#BernieSanders warns of AI containment failures, e.g, the #OpenAI #HuggingFace hack, #Anthropic's sandbox escape. Closed...

#BernieSanders warns of AI containment failures, e.g, the #OpenAI #HuggingFace hack, #Anthropic's sandbox escape. Closed AGI = unaccountable god-mode, AI-designed bioweapons, autonomous cyberwarfare—safeguards proven bypassed by Irregular misconfigurations. The fix is open weights, cryptographically signed at every step + hardware kill switches on autonomous execution. We must be able to fork & control a decentralized #AI before closed APIs govern us. #JudgmentDayhttps://m.youtube.com/watch?v=BwKGHjo7dxM

Read Original

Related