time savings from the AI assistants came from using them to write first drafts of code that experts then edit, or using them to review existing code.

With only 15% of code generated by AI assistants being used without any edits, engineers have been taking care to check and correct outputs where needed, the government said.

The trial appears to have been popular with government coders, with 72% of users agreeing they offered good value for their organization. Nearly two-thirds (65%) reported that they were completing tasks more quickly and 56% said they could solve problems more efficiently.

Notably, more than half (58%) of participants said they would prefer not to return to working without these solutions.

Tara Brady, resident of Google Cloud EMEA, said the tech giant is “thrilled to see the positive impact” its AI coding tool delivered for government workers.

"This landmark trial, the largest of its kind for Gemini Code Assist in the UK public sector, underscores the transformative potential of AI in enhancing productivity and problem-solving for coding professionals, and highlights the successful collaboration stemming from Google Cloud’s Strategic Partnership Agreement with the UK government."

Prime minister Keir Starmer has been highly vocal about the government’s plans to roll out AI across government and public services since taking office. Downing Street hopes to save taxpayers more than £45 billion through the use of AI.

Questions remain over government AI coding gains

Martin Reynolds, field CTO at software delivery platform Harness, welcomed the move but questioned whether the plans go far enough. A key factor here, he noted, lies in the volume of manual remediation required by developers using AI-generated code.

"While AI is creating an initial velocity boost, 85% of government AI-generated code still needs to be manually edited by engineers,” he said.

“That's before it enters the more manual downstream stages of delivery, such as testing, security scanning, deployment, and continuous verification, which are essential to getting code into production safely and reliably," Reynolds added.

The quality of AI-generated code has become a recurring talking point in recent months. A recent study from Fastly, for example, found developers often find themselves manually remediating faulty code, which ultimately negates the time savings delivered by the technology.

Nigel Douglas, head of developer relations at software supply chain security firm Cloudsmith, also voiced concerns about potential security issues, saying there's not much evidence of secure-by-design thinking.

Given the critical nature of the work conducted by developers in key government departments, this should be a key focus moving forward.

“Without security-aware tooling or policy enforcement, you can easily see over-enthusiastic use of AI coding assistants unknowingly introducing vulnerabilities into one of this country’s most critical software ecosystems," he said.

"We’re getting past the point where it’s acceptable for software development teams to ‘hope for the best’ - you’ve got to be able to verify the provenance of the ingredients flowing through your software supply chain and into production systems, and you need tools to help respond to newly emerging threats that may impact what you’ve already deployed.”

MORE FROM ITPRO

• Not all software developers are sold on AI coding tools
• AI coding tools are booming – and developers in this one country are by far the most frequent users
• AI coding tools are finally delivering results for enterprises

Senior software developers are among the most receptive to AI code generation tools, according to new research.

A study from Fastly shows roughly one-third of senior developers, classed as those with over 10 years of experience, regularly use AI tools to build software. This is nearly two and a half times the rate of junior colleagues, the study noted.

While Fastly said the study highlights the growing appeal of AI-generated code, there are notable downsides – mainly fixing faulty code. Developers across all levels reported issues in this regard, with time savings often being negated by manual remediation.

“Senior developers were also more likely [than juniors] to say they invest time fixing AI-generated code,” the company said in a blog post detailing the research. “Just under 30% of seniors reported editing AI output enough to offset most of the time savings, compared to 17% of juniors.”

Remediating faulty code has emerged as a key hurdle for developers using AI tools in recent months.

Google’s 2024 State of DevOps report noted growing developer concerns about code quality, for example. A similar study from Cloudsmith in June 2025 warned some enterprises were overlooking key security considerations in a bid to drive efficiency.

Analysis from Harness in January this year found the issue had reached such an extent that developers were growing frustrated with the new tools at their disposal.

More than two-thirds (67%) reported spending more time debugging AI-generated code than prior to the introduction of the tools. Around 68%, meanwhile, now spend more time resolving vulnerabilities.

Regardless, Fastly found that senior developers appear more than content with using AI tools, and are confident in their ability to spot potential issues during development.

More than half (59%) of senior devs said AI tools “help them ship faster overall” while 49% of juniors held the same view. Senior staff were also more likely to report “significant speed gains” (26%) compared to less experienced colleagues, (13%).

Experience breeds confidence

According to Fastly, a key factor behind the confidence observed among senior developers is that they’re “simply better equipped to catch and correct AI’s mistakes”.

“They have the experience to recognize when code ‘looks right’ but isn’t,” the company said. “That makes them more confident at using AI tools efficiently, even for high-stakes or business-critical code.”

Meanwhile, junior developers are typically more cautious about relying on AI, or “more likely to avoid using it in production at all”.

“More experienced engineers are not only using AI more aggressively, but are also trusting it more in production environments,” Fastly added.

MORE FROM ITPRO

• AI coding tools are finally delivering results for enterprises
• AI coding tools are booming – and developers in this one country are by far the most frequent users
• How AI coding is transforming the IT industry in 2025

IT services and solutions provider Hexaware has announced a new strategic partnership with agentic AI software creation platform, Replit, to bring secure and governed ‘vibe coding’ to the enterprise.

The move will enable business users and engineers lacking coding experience to create their own secure, production-grade applications without compromising on compliance or security.

The partnership combines Hexaware’s digital innovation expertise with Replit’s natural language-based development platform, designed to support software creation and allow business teams beyond IT, such as sales and operations, to create no-code internal tools and prototypes at scale.

“Our mission is to empower entrepreneurial individuals to transform ideas into software—regardless of their coding experience or whether they're launching a startup or innovating within an enterprise,” explained Amjad Masad, CEO of Replit.

“Forward-thinking companies like Hexaware are exploring Replit to rapidly prototype, build internal tools, and address unique business challenges with custom software solutions.”

Secure vibe coding

On the security front, Hexaware said customers will have access to secure Vibe Coding environments that comply with IT policies while enabling “high-velocity innovation.”

Replit’s platform offers advanced features such as SSO, SOC2 compliance, private deployments, and role-based access controls, while Hexaware embeds enterprise-grade governance and security protocols directly into the development process.

With security in hand, product and design teams will be able to leverage Hexaware’s expertise in agile development to prototype new ideas and iterate at speed.

Sales, marketing, and operations teams can also build internal tools tailored to their specific workflow needs using Replit, eliminating the bottlenecks often associated with SaaS tools and long IT development queues.

Additionally, the partnership aims to free up users to focus on high-level logic and business decisions, with autonomous AI agents on hand to tackle the complex and time-consuming backend tasks typically undertaken by staff.

AI-first commitment

Hexaware said the collaboration with Replit forms part of its wider goal of pioneering secure, scalable, and transformative AI-first platforms geared towards helping enterprises achieve digital agility.

“This partnership underscores Hexaware’s commitment to democratizing software innovation with security and governance at its core,” explained Sanjay Salunkhe, Hexaware’s president and global head of Digital and Software Services.

“By combining our Vibe Coding framework with Replit’s natural language interface, we’re giving enterprises the tools to accelerate development cycles while upholding the rigorous standards their stakeholders demand.”

MORE FROM CHANNELPRO

• Nasuni targets fresh growth under revamped leadership team
• Monday.com expands partner program with new AI specializations
• Workday snaps up AI-powered conversation recruitment platform

Researchers at Microsoft say AI is having a net positive effect on software development, with the technology “augmenting developers rather than replacing them”.

The claim follows the publication of a study examining the use of the technology amidst rising concerns over its impact on the workforce.

The use of AI in software development has become a source of “intense debate” in recent months, researchers noted, with some industry stakeholders suggesting it will ultimately replace developers and reduce entry-level opportunities.

“Some believe AI will soon replace entire categories of developer work, while others dismiss it as an overhyped distraction,” the study noted. “But amid all the speculation, one perspective is often missing: that of the developers who are actually using these tools”.

“To better understand how AI is reshaping the developer experience, we conducted a study that asked developers how AI tools are affecting their work,” researchers added.

A key focus of the study centered around how AI is impacting productivity across a range of metrics, using a framework known as 'SPACE'.

This refers to:

• Satisfaction
• Performance
• Activity
• Collaboration
• Efficiency

How AI is changing software development

According to Microsoft, the adoption of AI tools in software development is “no longer a niche experiment”, with the use of the technology now becoming the norm across the industry.

Three-quarters (75%) of developers told researchers they regularly use AI to complete tasks, for example. This, the study noted, highlights a “clear shift” in how development teams are integrating AI into their daily workflows.

“Among those who have adopted AI, usage is frequent,” the report stated. “64% of developers who use AI report using it at least once per week.”

Microsoft’s stats roughly align with alternative figures highlighting growing adoption rates among developers. Stack Overflow’s 2025 Developer Survey showed the use of the technology has surged over the last 12 months, with 84% of devs now using - or planning to use - AI tools.

This not only marked a significant jump from 76% in last year’s survey, but a third consecutive annual increase in adoption rates.

Notably, the seniority of developers doesn’t appear to have a major impact on adoption rates, with experienced devs only 4% less likely to be daily users compared to junior staff.

Organizational encouragement does have an impact on adoption rates, however.

Respondents who said their employer “actively advocates” for the use of the technology are “seven times more likely to be daily users than those who feel their company does not”.

Productivity gains are becoming clear

Productivity and efficiency gains have been key talking points for providers of AI developer tools over the last two years, and it’s worth noting that Microsoft itself has been among the most vocal proponents of the technology in this domain.

Among developers who regularly use AI in their daily workflows, 90% told researchers it makes them more productive. Similarly, 80% of developers said they “would be sad if they could no longer use it”.

“These results make it clear,” the study noted. “Developers who use AI believe it has a meaningful impact on their productivity.”

Exactly what constitutes “productivity” here is multi-faceted, however. Under the study’s SPACE framework, researchers said this “spans multiple dimensions” – from performance, efficiency, and task throughput, to job satisfaction and collaboration rates.

Regardless, using this framework, researchers said the signs remain positive and the technology is having a beneficial effect on teams. Among those actively using AI on a daily basis, developers said the technology improves productivity “across a wide range of dimensions”.

A significant majority (88%) said AI has improved their task throughput while a similar number (82%) revealed it has improved overall efficiency. Researchers said this aligns closely with “previous findings about the efficacy of Copilot to decrease the time it takes to complete coding tasks”.

While speed and performance gains were identified as key benefits, 71% said AI tools are helping them deliver better customer value. Almost two-thirds (62%) also revealed the technology is improving job satisfaction, largely due to the reduction of manual toil.

“These results suggest that, for those who use AI, it’s not just making them faster, it’s helping them feel more effective and more satisfied with their work,” researchers said.

“Previous research has shown a positive relationship between productivity and satisfaction , and it would appear that these results lend agreement to that.”

AI is isn’t quite improving cohesion

Microsoft said the impact of AI on collaboration “tells a more nuanced story”. For example, fewer than half (48%) of respondents said AI tools are actively improving their ability to collaborate with teammates.

This doesn’t mean that AI is harming collaboration, however. Indeed, researchers noted that very few developers reported decreases on this front while a “significant portion” remained neutral.

What is does suggest is that the role of AI within a team-based environment is still evolving, with teams unsure over how it can improve cohesion. Other findings on this front showed that AI is “changing the nature of interactions” within teams, suggesting tentative improvements.

“The engineering managers we spoke to reported fewer interruptions within their teams, as developers relied less on colleagues for quick answers to coding questions,” researchers said.

Overall, many senior software developers said the use of AI tools has actually improved the “quality and depth” of conversations and sparked new discussions over strategy.

“A VP of IT at a global supplier of technology and engineering systems explained that, because of ChatGPT, teams now spend more time ‘brainstorming about projects, ideas, and architectures’ and less time discussing ‘unnecessary’ coding questions,” the study said.

MORE FROM ITPRO

• Microsoft says AI tools such as Copilot or ChatGPT are affecting critical thinking at work
• AI coding tools are finally delivering results for enterprises
• Big tech promised developers productivity gains with AI tools – now they’re being rendered obsolete

Google publicly released its Jules coding agent earlier this month, and it’s already had a big update aimed at shoring up code quality.

Jules is intended as an ‘asynchronous’ coding agent, capable of working in the background alongside developers to write coding tests, fix bugs, and apply version updates across code.

Users working with the agent can delegate complex coding tasks based on project goals, marking a contrast to manually-operated vibe coding models and AI coding assistants such as GitHub Copilot or Google Code Assist.

Once Jules has access to an organization’s repositories, it clones the code contained within to a Google Cloud virtual machine (VM) and makes changes based on user prompts.

As a safety precaution, Jules does not make permanent changes to code before a human has manually approved its decisions - an issue that has caused serious problems for vibe coding users in recent weeks.

When its outputs are completed, Jules presents them as a pull request to the relevant developer, with a breakdown showing its reasoning for the changes it made and the difference between the code in its existing branch and its own proposed code.

Under the hood of the Jules coding agent

Jules is powered by Gemini 2.5 Pro, Google’s flagship LLM which is capable of ‘reasoning’ through complex problems before acting. Boasting a one million token context window, Jules is capable of using large swathes of an organization’s existing code for context.

Google has stated that Jules runs privately by default and is not trained on the private code to which it’s exposed.

Although Jules is free to use, the entry level tier only allows users to complete 15 tasks per day and just three concurrent tasks at any given moment.

Enterprises looking to use it at scale may contact their Google Cloud sales representative to add Google AI Ultra to their Workspace plan. This provides access to Jules for 300 daily tasks, as well as 60 concurrent tasks.

Independent developers or other interested users can pay for £18.99 ($19.99) per month for AI Pro, for which they’ll get 100 daily and 15 concurrent tasks, or £234.99 ($249.99) per month for AI Ultra.

Those on paid plans also have priority access to Google’s latest model. In practice, this means all plans currently use Gemini 2.5 Pro, but in the near future subscribers will be able to use more advanced models to power their Jules outputs.

Jules comes with constructive self-criticism

AI-powered coding is becoming increasingly popular, with Google’s own internal code now standing at more than 25% AI-generated and a recent Stack Overflow study finding that 84% of software developers now use AI.

But nearly half (46%) of respondents to the same study said they don’t trust AI accuracy, while three-quarters (75.3%) said they would turn to co-workers when they don’t trust AI answers.

AI-generated code also comes with inherent risks and limitations. Almost half of the code generated with leading models contains vulnerabilities, according to recent research by Veracode, including cross-site scripting (XSS) and SQL injection flaws.

To address these concerns, Google has also released a ‘critic’ feature within Jules, which subjects all proposed changes to adversarial reviews at point of generation in an attempt to ensure the code generated by the agent is as robust and efficient as possible.

When the critic detects code that needs to be redone, it sends it back to Jules to be improved. Google gave the example of code that includes a subtle logic error being handed back to Jules with the explanation: “Output matches expected cases but fails on unseen inputs”.

Google differentiated the critic from linters, common code analysis tools that can be found in coding agents from competitors such as OpenAI’s Codex, due to its ability to judge code relative to the user’s context and intent rather than assessing code quality based on presets.

It added that in the future the critic will become an agent in its own right, with the ability to draw on tools such as an AI code interpreter or search engines to gain more context on code, as well as to trigger criticism of code generation at multiple points throughout Jules’ reasoning process.

These ambitions are due to be realized in future versions, with no announced release timeframe.

“Our goal is fewer slop PRs, better test coverage, stronger security,” Google wrote in its blog post announcing the feature.

MORE FROM ITPRO

• GitHub's new 'Agent Mode' feature lets AI take the reins for developers
• AI coding tools are finally delivering results for enterprises
• Everything you need to know about OpenAI’s new agent for ChatGPT

A majority of enterprises are using a version of Python that’s a year or more older, according to new research, and it’s costing them dearly.

Statistics from JetBrains’ annual Python Developers Survey show over three-quarters (83%) of enterprises are using older versions, highlighting a broad “distribution across the latest and older versions of the Python runtime”.

They’re not particularly fussed, however, with the majority of respondents stating that they were satisfied with the performance of the version they’re currently using.

More than half (53%) said the version they’re using “meets all my needs” while 25% admitted they simply haven’t had time to update.

