Unsession: "AI coding tools"
summary
Discussion of Copilot, Cursor, ChatGPT and Claude with Clojure details: autocomplete delivers the biggest gains, often saves ~1 hour/day for active coding; Cursor’s multi-insertion and navigation convenience beat simple extensions for single-file edits; model-backed reformatting (JSON→EDN) succeeds about 95% success with one retry; longer model queries take ~15 seconds on Sonet 3.5 and can limit concurrency; multi-file edits and hallucinations remain common pitfalls; kill Cursor before presentations to avoid leaking local API keys.
AI-written summary · about
chapters (12)
- 0:00AI coding tools intro
- 2:12Who benefits most
- 3:15Productivity gains with autocomplete
- 8:05Live demo: AI code editing
- 13:13Multi-file changes with AI
- 19:18Why Cursor wins
- 21:16Senior devs and AI adoption
- 24:40Balancing AI integration
- 26:57AI in hiring and evaluation
- 28:29AI for creative tasks
- 32:26Security with AI tools
- 34:03Closing & Q&A
related talks
47mChasing the Main Thread - Adventures in AI Assisted Coding
57mAccelerate Your Developer Productivity with AI Embrace the Future Now!
45mPanel Discussion on AI for Code Generation
47mAI-Powered Software Development From the Trenches
51mLive Panel Discussion: Less Code? More Vibes? Web Dev in the AI Era
45mSupercharge your coding: Integrating ChatGPT and Copilot in long-term projects
32mBeyond the Hype: A Pragmatic Look at How Developers Actually Use AI Tools
01mThe AI Tools Kent C. Dodds Uses