We removed ~80% of the Claude Code system prompt for our newest models, this is what we've learned about writing system prompts, skills and Claude.MDs for them. https://t.co/6DZwSrZjE9
What I love about GPT-5.6 Sol in Codex: it turns messy context into verified work. This week it helped me build a cross-agent QA harness, trace a latency bug to root cause, and turn project evidence into stakeholder-ready decisions. It doesn’t just answer. It helps me operate.
Something's wrong when people start making McKinsey 2x2 charts to try to understand what model and effort to use. Hope all of this can be simplified for regular users.
My process for a new project this weekend:
-> Use Fable to build plan.html with design guidelines
-> Use Claude Design to make components + screens
-> Get gpt 5.6 to build