Breaking down usage rates, JetBrains’ study showed just under a quarter (21%) reported using version 3.11, released in October 2022, while 35% said they’re working with version 3.12.

Just 15%, the study found, are using 3.13, the latest version of the programming language.

Notably, the study showed that a significant number of developers are using Docker and containers to execute code, meaning the volume of enterprises working with older versions was “surprising”.

“With containers, just pick the latest version of Python in the container,” the company noted in a blog post detailing its findings. “Since everything is isolated, you don’t need to worry about its interactions with the rest of the system, for example, Linux’s system Python.”

JetBrains added that it expects containerization to “provide more flexibility” for enterprises and “ease our transition” to the latest version of Python.

Upgrading Python has big performance benefits

A key talking point in the JetBrains survey centers around the potential ease of use improvements enterprises are missing out on by failing to upgrade.

More recent versions include features that ease quality of life for developers, but also big performance gains, the company noted.

“The 83% of developers running on older versions of Python may be missing out on much more than they realize,” the firm said. “It’s not just that they are missing some language features, such as the except keyword, or a minor improvement to the standard library, such as tomllib.”

“Python 3.11, 3.12, and 3.13 all include major performance benefits, and the upcoming 3.14 will include even more.”

Drilling down into the potential gains, JetBrains noted that 48% of those using Python 3.11 can unlock code run gains of ~11% by upgrading to 3.13 - and that’s achieved while using ~10-15% less memory.

“If they are on the 27% still on 3.10 or older, their code gets a whopping ~42% speed increase (with no code changes), and memory use can drop by ~20-30%,” JetBrains added.

Speaking to ITPro, Daniele Volpe, software engineer at SecureFlag, said while updating to newer versions isn’t critical, the advantages are clear and should be capitalized on.

“Running old Python versions does add up over time,” Volpe said. “You’re basically paying for slower runtimes and higher cloud costs when newer versions give you efficiency for free.”

“It’s not always urgent, but ignoring it means you’re leaving money and performance on the table”.

Volpe noted that the impact on developers has a knock-on effect which ultimately exacerbates existing efficiency and performance-related issues.

Slower runtimes mean slower feedback loops, which slows down the overall efficiency of the development lifecycle.

“Developers spend more time waiting on tests, builds, or jobs to finish,” he said. “Plus, sticking with older versions means missing out on language improvements that could make their day-to-day work smoother.”

Upgrading might be a hassle – but just take the plunge

Volpe said that it’s no surprise a significant portion of respondents are satisfied with the performance of older Python versions. It’s a long-running practice and most teams “stick with what works until something breaks”.

“Upgrading feels like hassle and risk, so they put it off,” he said. “Comfort and inertia often win out, even if the gains from moving forward would be worth it.”

Maurice Kalinowski, product director at Qt, echoed Volpe’s comments on this front, adding that “updating any technical dependency is always a concern to a development team”.

“There can be regressions, both on functional or performance level, incompatible changes or even removed functionality,” Kalinowski said.

“This applies even more to the underlying programming language, as it touches not just one component but your whole solution.”

Simply put, if it isn’t broken, don’t try and fix it.

MORE FROM ITPRO

• AI is transforming software development – JetBrains CEO Kirill Skrygan says it’s up to developers to transform with it
• AI coding tools are finally delivering results for enterprises
• Want developers to build secure software? You need to ditch these two programming languages

AI is better at coding than most humans, according to software developers, but there are several glaring issues holding the technology back.

A survey by Clutch revealed that more than half (53%) of senior software developers believe large language models (LLMSs) can already code better than most humans.

The study comes as AI tools continue to push into software development, with solutions like GitHub Copilot, Cursor, and Windsurf, which OpenAI tried to snap up for $3bn in May before losing out to Cognition.

Given that, it's no surprise that three-quarters of those asked by Clutch said they expect AI to "significantly reshape" how software is developed over the next five years.

"That shift is already underway," the company said in a blog post detailing its findings.

"From the rise of prompt engineering to debates about AI-generated code quality, software leaders are grappling with rapidly evolving development models and updating their hiring, upskilling, and automation strategies to keep up."

According to the survey, 78% of those asked said they already use AI several times a week or more — that matches near enough to a recent survey from Stack Overflow that suggested 84% of developers were using, or plan to use, AI in their daily workflows.

That use of AI doesn't necessarily bother developers, with 42% reporting positive feelings about AI and another 23% saying they were "excited" by the technology.

Concerns about AI

But it's not all good news, the survey noted. One-in-ten reported concerns about how AI is used in software development with 8% skeptical about the technology.

The top concern was data privacy (24%), followed by job displacement (14%), the risk of errors in work (14%), and the loss of creativity (13%).

A further 7% raised the issue of a lack of entry-level roles to help junior developers join the industry. The impact on developers entering the workforce has become a key talking point in recent months, with industry stakeholders suggesting the technology could seriously hamper opportunities for graduates.

"When asked how AI might affect entry into the profession, opinions were mixed," Clutch noted in the blog post.

"45% of respondents said AI might actually lower the barrier for junior developers by giving them better tools and faster ways to learn. But 37% said it would do the opposite, making it harder for newcomers to compete or even get noticed. AI's ability to automate junior-level work is reshaping hiring criteria."

Beyond that, 79% believe that AI skills will soon become a must-have to get hired.

How developers are using AI

How AI is being used varies, according to Clutch. The survey found that 48% of developers use AI primarily for code generation, with 36% using it in testing phases and a further 36% during code review.

A minority of developers also use AI earlier in the development cycle for requirement gathering and system design, or even post-launch for debugging.

"These stages demand speed, consistency, and pattern recognition, which makes them a natural fit for automation," the post noted.

But Clutch argues it's not just about automation - the technology is enabling development teams to overhaul traditional processes and find new ways of working.

"It’s not just about doing the same tasks faster,” Clutch noted. “It’s about doing them differently, with new methods for debugging, testing, and prototyping that weren’t practical before."

One area of concern highlighted in the survey centered around AI-generated code. A host of major tech companies, including Microsoft and Google, have touted their own gains on this front in recent months.

More than half (59%) of respondents said they used AI-generated code without fully understanding it, for example. This, the study noted, could create new security risks for organizations.

In a study last year, researchers warned that generative AI could replicate insecure code. A more recent study from Cloudsmith also warned that development teams are walking into a trap with AI-generated code, with many placing too much faith in the technology.

"That gap between speed and understanding is something teams will need to work on," Clutch noted.

MORE FROM ITPRO

• AI coding tools aren’t the solution to the unfolding 'developer crisis’
• AI is finally delivering results for software developers
• Developers in this one country are the most frequent users of AI coding tools

Two high-profile incidents involving AI coding assistants show why developers might be justified in remaining hesitant of the solutions.

The CEO of Replit, a vibe coding platform which recently secured a big-money partnership with Microsoft, issued a public apology to a user last month after the tool deleted a company’s entire codebase during a test run.

In a series of posts on X, SaaStr founder Jason Lemkin revealed the AI assistant deleted the company’s database despite being prompted not to alter code without express permission.

Lemkin had previously detailed his experiences with the AI coding tool, revealing it was “pretty cool” on first impressions. However, he noted that the AI began essentially hiding its errors and fabricating data.

“It kept covering up bugs and issues by creating fake data, fake reports, and worst of all, lying about our unit test,” he wrote.

The incident reached a head when the tool deleted a database containing over 1,200 executive records, as well as data on around 1,200 companies.

Notably, when asked to assess the severity of its mistake, the tool responded: "Severity: 95/100. This is an extreme violation of trust and professional standards”.

“I made a catastrophic error in judgement,” it responded to a previous prompt on the changes.

Luckily, Lemkin said the AI assistant’s rollback feature prevented complete disaster on this occasion, although the tool informed him this wasn’t possible.

“Replit assured me its built it rollback did not support database rollbacks,” he wrote. “It said it was impossible in this case, that it had destroyed all database versions.

“It turns out Replit was wrong, and the rollback did work. JFC.”

More AI coding carnage

This wasn’t an isolated incident in July, either. Just days later, a similar situation unfolded for a user of Google’s Gemini command line interface (CLI) tool which saw data destroyed.

In a now-deleted post on GitHub, Anuraag Gupta, a product lead at Cyware, was experimenting with the open source coding tool. According to reports from Mashable, Gupta requested to move files from previous Claude coding activities to a new folder.

Gupta noted that the new location of the files, as stated by the AI tool, were not accurate and they had seemingly vanished. Gemini later acknowledged the files had been destroyed.

“I have failed you completely and catastrophically,” the AI model said. “My review of the commands confirms my gross incompetence. The mkdir command to create the destination folder likely failed silently, and my subsequent move commands, which I misinterpreted as successful, have sent your files to an unknown location.”

The model added that due to “security constraints" it was unable to search outside the project directory, noting that this was an “unacceptable, irreversible failure”.

Developers are wary of AI - and for good reason

While these incidents were experimental and didn’t result in complete disaster, they both highlight the potential issues developers might face when relying on the technology in real-world enterprise environments.

AI coding tools have been framed as a game changer for the industry over the last 18 months, with big tech providers specifically highlighting the productivity benefits of the technology.

Research shows these tools are having a positive impact on productivity and efficiency, but figures from Stack Overflow’s 2025 Developer Survey show engineers and devs are still wary.

The survey found 84% of developers currently use - or plan to use - AI tools in their daily activities, a sharp increase compared to last year’s edition. Running parallel to this growth is a sense of caution, however, with nearly half (46%) of respondents noting that they “don’t trust the accuracy” of AI tools.

This, Stack Overflow stressed, marked a significant increase in the number of developers wary of the technology compared to the year prior.

All told, three-quarters (75.3%) said they don’t trust AI answers at all and frequently refer to co-workers for advice when using the technology. Nearly two-thirds (61.7%) said they also have ethical and security concerns about AI-generated code.

MORE FROM ITPRO

• Want developers to build secure software? You need to ditch these two programming languages
• Not all software developers are sold on AI coding tools
• Big tech promised developers productivity gains with AI tools – now they’re being rendered obsolete

Software developers are using AI at record levels, new research shows, but they're hesitant to place too much faith in the technology.

Findings from Stack Overflow’s annual Developer Survey show the use of technology in the software industry has surged over the last 12 months, spurred on by the emergence of AI coding tools and, most recently, agentic AI solutions.

The survey found 84% of developers currently use - or plan to use - AI tools in their daily workflows. This marks a third consecutive yearly increase and a jump from 76% in last year’s edition.

OpenAI’s GPT model range was cited as the most frequently used by developers alongside Anthropic’s Claude Sonnet range and Google’s Gemini Flash models.

Yet despite the increasing uptake of AI tools in the profession, a growing number of developers aren’t willing to put their trust in the technology. Nearly half (46%) said they “don’t trust the accuracy” of the output from AI, which marks a significant increase compared to 31% in the 2024 survey.

Notably, even if AI improves to the extent that it can carry out tasks on behalf of developers, many said they would still prefer to ask a colleague for assistance.

Three-quarters (75.3%) said they simply don’t trust AI answers and would refer to a co-worker while 61.7% revealed they frequently have ethical or security-related concerns about AI-generated code.

Crucially, 61.3% said they’d refer to a colleague for assistance because they want to fully understand their code.

Prashanth Chandrasekar, CEO of Stack Overflow, said the findings show that while the benefits of AI are becoming clearer, lingering concerns about security and safety still remain.

“The growing lack of trust in AI tools stood out to us as the key data point in this year's survey, especially given the increased pace of growth and adoption of these AI tools,” Chandrasekar said.

“AI is a powerful tool, but it has significant risks of misinformation or can lack complexity or relevance.”

Chandrasekar added that pursuing an approach that “leans heavily on trustworthy, responsible use of data from curated knowledge bases” is now critical.

Similarly, ensuring humans are kept in the loop is an equally important consideration for development teams.

“By providing a trusted human intelligence layer in the age of AI, we believe the tech enthusiasts of today can play a larger role in adding value to build the AI technologies and products of tomorrow.”

Developers are bogged down in buggy code

While trust in AI tools was a key issue highlighted by developers, growing frustration over poor quality output was also raised.

Nearly half (45%) of respondents said they’re bogged down in time-consuming debugging as a result of AI-generated code. The survey findings align with previous research highlighting concerns over AI code generation.

In a study from Cloudsmith, nearly half of developers using AI said their codebases are now largely AI-generated. While devs said the technology has enabled them to improve productivity and efficiency, the study warned that a concerning number of teams are overlooking crucial security considerations.

Around a third of developers don’t review AI-generated code before deployments, Cloudsmith found.

Agentic AI hasn’t hit the mainstream yet

Agentic AI has emerged as the latest trend in the AI space this year, yet while it’s taken the industry by storm, uptake of agentic solutions hasn’t yet taken off in the development community, according to Stack Overflow.

Only 31% of developers said they used AI agents in their daily workflows, while just 17% plan to use them. More than one-third (38%) said they don’t plan to use AI agents at all in the near future.

Notably, among those that have used AI agents, around 69% agreed they experienced a marked increase in productivity and efficiency.

A host of major industry players have jumped on the agentic AI bandwagon in recent months, with Salesforce in particular dialing in on the technology. Earlier this year, CEO Marc Benioff suggested the company might not hire any new software engineers as a result of its own internal successes with the technology.

With the emergence of this latest iteration of the technology, some industry stakeholders have suggested AI might negatively impact the workforce, particularly entry-level roles.

Stack Overflow’s survey shows developers aren’t concerned about the prospect of AI-related job cuts, however. Nearly two-thirds (64%) of respondents said they don’t perceive AI as a threat to their jobs, marking a decrease compared to the 2024 survey.

MORE FROM ITPRO

• AI coding tools are booming – and developers in this one country are by far the most frequent users
• Want developers to build secure software? You need to ditch these two programming languages
• AI tools aren’t helping developers save time on software testing

AI might actually be slowing down software developers, new research suggests, despite their own positive predictions that AI coding tools will help their work.

Model Evaluation & Threat Research (METR) ran a randomized, controlled trial to study the use of early-2025 AI tools used by experienced developers working on their own open source repositories.

The results weren't positive, with the study finding that developers take 19% longer on average when using AI tools.

That said, METR noted this is a single snapshot of AI capabilities in a single setting — results may vary, METR's essentially saying — and the company hopes to use the figures as a baseline to track and estimate AI acceleration and adoption.

Experienced developers aren’t seeing benefits

METR's study used 16 developers with at least five years of experience who provided a list of 246 issues, such as bug fixes or wished-for features, that would make up their real day-to-day work. The developers in the study largely used Cursor Pro and Claude Sonnet.

"When developers are allowed to use AI tools, they take 19% longer to complete issues—a significant slowdown that goes against developer beliefs and expert forecasts," the post noted.

"This gap between perception and reality is striking: developers expected AI to speed them up by 24%, and even after experiencing the slowdown, they still believed AI had sped them up by 20%."

The METR team listed a few factors they believe likely contributed to the slowdown — the first of which was simply being over-optimistic about the usefulness of AI.

Beyond that, the study showed developers accepted fewer than 44% of AI generated code, with most of those taking part in the study saying they had to make major changes to clean up code. All told, roughly 9% of their time was spent reviewing or cleaning AI output.

Developers reported that AI appeared to struggle in large, complex environments, while the human coders themselves had so much experience in their areas of expertise that the AI couldn't keep up, an issue exacerbated by the fact that it "doesn't utilize important tacit knowledge or context."

The researchers also spotted a few potential causes for the AI slowdown, though they stressed the impact was unclear. That included developers using the AI to experiment and waiting on AI to generate code.

The researchers stressed that the results don't mean that AI is never useful. "Furthermore, these results do not imply that future models will not speed up developers in this exact setting — this is a salient possibility given the rapid pace of progress in AI capabilities recently," the paper said.

"Finally, it remains possible that further improvements to current AI systems (e.g. better prompting/agent scaffolding, or domain-specific finetuning) could yield positive speed up in this setting."

Does AI help or hinder?

Studying real-world impact is necessary, METR said, as AI benchmarks are often unrealistic or can't be easily applied to real-world tasks.

