Software engineering with GenAI
summary
Executives claim things like 30% of code written by AI and forecasts of near‑total automation, but engineers report mixed reality: agent experiments sometimes fail spectacularly (a C‑pilot run breaks tests), an agent added a bug costing $700, and a biotech team finds LLM output slower to fix than write. Tools with measurable uptake include Cloud Code with 40–160% usage growth and internal MCP servers, Cider at Google, Amazon Q, Cursor/Windsurf percentages, and reported time savings around three to five hours weekly.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI in dev
- 1:34Industry hype vs reality
- 4:24AI tool startup insights
- 7:48AI adoption in big tech
- 10:50The Amazon API philosophy
- 14:01Startup experimentation
- 16:56Independent engineer perspectives
- 21:17Critical questions on AI usage
- 24:26A shift in software development
related talks
33mAI assisted software delivery
26mAI Agents as Your Co-Developer
37mSoftware Engineering + AI = ? | The Future of Development
21mRevolutionizing software creation: Harnessing the power of GenAI
16mCoding Reinvented How AI Agents Are Shaping the Future of Software Development
01h 07mNew Methods of Working with AI.
01h 01mAre Your Engineering Teams Prepared for the GenAI Revolution
26mA reality check on AI’s impact on business transformation