What you'll take away
- Software 1.0 is code, 2.0 is trained weights, 3.0 is prompts written in English
- Treat LLMs like an operating system with odd strengths and odd memory gaps
- Build partial-autonomy tools with an autonomy slider, not fully hands-off agents
- Keep the generate-then-verify loop short so a human can check the work quickly
More talks
Workshop1h 37mFull Walkthrough: Workflow for AI Coding
Matt Pocock · AI Engineer Europe 2026
Matt Pocock's whole process for building real software with a coding agent, from the first question to reviewing the finished work.
Talk26mClaude Code best practices
Cal Rueb · Code w/ Claude 2025
One of the people who builds Claude Code at Anthropic on how to get the most out of it day to day.
Workshop25mPrompting 101
Hannah Moran and Christian Ryan · Code w/ Claude 2025
Anthropic's Applied AI team build a real prompt step by step, from a vague first try to one that works reliably.