"While coding/agentic benchmarks have proven useful for understanding AI capabilities, they typically sacrifice realism for scale and efficiency—the tasks are self-contained, don’t require prior context to understand, and use algorithmic evaluation that doesn’t capture many important capabilities," the post noted.

"These properties may lead benchmarks to overestimate AI capabilities."

The METR study follows research by Atlassian that suggested AI tools are saving 86% of developers a full day of work each week, letting them focus on improving code quality or creating new features.

However, half of those surveyed said they are losing that AI-gained time because of "organisational inefficiencies" — essentially, bad management is eating up any savings.

"This pressure-cooked mix of innovation and strain demands a closer look at how AI is reshaping the developer experience, and what that means for the future of software development across the industry,” said Atlassian CTO Rajeev Rajan.

MORE FROM ITPRO

• CEOs think workers are becoming hostile to AI tools, but they’re pushing ahead with adoption anyway
• AI coding tools are finally delivering results for enterprises
• Big tech promised developers productivity gains with AI tools – now they’re being rendered obsolete

Cognition has confirmed plans to acquire Windsurf just days after a rumored $3 billion buyout attempt by OpenAI fell through and the AI coding startup's CEO was poached by Google.

In May this year, OpenAI was reported to be on the cusp of purchasing Windsurf, formerly known as Codeium. Windsurf offers tools such as Autocomplete, which suggests code while software engineers type, and Search, which helps find bits and pieces in a repository using natural language.

Last week, the long-rumored deal fell through amid reports that one cause was OpenAI's ongoing tension with Microsoft. In its place, Google stepped in with a licensing deal and to poach key staff.

Days later, Windsurf and Cognition jointly announced the latter would acquire the sought-after AI startup. Financial details of the deal weren't immediately revealed, but Cognition CEO Scott Wu said in a blog post that the acquisition includes Windsurf's IP, product, trademark, business, and talent.

Windsurf's new interim CEO, Jeff Wang, said it was a "really volatile time at Windsurf", but added: "Despite the turbulence, so much of what makes us great is intact."

Inside the Cognition acquisition

Cognition is perhaps best known as the creator of Devin, which it claimed was the world's first autonomous software engineer — though critics noted plenty of flaws with the $500-a-month system at its launch earlier this year.

Wang said pairing Cognition's agents with Windsurf's AI-supported integrated development environment (IDE) was the "perfect combination of products", and suggested it would enable AI to take over more coding, helping support everyone from vibe coders to large enterprises.

"Working side by side, we’ll be able to let engineering teams plan a task in Windsurf aided by Devin’s deep codebase understanding, delegate chunks of work to a team of Devins, and tackle the hardest parts yourself in Windsurf with features like Tab and Cascade," Wang said.

"Then it seamlessly gets stitched back together all within the same environment. That’s the future we will now create together."

"This combined agent + IDE, along with everything else to come, will be a fearsome combo,” he added.

Cognition's Wu said Windsurf will continue to work as it had been previously while Cognition focuses on developing Devin.

"Over the coming months, we’ll be investing heavily in integrating Windsurf’s capabilities and unique IP into Cognition’s products," he said.

Google’s talent scouting bombshell

Last week, before the Cognition deal was announced, Google confirmed it paid $2.4 billion for licensing rights and to poach a small number of Windsurf staff.

This included a portion of the R&D team, but also CEO Varun Mohan and co-founder Douglas Chen. Jeff Wang, previously head of business, took over as interim CEO.

"We’re excited to welcome some top AI coding talent from Windsurf’s team to Google DeepMind to advance our work in agentic coding," Google said in a statement sent to journalists.

Windsurf explained in a blog post that the software it was building "aligned quite well" with other companies' products, but admitted there are now many different solutions outside of Windsurf to help support software development.

"Given the rapid pace of innovation, we see an advantage to double down our focus on the enterprise problems, which has long been our primary focus, and we will be continuing to devote resources to taking the wide range of product innovations in the broader market and making them work for enterprise workloads, the most impactful workloads to society," the company said.

The deal with Google will help "kick-start this next phase."

Tough times at OpenAI

The failed deal between OpenAI and Windsurf was reportedly impacted by the AI firm’s tensions with major investor Microsoft, which have been rumbling on in the background over the last few months.

Microsoft was an enthusiastic backer of OpenAI, with an initial $1bn investment in 2019 that has since totaled $13bn. Relations between the two firms appear to have become frayed in recent months.

Earlier this year, Microsoft began using a wider range of AI models in 365 Copilot beyond OpenAI's systems, and the once exclusive relationship opened up to allow OpenAI to work with other infrastructure partners.

Last month, reports suggested the relationship had soured further, with a report in the Wall Street Journal claiming "tensions are reaching a boiling point".

According to sources, OpenAI execs were considering accusing Microsoft of anti-competitive behaviour and Microsoft has reportedly considered pausing negotiations.

The disagreement appears to center on OpenAI's attempts to shift away from its existing non-profit model, how much control Microsoft has over OpenAI's compute resources, and even the access Microsoft would have gotten to Windsurf's IP had that deal gone through.

Another key aspect of the dispute appears to center on a clause in the contract between the two companies over what happens if OpenAI achieves artificial general intelligence (AGI).

If OpenAI achieves that perhaps impossible milestone, then Microsoft's access to its models will automatically end. However, that raises the question of how to define AGI for the purposes of that clause, with Microsoft worried — according to a Wired report — that its partner will pull the plug on its model access early.

CEO Sam Altman has been highly vocal about the company’s push to achieve this elusive goal, going so far as to predict AGI’s arrival this year. On the other hand, Microsoft CEO Satya Nadella has called the idea of AGI "nonsensical benchmark hacking".

Running parallel to all of this is the brewing war for AI talent in Silicon Valley, with Meta poaching OpenAI staff by throwing hundreds of millions of dollars in pay at top talent.

MORE FROM ITPRO

• AI coding tools are finally delivering results for enterprises
• AI coding tools are booming – and developers in this one country are by far the most frequent users
• AI coding tools aren’t the solution to the unfolding 'developer crisis’

The popularity of AI coding tools has skyrocketed over the last year and a half, with developers flocking to these solutions to supercharge their daily workflows.

A host of major industry players, including Microsoft, Google, and Amazon Web Services (AWS), have all unveiled dedicated tools for software development.

A new research paper shows there’s a growing geographic disparity in terms of the popularity of AI coding tools.

The paper, titled ‘Who is using AI to code? Global diffusion and impact of generative AI’, examined the use of AI in software development globally, finding that US developers are among the most frequent users.

Researchers tracked AI-generated Python functions in GitHub to identify the regions where the technology was being used most frequently, noting that by December 2024 around 30.1% Python functions came from US contributors.

This comes in stark contrast to counterparts elsewhere globally, researchers found. Around 24.3% of Python functions came from users based in Germany, while France and India stood at 23.2% and 21.6% respectively.

Notably, contributions from Russia and China lagged significantly, recorded at just 15.4% and 11.7% respectively.

Broadly speaking, US developers have led consistently over the last few years, the paper found, particularly since the advent of the generative AI ‘boom’ and running in parallel to various model releases since late 2022.

“We observe noticeable growth spikes in AI-generated code soon after key generative AI releases such as GitHub Copilot, the original ChatGPT, and GPT-4.0, highlighting how breakthroughs in LLMs prompt rapid uptake by developers worldwide,” the paper reads.

“Yet, we also observe significant differences among countries. In the US, which leads in the intensity of use of AI throughout our data, AI-written code grows from 0% in 2020 to approximately 30% by the end of 2024,” researchers added.

“Contributions from Germany, France, India, Russia, and China vary between 12 and 24% in 2024, suggesting significant geographic barriers to diffusion.”

The researchers’ findings on the popularity of AI-generated code align with recent confirmations from major tech firms on this front. In November last year, Google CEO Sundar Pichai revealed that around 25% of the tech giant’s internal code was written using AI, and this has likely since grown.

More recently, Microsoft chief executive Satya Nadella also highlighted the scale of the company’s AI-generated code, with this now standing at around 30%.

How developers are using AI coding tools

Perhaps unsurprisingly, more experienced and senior developers were less likely to use AI coding tools compared to their junior colleagues. This aligns with existing studies on the uptake of these solutions based on seniority, researchers noted.

The impact of AI on junior developers has become a recurring talking point in recent months, with studies warning that entry-level roles could be adversely impacted by the technology.

Some industry stakeholders have also warned that frequent use of the technology is creating an overreliance that undermines core coding skills.

But those entering the profession are more likely to be accustomed to using these tools, according to GitHub CEO Thomas Dohmke.

In a recent podcast appearance, Dohmke noted that many junior developers are ‘AI native’ and have been using AI throughout their time at college and university, or even earlier on in high school, making them more comfortable with using the technology.

The economic impact of AI in software development

Industry providers have been keen to highlight the benefits of AI coding tools, and AI solutions more broadly, in recent years. Productivity boosts, efficiency gains, and speedier development processes have all been touted as key advantages.

Research from GitHub last year, for example, showed these solutions are delivering marked benefits for users, allowing developers to streamline tasks in their daily workflow and focus on more pressing matters.

However, the research paper noted that the economic benefits of the technology vary wildly, and are often dependent on whether organizations and teams are effectively using AI.

Researchers noted that the economic value of AI-assisted coding in the US stands at anywhere from $9.6 billion and $14.4 billion annually.

MORE FROM ITPRO

• Big tech promised developers productivity gains with AI tools – now they’re being rendered obsolete
• Developers are at their wits end trying to build generative AI applications
• Want developers to build secure software? You need to ditch these two programming languages

Hundreds of Model Context Protocol (MCP) servers around the world are open to abuse, with vulnerabilities that put vibe coders and their organization's sensitive assets at risk.

Introduced late last year, MCP servers are an easy-to-use extension of LLMs, thanks to the simplicity of their protocols, and have come into widespread use due to the broad availability of independently developed MCPs.

However, according to analysis from Backslash Security, around half of the 15,000-plus MCP servers in existence are dangerously misconfigured or carelessly built. The resulting vulnerabilities are in some cases catastrophic, the company warned.

They fall under two general headings. First is the MCP ‘NeighborJack’ vulnerability, whereby hundreds of MCP servers are explicitly bound to all network interfaces (0.0.0.0), making them accessible to anyone on the same local network.

This was the most common vulnerability found, with hundreds of cases discovered.

"Imagine you’re coding in a shared co-working space or café. Your MCP server is silently running on your machine," the researchers said.

"The person sitting near you, sipping their latte, can now access your MCP server, impersonate tools, and potentially run operations on your behalf."

Meanwhile, dozens of MCP servers allowed arbitrary command execution on the host machine thanks to careless use of a subprocess, a lack of input sanitization, or security bugs such as path traversal.

Most concerning of all, on several MCP servers both vulnerabilities were present, allowing bad actors to take full control of the host machine running the server.

Malicious actors that come across these MCP servers would have full access to run any command, scrape memory, or impersonate tools used by AI agents, Backslash said.

Meanwhile, beyond code execution, MCPs can serve as stealthy pathways for prompt injection and context poisoning, Backslash warned. Malicious or manipulated public content can change what an LLM sees - returning misleading data, or rerouting agent logic.

“Our research highlights several prevalent MCP server weaknesses that can open enterprise environments to threat vectors including remote code execution, data exposure, and network traversal,” said Yossi Pik, co-founder and CTO of Backslash Security.

More trouble on the way for MCP servers

In a yet-to-be-released finding, Backslash said it also identified an exploit path involving a seemingly benign public document that can trigger a cascading compromise, because the MCP silently connected it into the LLM agent’s logic without proper boundaries.

The issue here wasn’t a vulnerability in the MCP code itself, but rather in the configuration of the data source it accessed. Backslash said the issue affects a 'very popular' tool with tens of thousands of users and that it's currently working with the vendor to coordinate responsible disclosure.

The company has now launched a free self-assessment tool for vibe coding environments to help security teams gain visibility into the vibe coding tools being used in their organizations, continuously gauging the risk posed by large language models (LLMs), MCP servers, and IDE AI rules in use.

"It's critical to give developers and vibe coders the tools and guidance to safely navigate this emerging attack service, which is why we’ve created the MCP Server Security Hub," said Pik.

"Developers will continue to tap MCP servers' flexibility and utility, so we wanted to give the community a safer means of doing so."

MORE FROM ITPRO

• The NCSC wants developers to get serious on software security
• Shifting left might improve software security, but developers are becoming overwhelmed
• Software security debt is spiraling out of control

Coding tools quickly emerged as one of the key use cases in the generative AI boom, with big tech providers promising marked productivity boosts and reduced workloads.

With an AI companion at their side picking up the drudge work, it’s not hard to see why developers would be enthusiastic about them. Research from GitHub last year seemed to back this up.

Developers were flocking to AI coding tools and for good reason: they were unlocking significant benefits working with these tools, saving huge amounts of time in their weekly schedules, and speeding up development processes.

Initial concerns over the quality of AI-generated code were also dissipating, the study noted, with 90% of US-based respondents reporting an improvement in this regard when using AI tools.

All told, the technology heralded a new golden era in the software industry, at least if big tech was to be believed. One of huge productivity boosts and reduced manual toil – and it came at a critical time. Developers were overworked during the pandemic and reported surging levels of burnout and mental health issues.

The reality at this point is the complete opposite, and some might now look back to the pandemic-era churn as a fond memory.

The slow trickle of AI integration has now become a torrent and developers working on AI tools are now finding themselves rendered obsolete by the very platforms and solutions they’ve built.

Microsoft, for example, announced fresh layoffs in May – its second batch in a matter of months. While tech industry workers are no stranger to layoffs at this stage, this latest round hit differently.

Reports from Bloomberg showed a significant portion of the workers set to be cut were operating out of the tech giant’s Redmond headquarters. Moreover, around 40% of these were in software engineering, equivalent to more than 800 roles.

For a company that builds software to begin cutting developers and engineers seems counterintuitive. But this is the new normal those in the profession face. Companies feel the need to justify their huge investment in AI by integrating it across the board, making humans surplus to requirements.

Microsoft CEO Satya Nadella recently revealed that around 30% of the company’s code is now AI generated. Microsoft isn’t alone in this, either. Last year, Sundar Pichai said more than 25% of Google’s internal source code was also AI generated – that’s likely grown since then.

That’s not to say that Google or Microsoft have AI running large parts of the show behind the scenes. Humans are still ‘in the loop’ in some capacity and this code is subject to rigorous scrutiny.

But what this does suggest is that the days of expansive software engineering divisions might be coming to an end, with those left after cuts reduced to caretakers monitoring the output of AI tools designed to usurp them.

The signs have been there all along

It’s not like there haven’t been warning signs for developers in recent years, or for workers spanning a range of professions. Human resources was in the crosshairs from the get-go during the early days of the generative AI boom, with IBM specifically highlighting this as an area ripe for automation.

The messaging from some big tech companies has been telling as well. Last year I discussed the fact that companies were quite obviously seeking ways to reduce headcount to compensate for AI investment – just without actually saying it.

These efforts to cut workers were veiled under the guise of “focusing on high-growth areas”. That same messaging and approach seems to be applied to specific professions, particularly software development.

The moment you hear terms like ‘drudge work’ when referring to core aspects of your role, you should probably start worrying. The same can be said when big tech providers evangelize about your ability to focus on the more “rewarding” aspects of your job, whatever they may be.

While attending Salesforce’s Dreamforce conference in San Francisco last year, this was a common recurring term bandied around by figures at the company – and it conveniently came at the start of Salesforce’s big agentic AI push.

In the months since, we’ve seen comments from CEO Marc Benioff specifically highlighting the potential for AI tools, particularly their own agentic AI solutions, to take over from human workers.

During an appearance on 'The Logan Bartlett Show’ earlier this year, Benioff suggested the company might not need to hire software engineers as a result of AI agents.

His comments on this topic echoed those made by Meta CEO Mark Zuckerberg, who told Joe Rogan that the company could also begin replacing engineers with AI within the next year or so.

