"The amount of AI-generated code worldwide has grown sixfold over the past two years, from 5% in 2022 to nearly 30% by the end of 2024. Along with it has come a measurable increase in programmer productivity. These are some of the findings from a new study by the Complexity Science Hub (CSH), which examined the impact of AI tools and platforms on software development."
"There's a fissure in the data, however: less-experienced programmers use AI more frequently (37%), but productivity gains are seen almost exclusively among experienced developers. Gen AI reshapes both the volume and nature of programming work, according to the CSH research team, led by Simone Daniotti. "Comparing the same developer before and after adopting gen AI, we show that AI adoption substantially increases output."
AI-generated code worldwide rose from 5% in 2022 to nearly 30% by the end of 2024, a sixfold increase. Generative AI correlates with an approximate 4% increase in programmer productivity overall. Less-experienced programmers adopt AI more frequently (37%), yet measurable productivity improvements are concentrated among experienced developers. AI adoption increases individual output and encourages use of novel combinations of software libraries, supporting exploration of unfamiliar technical domains. The expansion of AI-generated code and productivity shifts imply significant cost implications for software labor, with US programming-related labor costs estimated to exceed $600 billion annually.
Read at ZDNET
Unable to calculate read time
Collection
[
|
...
]