Sam Blackshear, a co-founder of Mysten Labs, has moved to Anthropic to work on defensive security research. The shift, confirmed by the company, brings a veteran of blockchain infrastructure into the AI safety space.
What Blackshear will work on
At Anthropic, Blackshear will focus on defensive security — the side of cybersecurity that aims to protect systems rather than find flaws in them. The role sits inside the company's safety research group, which works on making large language models harder to misuse or attack.
Blackshear was a key figure at Mysten Labs, the developer behind the Sui blockchain. He helped design the Move programming language, which emphasizes safety and verifiability in smart contracts. That background in building secure-by-design systems is likely what drew Anthropic's attention.
Why the move matters
Anthropic has been hiring researchers with deep technical backgrounds as it scales up its safety efforts. The company's flagship model, Claude, is built around principles of helpfulness and harmlessness. Adding someone who spent years thinking about how to make blockchain code less vulnerable suggests Anthropic is looking beyond traditional AI safety methods.
Defensive security research at Anthropic could involve hardening models against adversarial attacks, preventing data leaks, or building guardrails that stop models from being used in harmful ways. Blackshear's experience with formal verification and secure programming languages fits that agenda.
Mysten Labs did not immediately respond to a request for comment about Blackshear's departure or who will take over his responsibilities.
Blackshear is expected to start at Anthropic in the coming weeks. The company has not said whether he will lead a team or work as an individual contributor. His hiring is part of a broader push by AI labs to bring in engineers from adjacent fields like blockchain and cryptography, where security is a first-class concern.
