After years of a relentless race towards bigger models, more capable agents and increasingly autonomous systems, some of the biggest players in the industry are now talking seriously about safety, guardrails and whether frontier AI development needs to slow down.
There are already some pretty compelling reasons for that conversation. High-profile incidents involving autonomous AI agents have shown how quickly systems can find their way around safeguards, while the extraordinary cost of training frontier models is raising increasingly practical questions about just how fast the industry can keep moving. And then there is the geopolitical problem: if the US slows down while China and other competitors continue developing and releasing increasingly capable models, can anyone really afford to hit the brakes?
But Episode 20 of The Signal is also about a much more immediate question: what happens when the thing we are trying to control starts controlling how we think?
Fergus Bell and Tom Trewinnard explore the risk of cognitive atrophy, the growing problem of AI-generated “slop grenades” at work, and why producing a 20-page document in five minutes is not necessarily productive.
Listen, subscribe and follow The Signal wherever you get your podcasts.
","author_name":"Fergus Bell, Tom Trewinnard"}