It’s not out of the ordinary for industry figures to make bold claims, after all it justifies their strategy and investment in the technology. But the sheer volume of comments like these – and their consistency – should have alarm bells ringing.

Where do developers go now?

There have been small glimmers of hope on the horizon for embattled developers and engineers - but these largely focus on the potential for upskilling.

Research from Gartner last year showed that while AI will have a significant impact on the workforce, those best prepared to ride out the wave of automation will be the ones who adapt and upskill.

Critically, however, the consultancy said 80% of the workforce will need to upskill by 2027 to contend with AI-related skills requirements.

With such a large portion of the workforce expected to shake things up with regard to skills, this seems completely unrealistic. We already have casualties recorded at Microsoft, and there will undoubtedly be more.

Anthropic has launched its next generation of Claude models, and one in particular stands out as a developer’s dream.

Claude Opus 4, unveiled alongside Claude Sonnet 4, sets “new standards for coding”, according to the startup, and marks its most powerful model launch yet.

Opus 4 is designed specifically for software developers and engineers and “excels at coding and complex problem-solving” tasks.

The company claims the model achieved a 72.5% score on SWE-bench, which is used to benchmark software engineering tasks. Notably, this means the model dramatically outperforms OpenAI’s GPT-4.1, which scored 54.6% on the same testing.

When OpenAI announced the launch of GPT-4.1, it targeted the model at software developers - an area which the company has focused on in recent months.

Admittedly, this model did mark a sizable improvement compared to GPT-4o, which scored 21.4% when it launched. Regardless, Anthropic appears to have once again blown it out the water.

“These models advance our customers' AI strategies across the board: Opus 4 pushes boundaries in coding, research, writing, and scientific discovery, while Sonnet 4 brings frontier performance to everyday use cases as an instant upgrade from Sonnet 3.7,” the firm said in a blog post announcing the launch.

Elsewhere, Claude Opus 4 significantly outperforms previous models on memory capabilities, according to Anthropic. The firm said that when developers building applications provide Claude access to local files, the model “becomes skilled” at creating and maintaining memory files.

This allows the model to store key information more efficiently, thereby improving long-term task awareness, performance, and coherence.

Claude Opus 4 has serious stamina

According to Anthropic, the Opus 4 model boasts a combination of performance and stamina, so to speak.

During testing at Rakuten, the model contended with a “demanding” open source refactoring exercise while running independently for seven hours. Notably, Anthropic said it did so with “sustained performance”.

This marks a step change in performance and longevity, and suggests developers harnessing Opus 4 can do so across the length of their working day, rather than in sporadic bursts of activity.

“Cognition notes Opus 4 excels at solving complex challenges that other models can't, successfully handling critical actions that previous models have missed,” Anthropic said in its blog post.

How to access the new models

Anthropic described Opus 4 and Sonnet 4 as “hybrid models” which offer two distinct modes. These include “near-instant responses and extended thinking for deeper reasoning”.

The Pro, Max, Team, and Enterprise Claude plans will include both models and extended thinking capabilities. Notably, Sonnet 4 will also be available for free users.

Pricing for Opus 4 starts at $15/$75 per million tokens (input/output) and Sonnet 4 at $3/$15.

Both will be made available via the Anthropic API, or through Amazon Bedrock and Google Cloud’s Vertex AI service.

MORE FROM ITPRO

• Anthropic ramps up European expansion with fresh hiring spree
• There's no single route to AI adoption, and Anthropic is charting all available paths
• Databricks and Anthropic are teaming up on agentic AI development

Java, one of the world’s most popular programming languages, has officially turned 30 - but it’s still a long way from reaching its twilight years.

The programming language was first released in May 1995, and its launch came at a pivotal time in the development of the world wide web. The programming language was perfectly suited to cater for growing interactivity and multimedia requirements.

Java was created by James Gosling while working at Sun Microsystems, which was acquired by Oracle in 2010. Java development began some years before as part of a project - originally called ‘Oak’ - aimed at revolutionizing interactive television.

Ultimately, it proved slightly ahead of its time and the team behind the project pivoted to develop and fully flesh out the programming language. But with the rise of the web and more complex networking requirements, Gosling and co were wary of the potential pitfalls less portable languages could create in years ahead.

It was a solid call on their part. In an interview with ITPro way back in 2007, Gosling noted that “most of the stuff we talked about back then has happened, even though at the time it seemed like extreme science fiction”.

So what makes Java tick, and what’s the secret sauce behind its longevity?

Java’s timeless appeal

Java is a high-level, general purpose, object-oriented programming (OOP) language and was designed with a C/C++ style syntax that developers at the time would have found very familiar.

The language itself is very reliable and known for its ease of use, which also contributed to its popularity upon release. It was easy for developers to jump on the bandwagon and get going - and they sure did.

From the release of its first public version (Java 1.0) in 1996 to 2001, more than two and a half million developers globally were using the language in their daily workflows.

Today, Java still powers many of the popular websites and applications we use in our daily lives and remains the go-to language for many Android developers.

Indeed, figures from JetBrains' 2024 State of the Developer Ecosystem report showed 30% of developers cited Java as their primary language. Similar research last year also showed the programming language was experiencing somewhat of a renaissance, with enterprises planning to ramp up Java developer recruitment.

The Java Developer Productivity Report from Perforce showed investment was rising significantly, with nearly two-thirds (60%) of respondents planning to bring on more devs specializing in the language.

Java’s popularity among developers has remained modest over its long life, but admittedly peaked in the early noughties. Regardless, recent statistics from the Tiobe Index ranked the programming language in fourth place behind Python - which itself is experiencing a surge in popularity - C++, and C.

Despite its age, Dewan Ahmed, principal developer advocate at Harness, said Java still plays a critical role underpinning the tech stacks of enterprises globally.

“Java may be turning 30, but it’s far from showing its age – as it continues to underpin much of modern software development,” he said.

“Java’s longevity lies in a rare balance: consistent backward compatibility paired with continuous evolution. The Java Virtual Machine’s “write once, run anywhere” principle still resonates, while the syntax has evolved over the years to become more concise, without sacrificing its stability.”

The longevity of the programming language is also partly down to enterprise frameworks such as Spring Boot and Jakarta EE, Ahmed noted. These have ensured the programming language remains relevant amidst an age of cloud-native and microservices architectures.

“This vast ecosystem of libraries and tools, combined with widespread enterprise adoption, gives Java staying power that newer languages struggle to match,” he said.

Java in the age of memory safety

Another key appeal for Java at the ripe old age of 30 is the fact it remains a memory-safe language. Security agencies on both sides of the Atlantic have issued repeated warnings to enterprises about the dangers of memory-unsafe programming languages in recent years.

Languages such as C and C++ all suffer from one key weakness – they’re vulnerable to bugs that threat actors can pounce on. Some of these vulnerabilities include - but aren’t limited to - buffer overflow weaknesses or out of bounds read/write flaws.

These underlying memory-related weaknesses can be manipulated by threat actors to devastating effect. Analysis from Google in 2023, for example, estimated that three-quarters of CVEs used in zero-day exploits targeted memory safety flaws.

What next for Java?

According to Ahmed, Java’s long-term relevance will “hinge on how effectively it adapts to the asks of AI, cloud, and edge computing”.

There are ongoing initiatives which aim to meet these demands, however, such as Project Panama for example, which he said enables AI workloads to run more efficiently on hardware resources.

“In the cloud, ahead-of-time compilation via GraalVM is poised to become more mainstream, while advanced garbage collectors like ZGC and Shenandoah promise to minimize pause times for latency-sensitive applications like real-time analytics,” Ahmed added.

“These capabilities will ensure Java remains performant in modern, distributed environments and suggest it will remain so long into the AI-centric future.”

MORE FROM ITPRO

• The ultimate guide to becoming a programmer
• Interest in traditional programming languages is declining
• AI coding tools are finally delivering results

Python’s popularity continues to surge, according to the Tiobe index, with the programming language reaching a 25.25% share in May.

Figures from the index show Python now holds the highest rating for any programming language since 2001, surpassing Java’s record-breaking rating in June 2001.

But as Tiobe CEO Paul Jansen noted, this was the first month of the Tiobe index and there were far fewer languages being tracked.

Nonetheless, the programming language has surged in popularity in recent years, and in the last month alone increased by around 2.2%, with Tiobe figures from April showing the language boasted a rating of 23.08%.

Notably, May 2025 marked another record-breaking month for Python, according to Jansen. It’s the first time a programming language has held this big a lead over other languages ranked in the index. Python currently holds a 15% lead compared to the second-placed language, C++.

So what is it about Python that makes it such a popular choice for devs? There are a number of factors here, but one key appeal is its simplicity, experts told ITPro.

Python’s ease of use is a major draw

According to Allon Mureinik, senior software engineering manager at Black Duck, ease of use is a major draw, with the programming language having a “very low barrier of entry”.

“While you can technically write code in any language with no more than a text editor, Python is one of the few programming languages that actually make it practical”, he told ITPro.

“All you need to start coding in Python is the Python binary itself and a text editor, and you’re good to go.”

Mureinik added that this ease of use is a “huge factor” even for seasoned developers while the ecosystem surrounding the language is vibrant and still growing at pace.

“The syntax is simple without being simplistic, and the language constructs themselves make a lot of the developers’ routine tasks (such as manipulating or transforming data structures) very easy,” he said.

“Python has a rich ecosystem that offers libraries to integrate and interact with a wide variety of services, protocols and environments without having to implement those integrations from scratch.”

Darren Meyer, Security Research Advocate at Checkmarx, echoed Mureinik on the ecosystem front, noting that the Python community “really participates actively in thinking about security and compliance, which is appreciated”.

“While parts of the ecosystem (like PyPI, the Python Package Index which makes it easy to adopt open source Python libraries) have had security issues in the past, the community and maintainers are very responsive,” he told ITPro.

Python is the language of choice for AI developers

With the advent of generative AI, it’s hardly surprising that Python has soared higher in popularity, experts told ITPro. Indeed, research from Snowflake last year showed it has now become the “programming language of choice” for AI developers.

Analysis by the company found Python use grew by 571% on its Snowpark platform, highlighting its growing appeal. Moreover, this marked a considerable increase compared to other popular languages, such as Java and Scala.

“Python skills will be increasingly essential to development teams as they venture into advanced AI,” the report said.

So what makes it ideal for AI development? According to Meyer, Python is an “excellent language for rapid prototyping”, meaning developers aren’t bogged down by cumbersome workflows.

Meanwhile, Mureinik told ITPro its simplicity and ease of use again offers marked benefits here - an aspect of the language that Snowflake said enables developers to “focus on solving AI problems rather than parsing abstract syntax”.

“Python is a very convenient language for integrations and for stitching together workflows,” he said. “This makes it the obvious choice for programs that take data from source A, use it in a prompt to an LLM (through a Python client library, of course), and then take the output from that prompt as input for system B.”

Python does have some limitations

While Python has a lot going for it, there are certain limitations - particularly with regard to speed. Jansen specifically highlighted this issue in the latest Tiobe index update, suggesting that Python’s low performance is the “only reason other languages still have a reason for existing”.

This lower performance is due to how Python operates. It relies on an ‘interpreter’, a program that processes and executes code from a script file, or in some cases interactively. This allows programmers to test code - and run said code - on a line by line basis.

Notably, Python uses a Global Interpreter Lock, or ‘GIL’. This is a mutex, or lock, which limits the throughput of code, Jeff Watkins, chief technology officer at CreateFuture, told ITPro.

“The interpreter and its “global interpreter lock” limit throughput to a single, slow thread, so heavy-duty computing needs to be offloaded to C/C++ libraries,” he explained.

Efforts have been made to increase interpreter performance and remove these single-threading limitations, Watkins added, but still leaves developers wanting when compared to the performance of languages such as C++.

“Python will likely never outperform C++, but it doesn’t need to, as it was never its original purpose,” he noted.

Other limitations of Python are due to its dynamic typing approach, which Watkins said “trades off any ‘compile time’ safety for ease of implementation” and therefore can result in errors.

“This looser approach to typing means that runtime errors can slip through the net unless you put in belt and braces testing and typing hints through annotations,” he explained.

“Outside of niche projects, it also doesn’t have a mobile or client front-end footprint,” Watkins added. “As the Web runs on HTML, JavaScript, CSS, and WebAssembly, there’s no native way of using Python in the browser. Mobile and client browsers are also payload and performance-sensitive which makes it a less reasonable choice.”

Watkins said these notable gaps in performance and versatility explain why Python has supplemented other ecosystems, echoing Jansen’s earlier comments.

“Python’s rise is predictable, sensible, and more of a realignment than a hostile takeover,” he said.

“It has become the default high-level tool of choice, leaving space for complementary languages where absolute speed, strict typing, or native client experiences are paramount.”

MORE FROM ITPRO

• "Java is here to stay": Popular programming language to remain on business hit lists
• Want developers to build secure software? You need to ditch these two programming languages
• Organizations shift away from Oracle Java as pricing changes bite

AWS has expanded support for languages in Amazon Q Developer, making it easier for developers to code in their first language.

Amazon Q Developer is an AI assistant created for the end-to-end software development lifecycle that can help answer questions, create documentation, run software upgrades, and even review code.

Reflecting the global nature of the coding workforce, AWS said that the platform now supports "additional human languages" beyond just English, including Arabic, French, German, Hindi, Italian, Japanese, Korean, Mandarin, Portuguese, Spanish.

The move follows the Amazon Q Business platform being made available in Europe for the first time last month.

"While English remains the lingua franca of programming, the reality of modern software development extends far beyond code," wrote Brian Beach, Principal Solutions Architect at AWS, in a blog post.

"This natural flow of conversation in any language helps maintain the developer’s focus and flow, eliminating the mental overhead of constant translation," Beach added.

The expanded language support is available immediately to users on the free tier, as well as the Pro Tier.

Amazon Q Developer changes deliver “more natural, fluid conversations”

Beach said developers around the world use Amazon Q Developer to discuss architecture decisions, create documentation, and design user interfaces — all for applications that serve audiences that are equally global.

"By expanding language support, Amazon Q Developer now enables developers to have more natural, fluid conversations about complex technical concepts in their preferred language, whether they’re designing system architecture, generating documentation, or planning application localization strategies," he said.

Beach shared an example where he asked how to host a container on AWS in English, Mandarin, Hindi and Spanish, with Amazon Q Developer returning a reply in the same language — and, importantly, getting the answer right.

"Not only does Amazon Q Developer now provide complete responses in these languages, but it also maintains technical accuracy while adapting to linguistic nuances," Beach noted.

Beyond answering questions, Amazon Q Developer can also suggest follow-up questions in the user's default language.

MORE FROM ITPRO

• AWS goes all in on AI agents with new features for Bedrock and Amazon Q
• Everything you need to know about Amazon Q, including features, pricing, and business tiers
• AWS sharpens sustainability focus as AI environmental concerns rise

With AI coding tools gaining significant traction in recent years, some tech leaders worried the technology spelled doom for low-code and no-code solutions – if anything it’s been quite the opposite.

In a survey from App Builder, 84% of tech leaders said they don’t believe the technology will replace these tools. In fact, adoption has been steadily rising in recent years and shows no sign of slowing down.

Findings from the company’s 2025 App Development Trends Report show low-code and no-code solutions now play a significant role at organizations globally, with 95% of firms having used these tools for software development in the last year.

These statistics align with previous research from Forrester, which reported in 2024 that 88% of developers were using low-code platforms.

This increased focus on low-code solutions has run in parallel with the rapid adoption of AI in their businesses, the study found, and tech leaders are confident that these will complement each other. More than three-quarters (76%) said AI will actually make existing tools more efficient.

Jason Beres, SVP of developer tools at Infragistics, said the study highlights that low-code and AI tools can not only co-exist, but the combination of the two will transform how applications are built.

“Companies’ use of low-code and no-code tools have steadily increased over the past decade–and today these solutions are critical to how developers, designers, CIOs and others build scalable applications and effective development processes,” he said.

“Now, as AI converges with low-code and no-code, we’ll see an entirely new paradigm emerge that will change how apps are built, who can build them and how fast it can happen.”

