Panel: Productivity gains through AI?!
summary
Panelists argue there is no rigorous evidence yet and small experiments beat big orgs: a widely cited 15-person Copilot study is underpowered, randomized trials show spotty gains, and effects fade after a burn-in period. Hands‑on use of agents and cloud code tools (Cursor, early Windsurf, Copilot extensions) powers solo projects — one volunteer rebuilds a frontend with a backend in 6 hours — but large legacy jobs (APL mainframe migrations, compiler rewrites) need human design and rarely fully auto‑succeed; teams report raising coding time from ~10–15% to 55–65% when reorganized.
AI-written summary · about
chapters (13)
- 0:00AI productivity debate
- 2:07Panelists' AI stances
- 8:10Measuring AI productivity gains
- 13:08AI's impact on teams and jobs
- 19:06Legacy code modernization with AI
- 24:27Future of software development
- 31:08CTOs' AI tool experience
- 36:00Adoption challenges and inertia
- 41:47AI in marketing and sales
- 46:21Client expectations and AI transparency
- 48:05Go-to-market with AI assistance
- 49:27AI-assisted vs AI-driven development
- 51:10Q&A on AI and software
related talks
01h 02mPANEL - Unleashing the Future: The Impact of Gen AI on Development Productivity
57mOne Team’s Journey to Using AI Without the Hype
41mPanel discussion: AI in code development – hype vs reality
26mA reality check on AI’s impact on business transformation
01h 02mDeveloping resilience for the future of work
25mPanel The State of Web Development in 2026
27mSoftware engineering with GenAI
42mPanel: Can AI transform business?