Anthropic delayed Mythos model release by four months for safety, says CEO
Anthropic chief executive Dario Amodei posted on X to defend the company’s “race to the top” safety strategy after criticism from Nathan Calvin, chief strategy officer at the advocacy group Encode AI. Amodei listed nine decisions that he said cost the firm commercially but made its AI safer, including a four‑month delay of its Mythos‑class models.
Key points
- Anthropic delayed its Mythos-class models from February to June, citing stronger cyber safeguards before general availability.
- Company restricted sales to firms controlled by the People’s Republic of China and supported export controls on chips.
- Anthropic’s new public benefit fund will hold at least 5% of the company’s shares, per Amodei.
Anthropic’s Mythos models were first ready in February but were not made generally available until June, when stronger cyber safeguards were in place and only a small group of partners working on cyber defence could use them. The company also restricted sales to firms controlled by the People’s Republic of China, backed export controls on chips, contributed to California’s SB 53 safety law, blocked potential bioweapon use, and created a public benefit fund that will hold at least 5% of Anthropic’s shares. Amodei said the firm likely sacrificed a several‑month model lead to China’s GLM‑5.3.
The post, viewed more than 800,000 times, acknowledged that three of Calvin’s points could have been handled better and promised a review of the Responsible Scaling Policy. Vox’s Sigal Samuel asked for evidence the strategy works, noting an outside view of a race to the bottom.
Model page: GLM 5.3 →
The story so far
4 episodes →- Anthropic delayed Mythos model release by four months for safety, says CEOthis story
Dario Amodei says Anthropic held Mythos back for four months for safety
thenextweb.com · 21 October 2026
Loading the full article…
This text was published by thenextweb.com and written by Ana Maria Constantin. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Policy & Regulation
All →- Donald Trump calls for AI data centers to pay energy dividends to local communities · 1 src
- DHS bans DeepSeek AI from civilian agencies, extends to third-party products developed using its models · 39 src
- Wikipedia bans AI-generated articles, volunteers say perhaps 15% still slip through in early 2026 · 1 src
- Linux Foundation survey finds AI-generated pull requests make up a quarter of contributions · 1 src
- Jensen Huang urges Europe to cut AI regulation to avoid falling behind · 3 src
Comments
via GitHub Discussions