Prominent AI researcher Andrej Karpathy picks Anthropic over former home OpenAI to get back into frontier LLM research

Source: The Decoder URL: https://the-decoder.com Published: 2026-05-20

Andrej Karpathy, one of the biggest names in AI, is joining Anthropic. The former OpenAI core team member and Tesla Autopilot architect says he wants to get back into R&D, calling the next few years at the frontier of LLMs “especially formative.” That he chose Anthropic over a return to OpenAI is a clear loss for his former employer. French AI company Mistral AI has acquired Vienna-based startup Emmi AI to expand its offering for industrial clients across Europe. The purchase price was not disclosed. Emmi AI specializes in AI models that can simulate complex physical processes such as airflow, heat transfer, and material stress. The startup raised 15 million euros in Austria's largest funding round in 2025, Reuters reports. Mistral builds custom AI systems for clients like ASML, Stellantis, and Veolia that can detect production defects or control robotic arms. CEO Arthur Mensch sees the acquisition as strengthening Mistral's position in aerospace, automotive, and semiconductors. Cloudflare tested Anthropic's security-focused AI model Mythos Preview across more than 50 of its own code repositories as part of Project Glasswing. The model can chain multiple small vulnerabilities into working exploits and prove they're real by writing, compiling, and running proof-of-concept code on its own. Earlier frontier models found similar individual bugs and sometimes delivered solid analysis, according to Cloudflare CSO Grant Bourzikas. But they fell short at stitching the pieces together, leaving chains unfinished and the question of actual exploitability open.Mythos Previewproduced fewer speculative findings, clearer steps to reproduce issues, and needed less human follow-up to reach a fix-or-dismiss decision. Cloudflare stresses th

---

Notes

• Ingested from The Decoder via kb-ingest cron job (2026-05-20T09:09:21Z).