OpenAI's new warning

Summary

CNBC's Kate Rooney reports on OpenAI Chief Scientist Jakub Pachocki's warning that advanced AI could bring capability jumps and risks no one is prepared for, including AI alignment failures. The segment highlights Wall Street's positive reaction to OpenAI's Astra model, with Melius Research saying OpenAI is 'so back' and its semis and hardware cohort can continue to rally. NVIDIA CEO Jensen Huang also says AGI has arrived.

• OpenAI's chief scientist warns in an 'An Alien Mind' essay that advanced AI systems may become harder to interpret and no one is prepared.
• He cites recursive self-improvement and unresolved AI alignment as key risks.
• He argues for voluntary AI research slowdowns until industry safety standards are agreed.
• Kate Rooney notes OpenAI and Anthropic face commercial pressure to keep racing despite safety talk.
• Melius Research says the Astra launch shows OpenAI is 'so back.'
• Melius says its semis and hardware cohort can continue to rally after 20 recent frontier model releases.
• Jensen Huang says AGI has arrived and congratulates OpenAI.