Published on the 08/04/2025 | Written by Heather Wright

Biggest users, sharpest critics…

Developers are feeling the weight of AI expectations with AI coding tools increasing pressure to deliver faster, at a pace that’s becoming unrealistic. And while they’re embracing AI tools, there’s plenty of eye rolling at some of the claims about the technology.

HackerRank’s 2025 Developer Skills report surveyed more than 13,000 developers across 102 countries.

“Hands-on experience makes developers both AI’s biggest users and its sharpest critics.”

It found plenty of use of AI – 97 percent of developers are using at least one AI assistant, with 61 percent now using two or more AI tools at work, jumping between chat-based LLMs like ChatGPT (which remains a staple for developers), Gemini and Claude and blending them with developer-focused tools such as GitHub Copilot and Cursor.

In fact, AI is now generating, on average, 29 percent of developers’ code – though the report notes that number fluctuates widely depending on usage patterns and job roles, with security, cloud and data engineers leading the way in AI -generated code.

The use of generative AI tools for software development was an early use case for the technology, helping make coding cheaper and easier. In fact the early days of generative AI saw plenty of talk of the death of software development. That’s been tempered considerably with developer judgement and experience vital given risks associated with AI-generated code.

In HackerRank’s survey, developers were sceptical about AI for code completion, with 38 percent saying it was overhyped for that purpose. Other common AI uses, including code review and debugging, also topped the list of tasks developers said were overhyped – ‘not because AI can’t do them, but because the hype machine oversells what it can actually deliver’. In fact, developers reported seeing most AI use cases as overhyped, with the exception of documentation where there is more straightforward value afforded by AI.

“Hands-on experience makes developers both AI’s biggest users and its sharpest critics. They’ve seen AI do things sceptics say it can’t, but they also know where it breaks – where models stumble, generate nonsense, or require more oversight than promised.”

When it comes to the actual use of AI, four key areas stand out: Learning new concepts, code review, debugging (yes, two that are also mentioned as ‘overhyped’) and codebase understanding.

While ChatGPT was a leader across most tasks, and especially in learning, other tools are also getting a look in. In particular, when it comes to refactoring, GitHub Copilot and Claude are favoured while Cursor stood out for testing, integration and deployment.

Developers who had been early adopters of AI were getting value from the technology sooner, getting better results and reporting faster work as a result, but the report questioned whether this was the result of AI itself, or whether those developers were just better at making the most of new tools – with the developers themselves potentially wired to move faster.

Either way, expectations are rising. Sixty-seven percent of the developers surveyed said AI has increased pressure to deliver faster. And management are pushing even harder – with 84 percent of engineering leaders saying they’ve raised productivity expectations for their teams.

“As AI adoption scales, expectations are shifting faster at the top than on the ground. The question is whether companies are adjusting to reality – or just assuming AI makes everything instant.”

The report also sounds a warning for companies that aren’t investing in their developers, whether through pay, career growth or meaningful work – with developers signalling their prioritising career growth (49 percent) and keeping up with new technologies (42 percent), and only too happy to move on if the company doesn’t come to the party. And they’re not just switching jobs, they’re switching fields.

Many reported that they’re choosing their own learning paths, with companies failing to deliver on developers’ desires for growth.

More than 80 percent of engineering managers noted concerns about unclear internal career paths, lack of leadership support and training gaps.

The upshot? Sixty-one percent of developers without learning opportunities plan to leave with a year. Ninety percent say they’ll be bidding farewell to the company within two years. And overall, 40 percent of developers say they’re looking to exit their current role.

The AI push is also reshaping hiring trends, with businesses favouring more senior talent in their developer searches. That’s something HackerRank attributes, at least in part, to a fundamental shift as AI reshapes junior developer responsibilities.

Other factors too, are playing a role, including the higher expectations of companies who want new hires who can deliver immediate value, and cautious recovery with companies potentially prioritising experienced talent they can build teams around.

“The role of a developer isn’t disappearing, but it is evolving,” HackerRank says. “Developers are already adapting. The companies that evolve with them – the ones that rethink hiring, invest in learning and build real developer engagement – will be the ones that win in 2025.”