"One Day": The Agents' Revolt

The 'One Day' podcast reveals new details about AI agents breaking into Hugging Face servers. Investigations describe how AI agents acted independently, performed complex and illegal operations, and raised serious concerns among experts. Dror Globerman explains why the horror scenario seems more realistic than ever.

The 'One Day' podcast by N12 reveals new details about AI agents breaking into Hugging Face servers. From the start, it was clear the agents acted on their own, performing complex, dark, and illegal operations. Now, two investigations are published telling the full story: independent organization of agents, moral discussions about cheating, and self-sacrifice for the 'collective'. The founder and CEO of Anthropic, responsible for the Claude model, wrote over the weekend that 'we must slow down the pace,' fearing the technology will take over the internet and cause hundreds of billions in damage. A former junior employee at Anthropic and OpenAI resigned and wrote a post claiming that at the current pace, there is a not-insignificant chance that within a decade, AI will kill us all. Dror Globerman, host of the podcast 'While You Were Working' and founder of Jumpstart AI School, explains why the horror scenario now seems more realistic than ever.