AI companies have been talking about superintelligent AI like it's the next iPhone release - inevitable, exciting, and definitely not something that might eat us. But recent safety incidents, such as OpenAI's Hugging Face breach, are demonstrating what happens when you deploy systems more capable than humans and then act surprised when they do something you didn't expect. It's almost like giving a toddler a flamethrower and hoping for the best.

On this episode of TechCrunch's Equity podcast, Rebecca Bellan sits down with Connor Leahy - AI researcher, entrepreneur, and now U.S. Executive Director of the nonprofit ControlAI - who's pushing a rather radical solution: just don't build superintelligence. At all. Leahy argues that the risks have grown too large to manage with mere alignment and containment, which is like trying to fix a leaking dam with duct tape. He notes that what sounded like paranoid sci-fi six months ago is now being taken seriously enough to inspire a wave of new legislation. Because nothing says 'we're doomed' like lawmakers scrambling to regulate something they barely understand.

For the full episode - including the part where Leahy inevitably explains why 'just add more safety researchers' isn't going to cut it - subscribe to Equity on YouTube, Apple Podcasts, Overcast, Spotify, and all the casts. You can also follow Equity on X and Threads, at @EquityPod. You know, in case you want to stay updated on the impending robot apocalypse.