Microsoft AI drafts Code of Conduct the same week the industry called for a slowdown
BenBajarin · x · 2026-09-20
Analyst Carolina Milanesi's essay connects two events: Microsoft AI published a draft Code of Conduct for its MAI models on September 14, open for six weeks of public comment, with a revision due by year-end to guide 2027 model development. Two days earlier, Anthropic CEO Dario Amodei published "We Must Pace the Frontier," urging the industry to slow capability gains — endorsed within hours by Sam Altman and Elon Musk.
- Amodei's three-part proposal: embedded third-party evaluators (METR etc.) with badges and access comparable to internal risk teams; common safety standards among leading democratic countries; eventual global coordination including narrow agreements with China. Anthropic unilaterally committed to the first; OpenAI said it would follow. Altman also told Fortune OpenAI's IPO slips to 2027 on safety grounds.
- Milanesi's critique: "slow down" is being treated as a single dial, but AI isn't one thing at one speed — risking both lost work with clear human benefit and ignored dangerous uses. Both the code and the slowdown call need "a decision that costs something."
More from AGI Musings
- Will 60,000 ICLR Submissions Kill Academia? Insiders Say Reputation, Not Papers, Rules — JundeMorsenWu · 2026-09-21
- Agents may discover million-times-cheaper training, making data centers look silly, predicts Steve Moraco — menhguin · 2026-09-21
- Always thinking one model ahead: exponential AI iteration means each release is already outdated — haider1 · 2026-09-21
- Hinton's call against private ownership of 'means of computation' gets the Pope-vs-printing-press treatment — Dan_Jeffries1 · 2026-09-21
- You must know slop to know craft: discernment is embodied judgment earned over time — floguo · 2026-09-21
- Deep learning can't patch its own blind spots, researcher argues — that's why it stalls — ryunuck · 2026-09-21