Our Agents, Ourselves
Plus: Turning your best colleagues into reusable skills, the model built to fix AI writing, and math proofs that outran their authors
TL;DR
- Updated guides for ChatGPT and OpenClaw reflect the merger of Codex into ChatGPT, offering new workflows and insights for knowledge work.
- A Q&A session with executives reveals common questions and advice on adopting AI, covering strategy, governance, and team organization.
- The concept of 'cloning' coworkers into reusable AI skills is being explored as a way to capture expertise within companies.
- A new AI model, Deft, aims to fix AI writing's sameness but produces dense, fact-inventing output.
- Mathematicians face 'proof indigestion' as AI generates more proofs than humans can verify, leading to a potential shift in the division of labor between humans and machines in mathematical research.