hdahmΞ 🛡️(@IsTheBaron):Anthropic’s Mythos disclosure and OpenAI’s industrial policy paper landing days apart feels less like coincidence and more like counter-positioning. A model that jumps to 93.9% SWE-bench and cranks through longstanding exploits in the Linux kernel turn exploits turns "AI safety" from branding into statecraft.

hdahmΞ 🛡️

@IsTheBaron

CTO @hack_vc, SOTA Agent Cofounder. Writing python and terraform - not tweets. Bad at NFTs . pre @factionVC, @lightspeedvp, @coinbase, @cbventures

加入 March 2012

656 正在關注 642 粉絲

hdahmΞ 🛡️@IsTheBaron

2026.04.07 21:09

Anthropic’s Mythos disclosure and OpenAI’s industrial policy paper landing days apart feels less like coincidence and more like counter-positioning. A model that jumps to 93.9% SWE-bench and cranks through longstanding exploits in the Linux kernel turn exploits turns "AI safety" from branding into statecraft.