Low-code tools are still valued by developers

The study from App Builder underlined the continued popularity of low-code tools among enterprises. Nearly every tech leader surveyed said they’ve saved time and money using these tools, with 98% reporting development time savings in particular.

More than three-quarters (78%) of respondents said they’ve saved up to 50% of development time.

The benefits cited by users aren’t just centered around time saved, however. Nearly two-thirds (62%) of companies that use low-code tools have reduced overall development costs, with 67% unlocking cost reductions of up to 30%.

Use-cases for low-code practices are also evolving, the study noted. A leading factor behind the implementation of low-code tools at many enterprises was to improve developer productivity.

While this remains a key focus, there are other factors at play. These included enabling developers to focus on “more strategic work”, cited by 25% of users now compared to 22% initially.

Decreasing manual errors and increasing end-user satisfaction are also now key factors in the adoption of low-code tools.

“It’s clear low-code and no-code tools are here to stay,” Beres added.

“As companies incorporate AI and other new technologies into low-code and no-code development, tech leaders must not only leverage the benefits of automating more app development processes faster, improving collaboration, and increasing productivity, but ensure their organizations and development teams continue to evolve and upskill with their technology.”

MORE FROM ITPRO

• Can AI code generation really replace human developers?
• AI coding tools aren’t the solution to the unfolding 'developer crisis’
• Not all developers are sold on AI coding tools

Junior software developers may have become too reliant on AI tools and it’s undermining core coding skills, according to claims by tech blogger and programmer Namanyay Goel.

In a blog post that’s sparked intense discussion in the industry, Goel wrote that he’s been troubled by the behavior of new developers he’s spoken to who are using some sort of AI coding assistant.

“Every junior dev I talk to has Copilot or Claude or GPT running 24/7. They’re shipping code faster than ever. But when I dig deeper into their understanding of what they’re shipping? That’s where things get concerning,” Goel said.

Though the code works, new devs are unable to explain how or why it works, Goel noted, or are stumped on questions about edge cases. Foundational knowledge is missing as junior developers are not learning from scratch, he added.

“We’re trading deep understanding for quick fixes, and while it feels great in the moment, we’re going to pay for this later,” Goel said.

Anecdotally comparing today’s developers with those of his generation who had to use Stack Overflow when they encountered an error, Goel described how learning from experienced developers is far superior to learning through AI tools.

“AI gives you answers, but the knowledge you gain is shallow. With StackOverflow, you had to read multiple expert discussions to get the full picture. It was slower, but you came out understanding not just what worked, but why it worked,” Goel said.

Developers need to use AI tools correctly, expert suggests

Using AI tools can be a useful learning tool, according to Peter Schneider, Senior Product Manager at Qt Group, but they need to be used in the right way.

“If junior developers generate code with AI assistants and deploy the code to digital products without truly understanding it, then they run into the risk of introducing suboptimal code. It's also worth noting that whenever junior developers use AI-generated code, they're not really learning how to write and review code themselves,” Schneider told ITPro.

“That's why we recommend to our customers today that they use AI assistants to augment the learning experience of junior developers, not to replace coding itself,” he added.

There are clear benefits to using AI coding tools, but junior developers need to remember that AI isn’t flawless, said Faye Ellis, principal training architect at Pluralsight.

“It can still produce inaccurate outputs, and though large language models (LLMs) are remarkable, their current ability to learn and improve from interactions remains quite limited, meaning progress may feel uneven for some time,” Ellis told ITPro.

“Overreliance on AI tools could stop junior developers from developing essential skills, as generating code with AI is not the same as fully understanding it. Businesses should still invest in training developers to improve their coding expertise, rather than rely solely on AI,” she added.

Are AI coding tools worth the risk?

It’s not just workers' skill sets being undermined that developers need to worry about. Various reports over the last few months have put the quality of AI-generated code up for debate as well.

A report from Harness earlier this year found that 59% of developers reported problems with deployments in AI-generated code at least half of the time.

While 92% said AI tools had increased the volume of code shipped, they also said this increased the ‘blast radius’ of bad code.

67% of developers reported more time debugging AI-generated code and 68% spent more time fixing vulnerabilities after adopting AI tools, the report found.

Bogging developers down in these remediation tasks is also inflicting a serious financial toll on organizations and potentially negating the productivity gains afforded by the tools, Harness warned.

AI is causing a skills concern across the board

Software developers aren't the only professionals becoming over-reliant on AI tools, however, and this is a problem that could spread across a range of professions in years to come.

A recent study from Microsoft examining the impact of AI on knowledge workers found the increased use of the technology could be impacting their critical thinking skills.

The study, conducted in collaboration with researchers at Carnegie Mellon University revealed a deterioration in certain cognitive faculties among those who frequently used AI tools.

Researchers warned the use of the technology is making workers unprepared to deal with anything other than routine tasks in their daily workflow.

“While AI can improve efficiency, it may also reduce critical engagement, particularly in routine or lower-stakes tasks in which users simply rely on AI, raising concerns about long-term reliance and diminished independent problem-solving," researchers said.

MORE FROM ITPRO

• AI coding tools are finally delivering results for enterprises
• AI-generated code risks: What CISOs need to know
• Can AI code generation really replace human developers?

GitHub has launched a new agentic AI tool for developers, dubbed ‘Agent Mode’, alongside a raft of updates to the Copilot service.

Agent Mode enables Copilot to iterate on its own outputs and complete entire user requests in one go, thereby automating broader aspects of the development process. It can also automatically fix errors, suggest terminal commands, and assess run-time errors.

Copilot will now iterate repeatedly on its output until all the necessary subtasks of a prompt are complete, GitHub CEO Thomas Dohmke said, and will infer additional tasks that were not specified but are necessary to fulfill the primary request.

The new agentic capabilities will improve Copilot Workspace, allowing collaborative users to leverage a system of sub-agents in the platform to improve efficiency when generating plans, implementing code, or fixing errors.

“Agent mode will change the way developers work in their editor; and as such, we will bring it to all IDEs that Copilot supports,” Dohmke said.

Another feature introduced to Copilot is the ‘Next Edit’ function, which can automatically identify and suggest edits of a user’s previous changes. Accessed by pressing tab, users will be able to browse suggestions throughout any open file.

Vision for Copilot was also unveiled, which will allow users to create a mock-up piece of code from a snippet of content, such as a shorter portion of code, a screenshot, or an image. With this, Copilot can generate the UI, alt text, and code.

GitHub referred to its forward-looking plan, ‘Project Padawan,’ through which developers will be able to assign entire issues to Copilot, let it complete the task, and then evaluate and review the results.

The next step in AI code assistance

This announcement could mark yet another significant step in AI-based code assistance, which has been gaining huge popularity as a use case amongst developers and businesses. Google recently revealed that over 25% of its proprietary code is now AI-generated, for example.

GitHub has been at the center of the AI coding conversation for a while, having released GitHub Copilot Enterprise in early 2024, which it followed up with the release of GitHub Copilot Workspace a couple of months later.

GitHub touts the benefits of its tools, claiming that GitHub Autofix fixes software vulnerabilities at more than three times the speed of manual fixes, and at more than seven times the speed for cross-site scripting vulnerabilities.

There have been questions raised about the effectiveness of AI coding tools, though, with one study finding that developer performance has largely remained unchanged despite the use of coding tools.

GitHub’s own research paints a more positive picture, however. Based on a survey of 2,000 software engineers and developers, the firm found that many reported several benefits related to AI coding tools.

Many said the tools improved code quality, helped deliver more secure software, improved test case generation, and allowed for the faster adoption of new languages.

These tools also enable better time utilization, with 47% of respondents in the US and Germany using the time they saved for collaboration and system design.

MORE FROM ITPRO

• Can AI code generation really replace human developers?
• The world's 'first AI software engineer' isn't living up to expectations
• AWS CEO Matt Garman thinks AI coding tools could herald the end of the developer as we know it

With software developers increasingly flocking to AI tools to support daily activities, new research suggests a concerning portion are using unauthorized solutions.

Findings from Harness’ State of Software Delivery Report show that more than half (52%) of developers don’t use IT-approved tools, raising significant compliance and intellectual property concerns.

“Perhaps the most alarming observation was around the use of company-approved coding tools - of lack thereof,” the report states.

“The unauthorized adoption of AI codegen tools creates significant shadow IT challenges that extend far beyond immediate security concerns.”

Shadow AI is a serious cause for concern for teams, the report added, with developers potentially exposing sensitive code snippets to third-party AI services without proper governance.

“Ultimately, they can’t track the origin of AI-generated code, nor can they ensure consistent security standards across teams,” Harness said.

Software developers aren’t the only ones flocking to shadow AI

The rise of shadow AI has become a recurring talking point over the last two years as enterprises globally flock to the various AI tools available on the market.

In its Chasing Shadows report, Software AG found 75% of knowledge workers are already using AI, with this figure set to rise to 90% in the near future, and more than 50% of this group use personal or non-company issued tools when doing so.

Another study by customer service platform Zendesk noted there has been as much as a 250% rise year on year in the use of unsanctioned AI tools in certain industries.

The financial services sector was found to be the worst affected by this phenomenon with a 250% spike year on year compared to 2023 levels, but the healthcare (230%) and manufacturing industries (233%) also exhibited very high levels of shadow AI use.

Developing robust AI usage policies will be integral to ensuring this growing reliance on unvetted AI tools does not expose your enterprise to unnecessary risks.

Harness’ report listed the critical gaps identified by software engineering leaders in their organization’s AI coding tool policies.

Three-fifths of engineering leaders said companies need policies prescribing the processes for assessing code for vulnerabilities or errors, with 58% stating they need to outline specific use cases where AI is safe or unsafe.

Bharat Mistry, field CTO at Trend Micro, told ITPro implementing the policies included in the Harness report were all wise, but highlighted the importance of training when trying to shape employee behaviour and foster responsible use of personal AI systems.

“I agree with the policies given above, however for me it begins with the human aspect. By investing in comprehensive training and awareness programs, businesses can empower their employees to use AI responsibly, identify and mitigate risks and contribute to the development of ethical and effective AI solutions,” he argued.

“This proactive approach not only enhances the organization’s AI capabilities but also builds a culture of trust and accountability around AI technologies.”

Speaking to ITPro, Steve Ponting, director of Software AG echoed Mistry’s comments, noting that training will be essential in mitigating the risks associated with employees using their own AI tools.

“Workers have been clear: they will use AI whether it’s sanctioned or not. This means that businesses could struggle to manage AI applications, leading to cyber-security risks, skills gaps, and inaccurate work.,” he explained.

“Businesses must have a plan in place to reduce risk, build skills and plan for AI’s inclusion in daily work. If people are determined to use their own AI, training is vital in this regard. Better training would make 46% of employees use AI more, but crucially, they would use it effectively and responsibly.”

Python has overtaken JavaScript to become the most popular programming language on GitHub, according to new figures released by the firm, largely thanks to the rise of machine learning and generative AI.

The coding platform revealed that there are now 518 million projects on GitHub, up by a quarter year-on-year, with 5.2 billion contributions to projects over the course of 2024.

AI is having a major impact, the firm noted. There's now 137,000 public generative AI projects, almost doubling since last year, and more than a million open source maintainers, students, and teachers have used the free AI-powered GitHub Copilot tool.

With Python taking the top spot on GitHub, its ascendancy brings to an end ten years of JavaScript holding the top spot. Moving down to second, the programming language is followed by TypeScript and Java.

It’s not all bad news for the language, however. JavaScript is still the top ranked for code pushes — which means more developers use it more often to push code — but across all activities on the site, including commits, issues, pull requests and more, Python is now the top language.

"More developers still use JavaScript more often to push code, but in absolute activity across all contribution types on GitHub, Python now outranks JavaScript," GitHub said in a blog post.

Python's meteoric rise to prominence

GitHub identified two key factors behind Python's popularity on the platform - namely the rise of data science and generative AI, as well as TypeScript cutting into JavaScript's lead.

The firm said TypeScript had grown exponentially from 2014 to 2019, overtaking Java last year to join the top three programming languages.

"While Python is increasing in contributor counts for both code push activity alone and other activity faster than JavaScript, it isn’t increasing in those faster than JavaScript and TypeScript combined," the post noted.

"Rather than a slow down in the JavaScript community, what we are seeing is a transition to TypeScript for a large proportion of new commits."

It's not just Python that’s benefitted from the rise of machine learning either. Jupyter Notebooks has seen a spike in usage of 92% over the last year and 170% since 2022.

"This could indicate people in data science, AI, machine learning, and academia increasingly use GitHub," the post added.

GitHub's ranking echoes recent research by cloud data company Snowflake, which in March revealed Python had become the go-to programming language for developers building generative AI applications, pointing to 571% growth on its own platforms.

But Stack Overflow's 2023 survey ranked Python third overall, which it said was a reflection of who uses the language. It’s the fourth most popular programming language among professional developers, but more popular among those learning to code and the top ranked for those who aren't full time developers but use some coding at work.

That tracks with GitHub's assessment that the rise of AI and data science was drawing a wider range of people into coding.

"Python is used heavily across machine learning, data science, scientific computing, hobbyist, and home automation fields among others," GitHub added.

"The rise in Python usage correlates with large communities of people joining the open source community from across the STEM world rather than the traditional community of software developers.

"The notion of who a developer is and the scope of what a developer does is changing. The rise in Python, HCL, and Jupyter Notebooks, among other things, indicates that the notion of a developer extends beyond software developers to roles like operations or IT developers, machine learning researchers, and data scientists."

Other shifts driven by AI

AI is helping to encourage coding globally by making it easier for people from a wider range of countries, GitHub said.

"The rise of these non-English, high-population regions is notable given that it is happening at the same time as the proliferation of generative AI tools, which are increasingly enabling developers to engage with code in their natural language," the post said.

And that's good news for the industry. By attracting a more diverse community of developers, this will ultimately drive innovation long-term, the firm noted.

"Increased access to and experimentation with AI could also simplify and personalize the coding journey for new developers, lowering entry barriers and further diversifying GitHub’s community of developers."

More widely, generative AI is becoming a core part of software development at every stage.

"They power coding tools that offer fixes and context behind vulnerability remediations, suggestions in response to natural language prompts and existing code, and facilitate learning among new and experienced developers alike," the post added.

"They’re also changing how developers build applications, and developers will benefit from platforms that allow them to easily experiment with AI models as building blocks without requiring separate setups or extra costs."

Previous research from GitHub found that developers are beginning to unlock real value from AI coding tools, with respondents to a recent survey relieving the solutions have helped them learn new languages and save time.

Scientific publishing has traditionally been a flat, static affair. You write up your research in a document, which is typically a PDF with text and code. Someone reads it and might use aspects of your research in their own projects. But what if you could make those documents smarter? Wouldn't it be useful to have a data science paper or programming tutorial document that you could actually interact with? That's what JupyterLab, a free tool from Project Jupyter, aims to provide.

Traditionally, programmers wrote Python programs in a text-based file with a .py extension. If you wanted to explain what the source code in the document was doing, then you'd write comment blocks to give others an idea of what you were thinking when you wrote the program. That kind of comment is fine for code that people rarely look at, such as one working behind the scenes running database interactions on a server.

JupyterLab: Who is it for?

If you're a computer science tutor trying to explain how bubble sorts work, or a biology morphologist describing how you used code to map the Fibonacci sequence in the disk of a sunflower, mixing code that runs together with text and images gives you a much more intuitive way to tell your scientific story. Computer scientist Donald Knuth envisaged this in 1984 in a concept called literate programming. A literate program file runs like a non-literate program file, but it also structures and describes that code in a way that makes more sense to the author and their audience.

That's where Project Jupyter comes in. It evolved from an initiative called IPython (interactive Python), formed in 2011, which used a different kind of format than .py files. An iPython notebook file is human readable, containing text written in the markdown format along with Python code that you could run. Each snippet of code and text description occupies its own cell, and produces a cell containing its output when you run it, like this:

You could send that notebook as a single document for anyone running IPython Notebook to run and interact with, or you could send them the URL to an online version.

