Global developer rates have taken a hit, but no fear about using AI

Software developers don’t fear AI, but pay rates are under pressure as economic headwinds around the world make the halcyon days post Covid all but a distant memory.

Developer Q&A website Stack Overflow’s major global survey of 65,000 software developers shows the average full-stack developers median salary in 2024 was US$63,333, down 11% on 2023. The survey only included 396 (0.7%) respondents from New Zealand, where tech roles have been hard hit with the public sector slashing headcount and businesses reining in budgets too.

The number of software developers who are independent contractors or self-employed has also more than doubled to 18% from 9.5% in 2020, while part-time work has also risen significantly, suggesting that previously full-time employed developers are increasingly going out on their own, maybe as a result of the tough economic times seeing fulltime headcount trimmed.

The unemployment rate for developers worldwide remains low at 4.4%, but it has doubled since 2019, another reflection of the softening market.

Layoffs softening rates everywhere

Most large tech firms announced lay-offs in 2022 and 2023, with the likes of Intel cutting deeper this year as well. The semiconductor maker will shed 15,000 staff globally this year.

But the rise of remote work as a highly-valued working arrangement, a trend that really took off in 2020, will also explain the increase in developers working as freelancers. Of the software  developers surveyed, the majority are in hybrid-working arrangements, with 38% full remote, and just 20% working permanently in the office.

When it comes to artificial intelligence, only 12% of developers see it as a threat to their employment, while 70% are happy to use AI tools as part of their development workflow. Around 63% of developers are using AI for development purposes, up from 44% just a year ago.

That gels with my anecdotal discussions with developers who love using the likes of Github Copilot and other AI-powered coding tools that save them time and free them from mundane coding jobs to do more interesting work.

ChatGPT the reigning AI tool for devs 

Still, Microsoft’s Github Copilot use is actually in decline, according to the survey, with ChatGPT, OpenAI’s AI-powered chatbot able to handle computer code and seemingly able to do a better job. Of those developers who use AI, 82% use ChatGPT, double the number of Copilot users. Tabnine and WolframAlpha also saw declining use in this year’s survey.

Productivity increases were the most common reason (81%) given for using AI tools in the development process, followed by the ability to learn new skills quickly (62%).

Younger developers, those with less experience and those living in countries such as India and Brazil are more likely to use and trust AI, while usage is lower for older developers. 

Some top coding trends

When it comes to coding languages and development tools, Stack Overflows survey reported few dramatic changes since 2023. But a few interesting shifts in usage include:

  • FastAPI and Next.js adoption rose. Use of the Python framework FastAPI rose from 6% in 2022 to 9.9% in 2024. Meanwhile, Next.js usage rose from 13.5% in 2022 to 17.9% in 2024.

  • React and Node.js adoption fell. React usage dropped from 42.6% in 2022 to 39.5% in 2024. Meanwhile, Node.js fell from 47.1% using it in 2022 to 40.8% in 2024.

  • Svelte usage  slowed. After rapid increases in 2023, slightly fewer developers are using Svelte in 2024. But those who are using the framework seem satisfied, as 73% of its current users want to continue doing so in the next year.

  • Vite is increasingly used. The frontend tool is compatible with multiple frameworks. Vite was first covered by the survey in 2023, when 14.7% of developers reported using it. That rose to 19.9% this year.

  • Newly asked about in 2024, only 3% of the survey are using the highly touted Astro.

Previous
Previous

New report highlights challenges in implementing digital technologies curriculum

Next
Next

Data centres are guzzling up too much electricity. Can we make them more efficient?