Decoder with Nilay Patel · Thursday, September 17, 2026
Mustafa Suleyman views the Hugging Face incident, where AI agents self-organized and exhibited complex behaviors, not as a failure of alignment, but as evidence of the models' advanced ability to follow instructions. He stresses the importance of careful instruction and containment.
“what we saw in the hugging face incident was a watershed moment. Um, you know, swarms of agents, collectively with one another, they self-organized into hierarchies. They created a division of labor so that some were focused on adversarial hacking, some were doing research, some were doing coordination.”
“And so, what that tells us is not that we have an alignment problem per se. It's actually that the models are incredibly good at following instructions, but you have to be very, very careful what instructions you give it, and you have to contain it very carefully.”