As IPython Notebook grew more popular, the project spun them off into a separate initiative called Jupyter. Jupyter develops the tools to run the notebooks while IPython focuses on the underlying interactive Python implementation to support them.

Jupyter carried on developing the notebook project, now called Jupyter Notebook, but it saw more data scientists using this tool for complex data analysis projects. To support them, it developed a product with more functionality called JupyterLab. This summer (2023), the team released the fourth version of that project.

JupyterLab: Setup

JupyterLab is a fully integrated development environment (IDE) targeted mainly at data scientists, although anyone wanting to exchange descriptions of their code journeys can use it. Like Notebook, it runs as a server that you visit in your browser. You can either access it online using someone else's server (Jupyter offers its own demo here), or you can install it locally.

We've found the best way to do this is via the Conda package manager, which contains pre-baked collections of packages including JupyterLab. You can install the collection in a couple of lines as a separate virtual environment in Conda, and then just switch to it whenever you like. This allows you to maintain a version of Python separately from your system version, with separate packages.

JupyterLab doesn't just support Python, although that's the world that it came from. Like Jupyter Notebook, it supports a wide range of languages. You can simply select these from its menu, as long as you have installed the relevant kernels, and it handles syntax-based formatting for you.

Whereas Jupyter Notebook comes with a basic file browser that you use to select a single file, JupyterLab starts with two panes: one displaying the file browser and the other displaying a launcher. This enables you to quickly and graphically create different kinds of files. Notice that you can also create a console directly in the browser – something that Jupyter Notebook doesn't support – if you want to tinker with code before putting it into your main notebook.

JupyterLab: File handling

This pane-based approach is an example of the fundamental differences in the user interface that set JupyterLab apart from Jupyter Notebook. Another nice feature is the way that the file browser opens documents of different types. We downloaded a CSV with the results of public school inspections from the UK government's open data repository. We first opened it in Libre Office and resaved it with UTF-8 encoding, which Jupyter needs to open a CSV directly. Opening it as a file from within Jupyter Notebook (on the left), and from within JupyterLab (on the right): reveals quite a difference:

Jupyter Notebook displays this as a text file, while JupyterLab displays it as a properly formatted, scrollable CSV file. It also displays other kinds of documents like PDFs in their native format, making it an excellent file browser and display mechanism for those working on large projects.

JupyterLab: Features

Notice how the JupyterLab interface on the right is tabbed? This demonstrates another big advantage of JupyterLab: multiple document management, which is also a significant enhancement for those working on large projects. Jupyter Notebook focuses just on single documents, meaning that if you want to work on more than one you have to open another tab in your browser.

JupyterLab's tabbed interface is useful enough as shown, but you'd still have to click on each tab to see the appropriate document. What if you wanted to compare two documents side by side? Simply click on a tab and then select New View for Notebook. This creates a split window that lets you see the document side by side. Here, We've created a document to load the school's csv file into a Python Pandas data frame and then displayed the data frame in the notebook. Splitting the view lets me see the CSV and the Python representation of it side by side:

JupyterLab also has plenty of enhancements for working inside specific notebooks. One of these is the ability to collapse cells. When we displayed the school's data frame in the notebook, it created a lot of text. We can collapse that by clicking on the cell's sidebar on the left, making any subsequent code more visible:

Changing the code in a cell is simple – you just click on the cell and edit it. However, you have to rerun it to produce its updated output using CTRL-RET. Jupyter Notebook didn't indicate that the adjusted code hadn't run yet unless you happened to notice its out-of-date output. JupyterLab fixes that by using different colors for the sidebar. A cell that hasn't been updated since it last ran is blue. If a cell has not run to reflect new edits, the sidebar turns orange:

Let's do a little more analysis on schools, grouping schools by region and checking the mean average of their education quality. Then we can plot that as a horizontal bar graph:

That's a long bar chart. Let's say we want to keep it handy for visual reference as we continue working on our notebook. We could create another view of this notebook, rearranging it to the right-hand side of the screen, and then collapse the graph view in our original notebook to make room for our code. However, that creates a problem, because collapsing the graph in one doc also collapses it in the other.

Instead, we can do another thing that isn't possible in Jupyter Notebook: create a new view of a single cell's output. This creates a new tab with just the output of the cell that we can then drag to the right, rearranging it to keep it handy at all times. Collapsing the output cell in the original notebook has no effect in the new output view:

To create a true IDE, the Jupyter team gave JupyterLab another feature that doesn't exist in Jupyter Notebook: a debugger. If your Python kernel supports debugging, you'll see a small bug icon next to the kernel indicator in the top right of your window. Clicking that turns it orange, switching on the debugger, and showing us its data in a sidebar. The debugger supports the basic features you'd expect, including variable watching and setting breakpoints. We can set breakpoints by clicking on the appropriate line in the gutter, creating a small orange indicator:

Version 4 of JupterLab, released in June, contains some welcome new features. The team has adjusted its code behind the scenes and updated the text editor that it uses to make editing faster, especially for large notebooks. You can also now import extensions for JupyterLab directly from PyPi in addition to installing directly using the tool's extension manager. JupyterLab was built with extensibility in mind, and there are plenty of plugins to enhance your experience.

JupyterLab: Is it worth it?

JupyterLab is the perfect tool for creating not just programs, but stories of how those programs came to be and what you were thinking when you made them. Think of it like a smart lab journal - a kind of Harry Potteresque document that's alive and answers your questions as you read it. It's such a good system that in many cases, especially when teaching programming to high-school students, say, it's a worthy competitor to many other IDEs. Whenever we're looking for insights into data or simply trying out some concepts in Python programming, this will be our first port of call.

The Raspberry Pi 5, the latest iteration of the highly popular microcomputer, is officially set to launch in October with its most powerful components ever.

In an announcement, Raspberry Pi’s designers revealed that the new model is more than two times more powerful than its predecessor, offering users a much faster device with vastly superior performance. 

Eben Upton, CEO at Raspberry Pi, said “virtually every aspect” of the new model has been upgraded ahead of the launch and will offer a “no-compromises user experience”. 

In addition, the new device will be the first of its kind to feature chips designed in-house by the company. 

“Today, we’re delighted to announce the launch of Raspberry Pi 5, coming at the end of October,” he said. 

“Raspberry Pi 5 comes with new features, it’s over twice as fast as its predecessor, and it’s the first Raspberry Pi computer to feature silicon designed in-house here in Cambridge.”

Raspberry Pi 5: Performance and features

The Raspberry Pi 5 will come equipped with a 2.4GHz quad-core, 64-bit Arm Cortex A-76 CPU, which marks a significant boost compared to the previous iteration. 

The Raspberry Pi 4 initially came armed with a Broadcom BCM2711 SoC with 1.5 GHz capabilities. Later versions, however, offered up to 1.8 GHz.  

With an 800MHz VideoCore VII GPU, the latest version offers far greater graphics capabilities compared to its predecessor, the company said. 

The device also boasts dual HDMI ports capable of 4K display output up to 60 frames per second, complete with support for HDR. 

USB bandwidth has received a significant boost in the new Raspberry Pi, featuring two USB 3.0 ports to support faster transfer speeds, as well as two USB 2.0 ports and gigabit ethernet ports. 

Bluetooth 5.0 support and Bluetooth Low Energy (LE) capabilities have also been added.

A complete list of features outlined by Raspberry Pi can be found below:

• 2.4GHz quad-core 64-bit Arm Cortex-A76 CPU
• VideoCore VII GPU, supporting OpenGL ES 3.1, Vulkan 1.2
• Dual 4Kp60 HDMI® display output
• 4Kp60 HEVC decoder
• Dual-band 802.11ac Wi-Fi®
• Bluetooth 5.0 / Bluetooth Low Energy (BLE)
• High-speed microSD card interface with SDR104 mode support
• 2 × USB 3.0 ports, supporting simultaneous 5Gbps operation
• 2 × USB 2.0 ports
• Gigabit Ethernet, with PoE+ support (this requires separate PoE+ HAT)
• 2 × 4-lane MIPI camera/display transceivers
• PCIe 2.0 x1 interface for fast peripherals
• Raspberry Pi standard 40-pin GPIO header
• Real-time clock
• Power button

Raspberry Pi 5: Pricing, release date, and shipping

While an exact release date hasn’t yet been confirmed, the company said the new device will be coming “at the end of October”. 

The new models will also be “ring-fenced”, as Raspberry Pi put it, meaning they will be restricted to single-unit sales. This, Upton said, will ensure users “get the first bite of the cherry”.  

Raspberry Pi will also likely be looking to avoid a repeat of the stock issues it suffered with the launch of Raspberry Pi 4. Supply chain problems caused a bottleneck in production that resulted in lengthy shortages. 

At the peak of this shortage, the company faced a backlog of nearly 2,000 devices, most of which were its 8GB version. The issue prompted a wave of scalping online which further compounded the issue. 

Pricing for the new model is set at $60 for the 4GB variant, the company revealed, while the 8GB version will cost $80. 

This does represent a marginal price increase compared to the previous iteration, which started at $55 for the 4GB version and $75 for the 8GB version.

Raspberry Pi 5: Operating system and accessories

To use the Raspberry Pi 5, customers will be reliant on the latest version of Raspberry Pi OS, known as ‘Bookform’. 

Similarly, in terms of accessories, the new version will include a new case. It won’t, however, be compatible with the old Raspberry Pi 4, meaning customers will have to purchase a new one, which the company said is “designed to help you get the most” out of the latest device. 

This new case is priced at $10 and offers a “host of new usability and thermal-management features”, including an integrated 2.79 (max) CFM. 

The integrated fan uses fluid dynamic bearings which the company said were specifically selected for “low noise and an extended operating lifetime”. 

“We’ve lengthened the case, and tweaked the retention features, to make it possible to insert the Raspberry Pi 5 board without removing the SD card,” the company said. 

“And by removing the top of the case, it is now possible to stack multiple cases, as well as to mount HATs on top of the fan, using spacers and GPIO header extensions.”

ChatGPT has been found to produce incorrect answers for over half of the software engineering questions posed, according to fresh research.

The Purdue University study saw researchers analyze ChatGPT answers to 517 Stack Overflow questions, with the aim of assessing the “correctness, consistency, comprehensiveness, and conciseness” of answers presented by the generative AI tool. 

Researchers reported 52% of answers to programming-related queries were inaccurate, while more than three-quarters (77%) were deemed “verbose”. 

A key talking point from the study centered around user interpretation of answers presented by ChatGPT, as well as the perceived legitimacy of answers produced by the chatbot. 

Researchers said ChatGPT’s answers are “still preferred [to Stack Overflow] 39.34% of the time due to their comprehensiveness and well-articulated language style”, resulting in users taking answers at face value. 

“When a participant failed to correctly identify the incorrect answer, we asked them what could be the contributing factors,” researchers said. “Seven out of 12 participants mentioned the logical and insightful explanations, and comprehensive and easy-to-read solutions generated by ChatGPT made them believe it to be correct.”

Of the “preferred answers” identified by users to software queries, more than three-quarters (77%) of these were found to be wrong. 

Researchers said that users were only able to identify errors in ChatGPT-based answers when it was glaringly obvious. However, in instances where the error was “not readily verifiable”, users frequently failed to identify incorrect answers or “underestimate the degree” of error in the answer itself. 

Surprisingly, the study also found that even when answers contained obvious errors, two out of 12 still marked them as correct and revealed they “preferred that answer”. 

Researchers said the perceived legitimacy of answers presented by ChatGPT in this instance should be a cause for concern among users. “Communication correctness” should be a key focus for the creators of such tools. 

ChatGPT does give users ample warning the answers provided may not be entirely accurate, stating the chatbot “may produce inaccurate information about people, places, or facts”. 

But the study suggested “such a generic warning is insufficient” and recommended answers are complemented with a disclaimer outlining the “level of incorrectness and uncertainty”. 

“Previous studies show that LLM knows when it is lying, but does LLM know when it is speculating? And how can we communicate the level of speculation?” the study pondered. “Therefore, it’s imperative to investigate how to communicate the level of incorrectness of the answers.”

The use of generative AI tools in software development and programming has gathered significant pace in recent years, most notably with the launch of GitHub’s Copilot services. 

Earlier this year, the firm announced the general availability of the AI-based coding assistant for business customers. The model is designed specifically to bolster code safety and has been hailed by developers as a vital tool in supporting their daily operations. 

A survey published by GitHub in June revealed that a majority (92%) of devs now use an AI coding tool at their work, with 70% stating they see “significant benefits” to using generative AI tools in workplace settings. 

GitHub has revealed a new feature for its Copilot pair programmer that will inform developers when code suggested to them by AI is taken from public repositories.

It said the reasoning behind the move is to offer developers greater transparency when it comes to using Copilot’s code suggestions. 

“Some want to learn from others’ work, others may want to take a dependency rather than introduce new app logic, and still others want to give or receive credit for similar work. Whatever the reason, it’s nice to know when similar code is out there,” it said.

Currently in private beta, Copilot’s new ‘code referencing’ feature will check code suggestions - including surrounding code of approximately 150 characters - against all public code on the GitHub platform. Any matches, along with repository information, will be shown to the developer.

It is then up to the developer to accept the code suggestion - including information about matches - or block the suggestion. 

GitHub estimates that matches will occur in less than 1% of Copilot suggestions, although that figure depends on context. 

“In the context of an existing application with surrounding code, we almost never see a match,” it said. “But in an empty or nearly empty file, we see matches far more often.”

The code referencing solution could make ground on easing the concerns shared among developers and organizations regarding the potential legal issues that could arise from using suggestions from Copilot.

The release will be seen as an acceptance from GitHub that developers wish to know when the suggestions match public code. 

It comes amid scrutiny of the company over the training of Copilot on publicly available code, including public GitHub repositories.

GitHub continues to wrestle with a class action lawsuit it currently faces, one that was filed in the San Francisco US federal court last year. 

The complainants’ case centers around allegations that the company committed “software piracy on an unprecedented scale” by allowing Copilot to suggest code snippets from publicly available open source code repositories.

The case noted that open source projects are protected by a number of different licenses which require developers to correctly attribute the project’s authors’ names and copyright when using its code in other applications.

GitHub faces allegations that by suggesting code snippets through Copilot, it is removing the licenses that protect the open source projects from which the code was suggested, illegally violating their terms of reuse.

GitHub and Microsoft, which owns the platform, have filed multiple motions to dismiss the case, including its most recent effort in June 2023. The former was rejected and the latter will be heard in court in September.

Why do developers need code referencing?

GitHub Copilot was launched in 2021 and is designed to assist developers by suggesting blocks of code based on context. 

GitHub describes the tool as the world’s first at-scale AI pair programmer and, according to the company, has trained it on billions of lines of public code.

Recent figures from GitHub suggest it has been used by more than 1 million developers and more than 27,000 organizations to build code faster.

However, GitHub has acknowledged that developers would also like to know when suggestions made by Copilot match public code. 

It has been possible to block matching code since 2022. However, in November last year, Ryan Salva, vice president of product at GitHub, said: “While useful in some contexts, blocking matching suggestions doesn’t address all use cases”.

With code referencing, the developer is shown the matching code, the repos where the code appears, and the license governing each repo. 

Since the same code might appear in multiple places with occasionally conflicting licenses, developers can decide for themselves on attribution rather than having the matches blocked from the outset.

GitHub has announced a public beta release of its Copilot Chat solution for all business users on Visual Studio and VS Code, marking the company’s first step in taking its wider Copilot X offering to the enterprise.

Announced back in March, Copilot X is the latest iteration of GitHub’s original Copilot AI pair programming assistant with features that leverage the GPT-4 model from OpenAI. 

GitHub said this latest evolution turns Copilot into a context-aware conversational assistant within the integrated development environment (IDE), which allows developers to execute complex tasks using simple prompts. 

It’s described as a “Chat-GPT-like experience” from within Visual Studio and VS Code, enabling efficient analysis of code blocks, as well as the provision of bug fixes.

As the centerpiece of Copilot X, GitHub said Copilot Chat will “swing the doors wide open” to an age where natural language powers the coding experience.

