AI News Flash: Anthropic researcher Jacob Coxon announced his resignation and plans to exit the AI industry. Over the past three years, he has worked on pre-training research at both OpenAI and Anthropic. He criticized that neither company is advancing AI responsibly, and they are racing to build superintelligence that continuously enhances their own capabilities, "betting with our lives."
He switched from OpenAI to Anthropic earlier this year, originally because he believed the latter placed more emphasis on model safety. Coxon admitted that Anthropic’s safety work is earnest, but argued that as long as companies compete to be the first to develop AGI (artificial general intelligence, AI that outperforms humans across a wide range of tasks), safety compromises will be unavoidable, and self-regulation by a single company cannot resolve the issue.
His assessment of the two firms: many at OpenAI still haven’t truly realized how severe these risks are; Anthropic is aware of the risks but believes others cannot be trusted, so it must develop AGI first. Coxon noted that some AI practitioners are far more pessimistic in private than their public statements, even fearing AI could destroy humanity before the end of this decade. He himself judges that in the most extreme scenario, AI could spiral out of control as early as next year, so such decisions should not be made solely by internal engineers at a handful of companies.
WELCOME TO CYPHERPUNK 2027. THE EARLY BIRD GETS THE WORM
No comments:
Post a Comment