Jellyfish AI development study: The real sting has yet to land

Jellyfish's AI Engineering Trends study, drawing on data from 700+ companies, 200,000 engineers, and 20 million pull requests, finds AI coding tools are now standard across engineering teams. Top AI adopters have doubled pull request throughput, but fully autonomous code agents are growing exponentially and bring new risks. Experts warn that increased code output is straining code review processes, with AI-generated code often looking correct until it isn't. Open source maintainers are also feeling the burden of low-quality AI-assisted submissions. The study highlights a widening gap between aggressive AI adopters and laggards, while cautioning that productivity gains come with new verification overhead.