A Jellyfish survey of 636 software development professionals finds 64% report at least a 25% increase in developer velocity from AI, with 80% viewing AI as a net positive for productivity. Top AI use cases are code writing (53%), code review (49%), and code explanation (43%), with Claude Code, Gemini Code Assist, and GitHub Copilot leading adoption. Challenges persist: only 53% say AI improves code quality, 42% cite rising tool costs, and 36% face resistance from senior engineers. A growing gap is emerging between organizations that have mastered prompt and context engineering and those still experimenting. Only 43% rate their AI adoption as high or very high, and 64% say they still need better data to diagnose engineering productivity issues.
Sort: