All editions.
Every briefing, fully archived. Five to six new ones every week.
9 briefings. Filtered by openai — clear filter.
- EDITION #333
He Helped Build ChatGPT. Now He Says Chat Was the Wrong Bet.
TypeSafe's Jev skips chat entirely — typed, confidence-scored decisions in under a second, now open for early access.
- EDITION #310
Two Models From the Same Base Scored 1.5% and 95% on the Same Test
OpenAI's GPT-5.6-Cyber found two Chrome zero-days. The number that matters measures willingness, not skill.
- EDITION #284
Why Your ChatGPT Agent Now Only Needs Your Password Once
ChatGPT Work agents can now log into sites for you — and remember it, so you never repeat that step again.
- EDITION #277
Why Your Two-Day Report Now Takes Minutes
OpenAI and Anthropic both shipped office AI agents this month. One venture firm's competitor research already went from two days to minutes. Here's which tool to use for what.
- EDITION #247
Why OpenAI's Best Coder Gave Up Writing Code
He's a "10x engineer" worth nearly $30 billion who loved nothing more than shipping code at 2 a.m. Then he took the hardest job in AI — and the trade he made is one every high performer eventually faces.
- EDITION #186
Wall St Just Spent $5.5B to Skip McKinsey
Anthropic and OpenAI both quietly raised consulting-killer war chests last week. Your industry might be the first stop.
- EDITION #183
Why AI Just Beat ER Doctors at the Hardest Part of Their Job
OpenAI's cheapest reasoning model hit 67%. Two attending physicians hit 50 and 55. The gap widened when info got thin.
- EDITION #171
Why Your $20 AI Plan Is Already Dead
Anthropic stripped Claude Code from Pro for 24 hours.
- EDITION #165
Why the Tell-Tale Sign of AI Images Just Disappeared
For three years, you could spot an AI image by the garbled text. That tell is gone — and the implications go way beyond marketing.
