Crypto Firms Left Behind in AI Arms Race as Frontier Models Remain Out of Reach

Crypto exchanges and companies are eager to access powerful new AI models to strengthen their code against attacks, but only a select few have gained access so far. Coinbase secured access to Anthropic's restricted Mythos model in June, while Zcash used it to audit its protocol at the request of Shielded Labs. However, other major players like Binance are still waiting for access.

Binance's chief security officer Jimmy Su said that his company has been trying to get access to frontier AI models but hasn't succeeded yet. He noted that Anthropic's controlled rollout is a responsible approach because newly released models may benefit attackers faster than defenders.

However, as publicly available models become more powerful, the pressure will be on model developers to make them more widely available. Solana Foundation chief information security officer Michael Coates supported safeguards but argued that legitimate defenders need a faster route to these models. He said that '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.'

Blockchain Capital's Sean Cheetham also supported eventually opening up restrictions, saying that broader availability could ultimately favor defenders as legitimate security researchers outnumber sophisticated attackers.