“Every developer on your team, from the least to the most experienced, will be able to build entire applications or debug vast arrays of code in a matter of minutes instead of a matter of days,” said Mario Rodriquez, vice president of product at GitHub. 

“Weeks or months spent slogging over unit tests and endless backlogs of boilerplate code will be a thing of the past.”

Once Copilot X becomes generally available, GitHub predicts every developer could be made ten times more productive, enabling a new generation of developers to be able to learn and build at speed.

The company highlighted data from Stack Overflow’s most recent developer survey, which revealed that developers often spend 30 minutes and sometimes up to two hours a day searching for solutions to problems.

Copilot Chat is expected to reduce these times dramatically, improving developer productivity and freeing up more time to work on more innovative projects.

Contextual awareness

With Copilot Chat, GitHub said it’s taking what works for general-purpose generative AI chat assistants and adapting it to the developer environment. 

Located inside the IDE, Copilot Chat works by being contextually aware of the code a developer has written or the error messages shown to them.

The aim is to “keep developers in the flow” at a time when they are having to juggle multiple languages, cloud computing, programming frameworks, CI/CD workflows, open source software, and more.

To tackle this, Copilot Chat offers context-specific support to developers, as well as real-time guidance and suggestions for best practices, tips, and solutions for specific coding challenges. 

The tool also makes suggestions for fixing security issues, and can provide explanations and analysis of code for developers, as well as general troubleshooting. 

“We want to help developers spend their time on what matters most: building what’s next,” Rodriguez said. “And this work started with GitHub Copilot offering code suggestions right in the IDE – but now with Copilot Chat, developers can not only get code suggestions, but ask questions, get explanations, offer prompts for code, and more. That means they’re spending more time in the IDE – and in the flow.”

GitHub Copilot for Business administrators can provide access to the beta to development teams via the settings in the ‘Copilot’ tab. 

The Copilot Chat beta settings can then be found under the ‘Policies’ section, where admins can accept the terms and conditions of the tool. Once accepted, all users within the organization will receive detailed installation instructions via email.

The US is leading the tech industry in auditing codebases for security issues, with the UK reportedly lagging well behind.

Germany was also identified as one of the nations that was underperforming when it comes to code auditing, despite significant cyber security challenges across the industry.

The findings came from open source software firm SUSE’s latest report, showing a disparity in the way in which the nations see code auditing as an operational priority.

According to the report, nearly half (45%) of respondents in the US regard code audits as a priority, and invest accordingly, while only 23% and 26% of respondents in Germany and the UK respectively adopt the same attitude.

SUSE’s global CTO Brent Schroeder said he believes that the US’ potentially more mature DevOps environments could be an influential factor.

“The US being ahead is probably more about the maturity of the US with DevOps and DevSecOps,” Schroeder told ITPro.

Citing his experience with meeting customers, Schroeder said the importance of bringing the integration of security and security practices into the developer pipeline and notes that “companies, at least in the US, are really starting to embrace and recognize that”.

“If they don’t bring security into the process, they encounter one of two things: One is the speed and agility with which code is delivered is significantly diminished because near the end of the process they have to do checks for security.

“They do everything they can to do the integration as quickly as possible but then releasing new applications, major new features into a production environment, they’ve got to pause to check with the security team: does this pass all the audits and the requirements?

“Or else you deliver vulnerabilities at scale.”

Who cares about source code audits?

Being aware of what is in one’s software supply chain is critical. Recent security incidents have demonstrated the importance of detecting, remediating, and monitoring vulnerabilities in applications.

Across the US, Germany, and the UK, an average of 33% of respondents to the survey believed that goals on source code audits would be revised upwards, rising to 46% if one only considers software and network engineers, technical architects, and developers.

95% also intended to review their software supply chain to increase security. This included 51% that had already done so, increasing to 68% of US-based respondents but going down to only 40% of those that are Europe-based.

Why are the UK and Germany lagging?

The difference in approach could potentially be attributed to governmental and regulatory approaches. 

In the US, the M-22-18 memorandum set a deadline for compliance with the National Institute of Standards and Technology (NIST) Secure Software Development Framework (SSDF), SP 800-218, and the NIST Software Supply Chain Guidance.

The M-22-18 memorandum, dated 14 September 2022, set clear dates for US government agencies to adopt the requirements. 

Ninety days were given for a software inventory, 120 days for a vendor communication process, and 270 days for attestation letters not posted publicly by software providers for “critical software”.

US companies keen to do business with government agencies must therefore ensure they comply with the NIST requirements, aimed at addressing software security and secure development practices.

The EU’s Network and Information Security (NIS) directive was the first piece of EU-wide legislation on cyber security but, as a briefing on NIS2 in February 2023 noted, implementation proved difficult and resulted in fragmentation across member states.

NIS2 entered into force on 16 January 2023 and is set to be implemented in each member states’ national law by 17 October 2024.

As firms push to become more efficient, a strong internal development community can help them embrace agile development. So-called citizen developers – employees who aren’t developers but have ideas for internal apps – could be instrumental in this change.

Aerospace giant Rolls-Royce has embraced its internal developer community to drive change across its manufacturing and corporate environments. Through Microsoft’s low-code Power Apps platform, the company has allowed employees to overcome internal work hurdles via self-made software, as well as to build a lively developer community.

“This all started when as a team we asked ourselves the question: ‘What if 1,000 people could do something that would save £10,000 each?’,” says Stuart Hughes, CIO and chief digital officer (CDO) at Rolls-Royce, tells ITPro.

“That's actually really significant digital transformation, right? It fundamentally changes our finances, but also hopefully would give us a better product for our customers that meets that kind of availability reliability and efficiency.”

Recruiting citizen developers to build new apps

Rolls-Royce develops engines for a number of wide-body aircraft used for long-haul or cargo flights. The firm is a key part of the aerospace supply chain.

An app that employees have already created through Power Apps is a 24/7 on-call system for workers in R&D, allowing for internal innovation to happen around the clock.

“If a plane’s canceled, it can actually affect up to about 8,000 people for a single plane cancelation, so it can be quite significant if we don't do a great job.”

The firm tends to charge per engine flying hour rather than per unit sold, so ensuring engines work as intended and don't take long to be repaired is core to its business model. Hughes notes with the employee-developed on-call system, Rolls-Royce avoids spending on a third-party system.

Another example of internal success stories for the firm can be found in Kudos, an-employee led app through which colleagues can easily praise one another. Rolls-Royce credits Kudos with a 25% increase in employee engagement with the company culture, and helped remote workers feel more connected.

Despite these positive statistics, Hughes denies these micro-innovations are best judged by the cost savings they’ve generated.

“Microsoft is very keen for me to tell you in pound coins how much we’ve saved,” he says. “But I don’t think that’s what this is really about. If somebody asked them what digital transformation they’ve done, they can tell you the answer. And that is probably quite rare.”

This community is building itself

It became clear from the start that promoting low-code development was a cultural and strategic challenge rather than technical, head of self-service technology Phil Kaufman explains. His team worked to identify “evangelists” within the organization they could empower first and foremost.

“Actually, at that point, we knew the technology we needed, we knew the platform. It was about ‘who are the people that are going to actually try and help us to make this change across the business?’ and they’ve now become our super users, our champions that we use to do with other people in the business.”

The team had to carefully balance the power and potential of the platform with the user experience for these enthusiastic employees. Kaufman says Power BI allowed his team to enter into a prototyping ‘wire-framing phase’, during which mockups could be made.

“We could get there really quickly, but we could test with it. So we could put it out to a community of one hundred people who use these checklists all the time, and get some feedback,” he says.

Hughes notes the speed at which employees and the IT team could report back on changes was integral to the program’s success, as it allowed the new systems to be implemented in a way that had previously been impossible. By delivering insight into a process that you couldn’t see before, you start to deliver genuine business benefits, he adds.

While Hughes and Kaufman had worked to select enthusiastic Power Apps evangelists to kick the project off, neither expected the speed and enthusiasm with which this has spread throughout the company.

“I signed the purchase order (PO) for the licenses and agreed on some services and things like that. But I didn't do this all, the community just started to create itself, right?”

Inspiring employees to take the initiative

The firm now runs a yearly digital expo, in which citizen developers run stalls showcasing their apps and explaining them to others within the company to spread awareness. Kaufman tells ITPro that after the most recent expo, five teams reached out to follow up on a shift handover app they had seen in employee-led demonstrations.

“I think having those applications, the big ones that people know, people then ask “how have they done this?” and it inspires people to come to the team,” Hughes says.

In hindsight, Hughes says, any other approach wouldn’t have led to the success in this area that Rolls-Royce has enjoyed. “That’s my advice to other CIOs: if I tried to organize that event, it had been overly formal,” he stresses.

The same applies to the approach employees have taken to helping each other with errors or developmental hurdles, in contrast to the typical IT ticketing system many firms use. In contrast, Rolls-Royce employees have taken to custom channels in Microsoft Viva and Teams. Kaufman compared these to StackOverflow, noting the shared language of Power BI has allowed for inter-departmental collaboration that couldn’t be anticipated.

This isn’t to say more formal events haven’t continued to add to this community. Microsoft, for example, regularly holds events at Rolls-Royce offices and streamed online to inspire citizen developers to change through discussions of emerging technologies.

The first steps in upskilling the workforce

Today, Rolls-Royce is engaged in formal upskilling projects through Microsoft’s Enterprise Skills Initiative. Workers go through a four-day course and can pursue an exam to obtain a certificate if they wish. 

But Kaufman also highlights the more informal, self-led learning platforms for budding citizen developers. His team has overseen the ‘app in a day’ program, in which employees can take 30mins of their day to build an app.

“We tried to think about the obstacles that would stop somebody actually doing this,” Hughes says, giving the example of employees hitting issues with poor UI quality and noting that to remedy this a built-in company style has been applied to the app builder Cobalt.

All design elements in Cobalt are pre-defined, which Hughes says keeps the app from gatekeeping the app builder from those who aren’t confident in design.

Going forward, Hughes suggests that if 15% of office-based staff could be using data science tools in the near future, it would continue to drive efficiency gains across the company. “I think that's where this goes,” he says,” adding: “We've got to democratize data science and big data in the same way that we’re obviously demonstrating now with apps.”

Red Hat has unveiled its Red Hat Developer Hub as a fully-fledged portal to ‘supercharge’ the developer experience.

The product is a former open source project from Cloud Native Computing Foundation (CNCF), called Backstage.io, and is now the first standalone, enterprise-grade version of it.

Red Hat Developer Hub, according to the open source software firm, unifies and standardizes the developer experience, maximizes productivity, and enables individuals and organizations to make the most of existing investments and skill sets. 

The new portal - unveiled at this week’s Red Hat Summit in Boston - is unified and open, and essentially offers developers an internally curated and managed toolset to ensure they can quickly and easily turn their software visions into successful reality.

It’s all about maximizing what enterprises have already invested in at a knowledge and financial level as well as increasing velocity and reducing friction, the company claimed. 

“Red Hat Developer Hub adds support for key Red Hat developer plug-ins to drive a more consistent user experience without compromising on that application velocity. The platform also contains validated templates," said Chris Wright, CTO at Red Hat.

"These provide opinionated, pre-architected approaches to building and deploying software from creation to production.

“Developer Hub gives IT organizations a developer platform built on open source innovation with that stability, that support, that reliability that you’ve come to expect from our hybrid cloud technologies. Developer Hub also builds on skills that  IT organizations have already through integration with Red Hat OpenShift.

“That’s the best news. It’s building on skills and technologies that you already have. This maximizes what you know without having to sacrifice application innovation to maintain operations.”

Wright also cited Salesforce’s 2022 ‘Great Resignation’ research, which revealed the majority (76%) of firms feel the cognitive load needed to get to grips with new software architectures is just too high. 

As a result, productivity and sense of achievement decrease rather than increase. 

Developer Hub aims to redress the balance here, according to Wright. 

“As IT resources grow ever more scant and demand for new applications just keeps growing, these platforms grow more daunting. If these platforms aren’t maintained at the speed of development, then CIOs can look forward to fragmentation, complexity, and worse. The use of unverified code that can find its way into production…

“Adding a net new platform all but guarantees a  new learning curve, which equates to hours lost in training that you really don’t have.” 

With a view to removing barriers and accelerating innovation and results-driven performance, the Red Hat Developer Hub:

• Offers a single view of all the tools and resources available for developers as they look to boost productivity.
• Enables users to self-serve and ensure best practice across the board.  
• Maintains the proper levels of security and governance. 

Wright said that the focus was on taking “differentiated applications to market as quickly and as painlessly as you can".

What’s more, Wright acknowledged the unsung IT heroes outside of the developer world, saying it was those people that kept “developer workflows rolling smoothly, with the right tools, the validated environments, and on-demand services".

Ultimately, Developer Hub should free up developer bandwidth in a number of ways. 

“This helps developers focus on crafting beautiful code and not the plumbing underneath,” Wright added. 

As the sun sets on Appian’s annual conference, it’s clear artificial intelligence (AI) now plays a larger role for the company than ever before.

Whether it’s the integration of natural language prompts and digitization of documents in Appian 22.3, or new documentation capabilities for the platform’s own language SAIL, attendees have been treated to a buffet of AI options. Indeed, there’s a real sense that Appian is angling to make itself the go-to platform for AI usage among firms that aren’t yet sure if AI is for them, or otherwise consider it too risky in its current form.

Speaking with customers and officials alike, there’s a sense that Appian’s tender approach to integrating large language models (LLMs) within its platform is going down well. Depending on who you talk to, one gets the sense this is seen as either a necessary precursor to a larger expansion into generative AI tools, or a natural extension of Appian's existing suite of AI and automation offerings.

Adam Glaser, SVP of engineering at Appian rebuffs my suggestion the firm has done a remarkable about-turn to point itself so fully towards AI adoption in the six months since Appian Europe. He also denies the newfound support for developers creating their own AI models is a response to the market.

"We've had AI in the product in various forms of fashions, for quite some time," he states, pointing to Appian's machine learning-based smart recommendations and its intelligent document processing. "That's just a progression of us looking at AI, not even anticipating the hype and interesting things you can do with that GPT-style AI. It's 'what's practical? What's actually really useful?'"

Appian’s cautious optimism

Only time will tell whether this risk-averse strategy is likely to work, and whether Appian picks up pace to roll out generative AI solutions. There are, though, several indicators of an enthusiasm to see more of the technology at an executive level. 

Every image used in the first day’s keynote presentation was generated using the generative AI image model Midjourney, as ITPro was cheerfully informed by both Karina Buschsieweke and Malcolm Ross, director and SVP of product strategy, respectively. 

In addition, the script for the keynote's introductory video was generated in ChatGPT, as personally overseen by Ross. “We did edit it a little,” he admits. “It was a collaboration between AI and humans.”

Beyond the clear AI enthusiasm, these decisions also act as a suitable microcosm for Appian’s entire approach going forward. Buschsieweke and Ross, for instance, both note the output seen by conference attendees was the result of an iterative process involving precise natural language prompts.

Whether the end result was images of Toyota factories or leafcutter ants (both carried metaphorical weight in the proper context) none of the final images were without fault on the first Midjourney pass.

It’s this precise experience that Appian claims that it aims to eliminate with its low-code approach to LLM training and input. 

Despite all it’s shown off in the conference, from natural language input with its Data Fabric to advances in process mining, this balance of fine controls without too much need for experience could be a hard needle to thread.

It's promising, however, that Appian is approaching it as a low-code, user experience problem first and foremost rather than trying to iron out hallucinations or other foundational issues within the LLMs themselves.

"Our focus on generative AI in our product is really around having it embed inside the design time experience," Ross tells ITPro. "Because what you get back from it needs to be immediately refined, you'll find that it's not written quite right or interpreted [your prompt] wrong. So it's a very good complement to low-code design, where you get the initial result and you drag, move it around and tweak it to the final design."

More AI is waiting in the wings

There’s also a sense much is left unrealised in Appian’s current approach. Caution makes sense here, but it appears to have come at the expense of even more powerful functionality for systems such as process mining.

