Global exchanges left on the sidelines are moving quickly to address the AI security gap that emerged after Coinbase secured exclusive rights to Anthropic's powerhouse AI model, 'Claude Mythos'! Rather than sitting idly by, exchanges like Binance are building a united front to exert pressure while simultaneously finding their own workarounds.

Binance's Chief Security Officer (CSO) Jimmy Su announced that while they are uniting with other excluded platforms, they are also revamping their entire security architecture into a 'Breach-Assumed' system. A particularly notable workaround involves deploying offline open-source models, such as Meta's Llama, directly into local environments to conduct security analysis. This strategy allows them to run a high-level self-defense system without exposing sensitive internal code to external commercial AI APIs. OKX is also fortifying its system-layer defenses by introducing trade simulations for user protection within its 'On-chain OS'.

It looked like the control and monopoly of massive AI companies would dominate the market, but it’s actually leading to an interesting trend that is stimulating the development of local open-source multi-agent architectures across the entire ecosystem. It’s going to be really exciting to see if this open-source coalition can neutralize the monopoly of Claude Mythos.

Related Links