Jacob Coxon says this is why pacing the frontier might not work
Coxon said in an interview that he was "almost certain" Chinese spies had infiltrated US AI labs, making it difficult to keep AI advances from China.
TL;DR
- AI researcher Jacob Coxon believes slowing US AI development is ineffective.
- Coxon suggests Chinese spies may have infiltrated US AI labs (OpenAI and Anthropic).
- He believes China could "catch up the lead" if spies steal progress or if China develops AI independently.
- Distillation attacks, where advanced AI trains less advanced models, are another reason for China's AI progress.
- Coxon compared Anthropic staff to worker ants sacrificing themselves for superintelligence.
- This comes after calls from AI executives to slow development for safety, which some critics argue could benefit companies like Anthropic and OpenAI financially and disadvantage smaller rivals.