Mysten Labs tech chief joins Anthropic to work on AI security

When the co-founder of a multi-billion dollar blockchain project steps away to join an AI lab, builders should stop and pay attention. Sam Blackshear, the CTO of Mysten Labs and the architect behind the Move programming language, is moving to Anthropic. He isn't going there to build a new token or a decentralized exchange. He is going there to work on AI security.

This move isn't just another LinkedIn update. It is a signal of where the real technical war is being fought. For years, the smartest minds in cryptography and systems architecture were obsessed with securing distributed ledgers. Now, that gravity is shifting toward securing large language models and the agents that run on them.

The Security Balance has Shifted

In the crypto world, we have spent a decade dealing with the asymmetry of security. An attacker only needs to find one flaw in a smart contract to drain a protocol. The defender has to be perfect every single time. Blackshear understands this dynamic better than almost anyone. He helped build Move specifically to solve the inherent safety flaws in Solidity.

But AI introduces a different kind of asymmetry. We are moving from a world of deterministic code to a world of probabilistic logic. In a blockchain, 2+2 always equals 4. In a large language model, the output depends on the prompt, the weights, and the temperature. Securing a system that thinks for itself is infinitely harder than securing a system that follows a script.

Blackshear noted that AI is fundamentally changing the balance between attackers and defenders. Right now, attackers are using AI to find vulnerabilities faster than humans can patch them. If the defenders don't have better AI-driven tools, the entire digital infrastructure becomes a house of cards.

Why Move to Anthropic?

Anthropic has positioned itself as the "safety-first" AI company. While OpenAI and Google are racing for raw performance and market share, Anthropic has focused heavily on constitutional AI and alignment. For a systems builder like Blackshear, this is the logical next step. If you want to build a secure foundation for the next thirty years of computing, you don't do it on the blockchain anymore; you do it at the model level.

This doesn't mean Sui or Mysten Labs are in trouble. Blackshear is staying on as an advisor, and the Move language is mature enough to stand on its own. But it does show that the most pressing problems in tech have migrated. The "hard problems" are no longer about transaction throughput or consensus mechanisms. They are about ensuring that the models we are giving control of our lives don't become weapons in the hands of bad actors.

What This Means for Web3 Builders

If you are building in the crypto space, you need to realize that the fence between AI and Web3 has been torn down. We are seeing a brain drain of technical talent from pure-play crypto into AI safety and infrastructure. This isn't a betrayal of the industry; it is an evolution.

• Smart Contracts get smarter: Expect to see AI-driven formal verification tools that are far more advanced than what we have today.
• Identity is the new battlefield: As AI gets better at social engineering, the need for decentralized identity and proof-of-personhood becomes a survival requirement, not a luxury.
• Agentic Security: We are moving toward a world where security isn't a static firewall, but a dynamic agent that monitors network health in real-time.

Builders should be looking at how to integrate AI security protocols into their existing stacks. If your project relies on manual audits and bug bounties alone, you are already behind. The attackers are using LLMs to scan your bytecode for reentrancy attacks and logic errors. You need the same caliber of intelligence on your side.

The Skeptic's View

As a founder, I have to ask: Is this just another pivot to the current hype cycle? We saw it in 2021 with the Metaverse, and we saw it in 2017 with everything becoming a blockchain company. But this feels different. Blackshear isn't a marketing guy chasing a trend; he is a compiler and language expert. These people don't move unless there is a legitimate technical mountain to climb.

The skepticism shouldn't be about whether AI security is important—it clearly is. The skepticism should be about whether these centralized AI labs can actually solve the problem. Anthropic is a private company. If they build the ultimate security layer for AI, they own it. This is why the intersection of crypto and AI is so critical. We need the security of AI, but we need the decentralization and transparency of crypto to ensure that security isn't just a walled garden controlled by a few elites.

The real risk isn't that AI will become sentient and kill us; it's that human attackers will use AI to break every digital lock we've spent the last twenty years building.

The Founder's Takeaway

The takeaway for founders is clear: stop treating AI as a side feature. It is the new base layer of the stack. If your CTO isn't thinking about how AI shifts your threat model, you are sitting on a time bomb. Blackshear’s move to Anthropic is a wake-up call that the security landscape has permanently changed.

We are entering an era of "active defense." Static code is no longer enough. The systems we build must be as dynamic as the threats they face. For those of us still in the trenches of Web3, the mission is to make sure these AI-powered security tools stay open and accessible, rather than being locked behind corporate APIs.

Read the original at Cointelegraph →