Mysten Labs Co-Founder Sam Blackshear Steps Down to Join Anthropic

Sam Blackshear, the co-founder and chief technology officer of Mysten Labs, has announced he will be stepping down to join Anthropic to work on defensive security research.

Blackshear, along with four former Meta executives, founded Mysten Labs — the creator of the Sui blockchain — in September 2021. Blackshear said Mysten Labs co-founder and CEO Evan Cheng will step in to set the “technical vision” for Mysten going forward.

“When I reflect on my career, I have been happiest doing hands-on technical work while exploring new problem domains,” said Blackshear. “This is an opportunity to do both, and I’m motivated to work on security at a time when the balance of power between attackers and defenders is shifting.”

His departure comes as AI is increasingly being used to identify software vulnerabilities, automate phishing campaigns and accelerate attacks on blockchain infrastructure.

AI arms race is “getting scary:” Blackshear

In April, Blackshear hosted a panel with security experts about how AI coding agents are reshaping smart contract security.

He said for him, a turning point that led him to realize that something had fundamentally changed was when he ported a static analysis tool he built at Facebook written in OCaml-Java over to Move, the programming language used to build smart contracts on Sui.

Sam said he used Claude to make the move and flag any potential vulnerabilities, and “it just did it,” he said.

“And then when I saw the results, I was like, ‘whoa,’ this is a new world that we’re in terms of like the speed and the capabilities.”

“The triage results was completely manual before, and the coding would have taken a long, long time,” he added.

Crypto firms seeking frontier AI access

Many of crypto’s biggest players have been seeking access to powerful new AI models to strengthen their code from attacks, with many of these models, including those at Anthropic, restricted to only a small group of companies.

Crypto executives speaking to Cointelegraph agreed that while restricting these frontier models initially makes sense, it no longer does once competing, publicly available AI models hit the same level of cybersecurity capability.

“I fully understand guardrails for advanced models, but we need to streamline the verification programs, the acceptance programs, to give these models to legitimate defenders,” said Solana Foundation chief information security officer Michael Coates.

“We need to make sure that the best models we can get are in the hands of defenders because attackers will have something capable enough.”