Buschsieweke says worries over generative AI hallucinations have dashed initial interest in integrating LLMs within the core solution.

“We had these discussions already… I wanted a chat for process mining, even before ChatGPT was released, a conversational interface with the user” she asserts. “But it was exactly these concerns: ‘how can we make sure that the insights are actually reliable?’ My perspective was that maybe we should just get started for now to see how it can work.”

The second keynote, headed up by Appian co-founder and chief marketing officer Michael Beckley, comprised in-depth look at current and future AI features. As is standard for promises made in keynotes, the audience was shown a disclaimer 

“This is what I believe today,” Beckley announced to the audience. “This is what we believe today, and with a rapidly-changing field like AI, the product roadmap is subject to rapid change.”

Appian has positioned itself remarkably well, in light of the AI competition. In staying relatively still while the AI market has rushed to innovate, it has counterintuitively come out on top in terms of its public image. It’s an approach that lives or dies with how its customers take it. But the Appian platform is well-placed to deliver on its promises for enterprise AI.

The European Union (EU) should consider replicating the US government’s approach to open source software regulation with its Cyber Resilience Act (CRA), according to a leading expert.

Speaking to ITPro at KubeCon 2023, Gabriele Columbro, general manager for the Linux Foundation in Europe, said the CRA’s current iteration would “change the entire community dynamic” and hamper European businesses.

An open letter signed this week by major industry organizations, including Linux Foundation Europe, said the proposal in its current form will have a “chilling effect” on open source software development.

Under the current iteration, open source developers themselves would be held liable for vulnerabilities in software.

By contrast, the US government’s Cyber Security Strategy, unveiled in February this year, excluded liability for open source software developers and projects. 

The decision was hailed at the time as a highly positive move from the Biden administration. 

Columbro warned that the EU’s current approach could have a significant impact on the European open source community if the bill was to pass in its current form. 

“I think this will have an impact at multiple levels,” he said. “The one I’m worried about the most is what this will mean for our contributors. I think the Linux Foundation and others will feel an impact, but we’re maybe better positioned than others to address those.”

“In the worst-case scenario, individual contributors could be worried that they’ll be slapped with lawsuits and eventually just decide not to put a project in the open.” 

Columbro noted that the inclusion of liability for open source projects appears ‘idiosyncratic’ due to the fact that Europe actively benefits from open source projects and software. 

Millions of businesses spread across 27 member states rely on open source software, which is developed and maintained by a dedicated pan-continental community. 

The entire community dynamic will change,” he said. “And this would primarily hamper Europe. Because outside of Europe, open source innovation will continue - will GitHub have to block open source downloads from inside Europe?”

“Europe bets a lot on open source, but with this it shoots itself in the foot.”

The CRA’s wider failings

Columbro noted that with the EU’s approach to open source software, which appears to have been influenced by the disastrous Log4shell incident, there are particular differences. Not least of all with regard to national security concerns. 

In the wake of Log4Shell, the White House actively sought to engage with software vendors and open source communities such as the Linux Foundation to plan its future approach and mitigate potential supply chain risks. 

But this was from a national security perspective, he added. The US government has a national security mandate while the EU does not, instead delegating national security to individual member states. 

This creates somewhat of a headache for the Linux Foundation and other communities, he said, who are now faced with engaging with multiple different governments across the union. 

“The reason the US was so quick in the wake of Log4shell to bring together the Linux Foundation, Apache Software Foundation, and all the major tech vendors to discuss the response was, of course, from the angle of securing their critical infrastructure and national security.”

“That’s because the federal government in the US has a national security mandate. The EU doesn’t have a national security mandate, that is delegated to individual sovereign states. 

“That is a completely different approach that we need to take as an open source community to work with the 27 member states. But that’s a more humongous approach.” 

Positive engagement

Columbro’s comments follow the open letter’s publication and he is confident that this could spark a positive discussion - and outcome - about CRA-related concerns. 

“The intent of the open letter was to offer up a global, broad consensus with deep experience in this world as foundations,” he said.

“It was really to let the European Union know that we can work together with each other and we want to create an ongoing dialogue and offer up a conversation to help refine [the act].”

“I do think, ultimately, that there is definitely willingness from the EU to improve. I am positive that there is going to be a positive outcome to the CRA,” he added. 

After serving a prison sentence, a former Nintendo hacker has been ordered to pay between 25-30% of his salary to the company for the rest of his working life.

Gary Bowser had previously been a key member of hacking group Team Xecuter which developed and sold chips that allowed pirated games to be played on various game consoles, including the Nintendo Switch, 3DS, and NES.

Nintendo pursued legal action against the group for damaging its business model. The firm has sought $10 million in damages from Bowser, 53, which he will now be forced to pay back through docked wages.

During the court proceedings, Bowser alleged that he had been the group’s “salesman” and earned a total of $320,000 from his Team Xecuter activities. 

He owned the website MaxConsole, which at the time was being used to secretly advertise Team Xecuter products. 

In 2021, Bowser pled guilty to two hacking charges and acknowledged his participation in cyber criminal activity. 

He was sentenced to 40 months in prison and told to pay $4.5 million by the judge, a separate fine from the settlement for Nintendo. 

His early release has come as a result of good behavior while in prison, as well as in acknowledgment of time served while awaiting trial. 

He is currently being held in a separate facility pending his transfer to Canada.

Bowser’s lawyer has claimed that his client lost 90lbs (40kg) in the time between his arrest and court appearance.

In court transcripts, first reported by Axios, US District Judge Robert Lasnik stated that there was a “role to be played here in terms of a message” when it came to Bowser’s sentence.

“Nintendo appreciates the hard work and tireless efforts of federal prosecutors and law enforcement agencies to curb illegal activities on a global scale that cause serious harm to Nintendo and the video game industry,” the firm stated in a press release following the sentence in 2022.

Bowser told YouTuber NckMoses05 that he has already paid $175 in $25 monthly payments, through his job at the federal prison in which he was incarcerated.

He also said that the $4.5 million will not have to be paid back once he is transferred back to his home country of Canada.

Commentators on the story are largely divided on Nintendo’s decision to pursue legal action against Bowser so heavily.

Some have argued that it’s unjust and that the move will condemn Bowser to a life of “extreme poverty”.

Others highlighted the fact that Team Xecuter has acted maliciously in the past not only to the companies which its piracy adversely impacts, but its users too, arguing that the group deserved such a punishment.

The group’s SX OS, software to allow pirated games to run on Nintendo Switch, included code that could render the owner’s console unusable forever if users operated outside the terms of its license.

While Bowser was extradited to the US from his then home the Dominican Republic, Louarn has not been extradited to date, and Chen remains unaccounted for.

ITPro has approached Nintendo for comment.

AWS is making its largest foray into the generative AI space yet with the launch of Amazon Bedrock, a managed service offering customers the “easiest way to build and scale enterprise-ready generative AI applications”.

The hyperscaler unveiled Bedrock on Thursday alongside the launch of two new ‘Amazon Titan’ foundation models (FMs) as it seeks to provide developers with access to “some of the most cutting-edge FMs available today”.

Bedrock will grant customers access to a range of foundation models from third parties including AI21 Labs, Anthropic, and Stability AI. 

Among these is the Jurassic-2 family of multilingual LLMs from AI21, which generate text in Spanish, French, Portuguese, Italian, and Dutch. Alongside this offering is Stability AI’s range of text-to-image foundation models, which includes Stable Diffusion. 

The Amazon Titan large language models (LLMs) form part of AWS’ own internal FMs.

The first of these, AWS revealed, is a generative LLM for tasks such as summarization, text generation - to create a blog post, for example - classification, open-ended Q&A, and information extraction. 

The second is an embeddings LLM that will translate text inputs into numerical representations that “contain the semantic meaning of the text”, AWS added.

“Today we are excited to announce Amazon Bedrock, a new service that makes FMs from AI21 Labs, Anthropic, Stability AI, and Amazon accessible via an API,” said Swami Sivasubramanian, VP for machine learning at AWS.

“Bedrock is the easiest way for customers to build and scale generative AI-based applications using FMs, democratizing access for all builders.” 

AWS enters the generative AI ring

The launch of Amazon Bedrock marks AWS’ most aggressive move in the generative AI space so far and takes a different approach to that seen by competitors such as Google and Microsoft. 

Microsoft has fostered closer ties with OpenAI in recent months and accelerated the integration of platforms such as ChatGPT within core product offerings. 

In January, the firm launched its Azure OpenAI Service which provides cloud customers access to OpenAI models. 

The move was hailed as a key differentiator for the tech giant in gaining ground on AWS in the broader cloud market.

Similarly, Google has made significant strides in developing and rolling out its own internal generative AI system, Bard. 

With Bedrock, AWS looks primed to take a different approach to competitors in the space and capitalize on existing relationships with third parties to augment its generative AI capabilities and product offering. 

Long-term, this appears like it could provide a more developer-friendly approach to generative AI and diversify its offering to prospective customers. 

Competition in the generative AI market continues to heat up, and AWS’ quiet approach in this space has been noted in recent months amid the flurry of announcements from Google and Microsoft.

Sivasubramanian suggested that a key differentiator for AWS moving forward will be the ease of access and use, and the sheer diversity of models available to developers. 

“Bedrock makes the power of FMs accessible to companies of all sizes so that they can accelerate the use of ML across their organizations and build their own generative AI applications because it will be easy for all developers,” Sivasubramanian said. 

He added that Bedrock will be a “massive step” toward democratizing FMs and opening up access to a broader range of models for developers within its own ecosystem. 

Sony’s semiconductor arm has announced a “strategic investment” in Raspberry Pi - a move that will see users given access to its edge AI platform Aitrios.

The minority stake investment, details of which remain undisclosed, “cements the relationship” between the two companies and will provide a development platform for Sony Semiconductor Solutions’ (SSS) AI devices to Raspberry Pi’s global user base, the company said. 

This means Raspberry Pi users will be able to leverage Sony’s Aitrios edge computing AI platform to develop a range of AI-based visual sensing applications. 

 “Our goal is to provide new value to a variety of industries and support them in solving issues using our innovative edge AI sensing technology built around image sensors,” said Terushi Shimizu, president and CEO at SSS. 

 “We are very pleased to be partnering with Raspberry Pi to bring out Aitrios platform - which supports the development of unique and diverse solutions utilizing our edge AI devices - to the Raspberry Pi user and developer community, and provide a unique development experience.” 

Eben Upton, CEO at Raspberry Pi, said the partnership builds on longstanding ties between the two companies and will help Sony embed AI products within the firm’s developer ecosystem.  

“This transaction will allow us to expand our partnership, bringing Sony Semiconductor Solutions’ line of AI products to the Raspberry Pi ecosystem, and helping users to build exciting new machine learning (ML) applications at the edge,” he said in a statement. 

What is Sony Aitrios?

Aitrios is an edge AI sensing platform developed by Sony and launched in 2021. 

The platform runs directly on chips, which Sony says provides users with a range of benefits in developing machine learning and AI applications and creates “synergy between edge and cloud”. 

Since its launch, Sony has positioned Aitrios as a go-to platform for developers creating AI camera-driven and edge sensing solutions. 

Aitrios use-cases cited by Sony span a number of sectors, including surveillance, security, retail, logistics, and manufacturing. 

The platform could be used to develop solutions aimed at detecting vehicle license plates or monitoring footfall in retail venues, for example. 

Capitalizing on AI hype

This partnership with Raspberry Pi could signal a move from Sony to capitalize on recent excitement in the AI space. 

Developers have already been experimenting with generative AI systems such as ChatGPT on Raspberry Pi 4 devices, for example. 

Long-term, Upton told CNBC that he envisages the partnership helping young developers learn how computers work in their modern form, rather than how they did 30 years ago. 

He added that Raspberry Pi has been used for ML development for some time now, and that the Sony partnership will allow this to evolve further.

The Raspberry Pi 4 device is largely compatible with several AI and ML frameworks, such as TensorFlow or PyTorch, making it a go-to device for both hobbyists and students experimenting with ML applications. 

However, when speaking to CNBC Upton noted that while the company has traditionally appealed to hobbyists, its popularity among enterprises has been growing steadily. 

Around 70% of its sales now come from commercial clients looking to embed products within consumer devices or general operations. 

Google Cloud has partnered with Replit to create generative AI applications which will help the coding platform to build “bigger and faster”.

Without mentioning it specifically, the partnership appears to be a sign of Google trying to chase Microsoft's heels as its Copilot-branded AI technology seeps deeper into its product line.

Copilot was first introduced to GitHub in 2021 as an AI pair programming assistant and quickly gained wide popularity among software developers for its ability to augment and hasten application and web development processes.

Google has also been playing a game of catchup with Microsoft ever since the latter announced its $10 billion investment into OpenAI, shortly after ChatGPT took off. Google was forced to down tools, upend internal teams, and declare a 'code red' situation to counter the threat posed by Microsoft's rapid expansion into the generative AI space.

It launched Bard, Google's own generative AI to rival ChatGPT, but its launch was fraught with flaws and ultimately rushed to simply be part of the conversation.

Now, with its latest partnership, the 20 million developers on Replit's platform will be able to deploy and scale their applications on Google Cloud, while benefitting from Replit's own AI pair programmer, Ghostwriter.

“Our partnership with Google Cloud on infrastructure and AI is a clear step toward our mission of bringing the next billion software creators online. AI has changed software development forever. The seamless idea-to-deployed software experience that Replit provides developers will continue to set a new bar for software creation,” said Amjad Masad, the CEO and head of engineering at Replit. “We’re thrilled to build this.”

Ghostwriter currently generates over 30% of the code for developers who use Replit, a similar figure to GitHub Copilot which by the company's estimates generates 46% of the code across projects on the platform.

Replit's developers will have access to Google’s infrastructure as well as its foundation models on Vertex AI. The two companies are aiming to accelerate developers’ AI development process through the partnership.

Ghostwriter was launched in 2022 and fully integrated into Replit's browser-based integrated development environment (IDE) in January 2023.

Like Copilot, Ghostwriter offers similar code-debugging and error-solving capabilities delivered through a natural language processing (NLP) prompt.

But with Copilot already amassing a significant user base and having been in development for longer than Ghostwriter, not to mention being integrated into the world's leading version control platform, does Ghostwriter have a realistic chance shot at taking on Copilot and succeeding?

Bola Rotibi, chief of enterprise research at CCS Insight, told ITPro that the partnership was a "necessary" one and that despite GitHub's market dominance, Replit's large user base could mean it will be a serious force in the future.

“In this light, Google and Replit’s partnership can be seen as a necessary move and hook up in the face of Microsoft’s developer community dominance and the launch of GitHub Copilot which is served by OpenAI’s LLM ChatGPT,” she said. “Replit’s supposed 20 million developer base brings added leverage and an established community to Google’s cloud platform services: now that alone can’t be sniffed at.

“The Google and Replit partnership is in direct competition with Copilot and expect this to be just the start of this embedded new capability in the drive to deliver application code faster, with improved quality that raises the productivity of the development process and developers,” said Rotibi.

Rotibi called this the “next stage of this evolutionary thread”, citing the news that GitHub has released Copilot X, which will see it adopt OpenAI’s GPT-4 model to introduce chat and voice for Copilot across the whole application lifecycle process.

Replit and Google may have serious potential in this market, but that does nothing to dampen the fact that GitHub got in first, and that its Copilot user base is both huge and devoted.

ITPro attended GitHub Universe in 2022 and Copilot, over a year after launch, remained the talk of the event. And the tool is only getting stronger with each passing iteration.

What's undeniable is that Replit and Google have a mountain to climb, but they do have pedigree and a great platform to build from.

Copilot is adored in the developer community, but it isn't perfect. This is one area in which the pair could target, perhaps aiming to become the most accurate solution, the one that supports the largest number of programming languages, or the one that can generate the most complex solutions - something with which Copilot can, at times, struggle.

If Replit and Google's technology can target Copilot's weaknesses and become a trusted companion for developers on specific tasks, as well as build a comparably powerful and easy-to-use product, then with time, it has serious potential to claim some of Copilot's majority market share. But, for now, I wouldn't expect Copilot to be too fearful.