E188 - Emad Mostaque: AI Expert: “We Have 800 Days Left”

**SPEAKER_1** (0:00)
We're heading towards a point where human cognitive labor is effectively negative. Not zero, but even negative.

**SPEAKER_2** (0:05)
How many years that could actually be far out? Like 10 years most?

**SPEAKER_1** (0:08)
Um, 100 days as of today. We built machines that can substitute for muscles. Now we're building some machines that can substitute for brains. Where do we have left to turn? Your job will be able to be replicated every single email you've sent, every single thing you've worked on, and it'll be just like you, except for it won't make mistakes, it won't go to sleep, and it's tax deductible. You'll be like, hey, there's Bob. It's a frickin GPU, not Bob anymore. It's a little bit creepy.

**SPEAKER_2** (0:32)
But what happens to the human meaning when so much of our perceived worth is derived from our work?

**SPEAKER_1** (0:37)
Well, the last few human-only breakthroughs are literally in the next few years, and that's it. The current race is a race of power to create machine god that stops all other gods and somehow works for you. I mean, this might be the great filter. Maybe all societies get to this point and they don't get past it. We're only going two ways, complete destruction or we're going to the world of abundance. It's a coin toss right now. But humans are good enough. There's no problem in the universe that we can't solve if we ordinate with common positive stories. This is the last chance. Something else comes, but not what we knew before.

**SPEAKER_2** (1:17)
Hey, everyone, welcome back to the Know Thyself podcast. Our guest today is a mathematician, a former hedge fund manager and founder of Stability AI, the company behind Stable Diffusion, and now building the Intelligent Internet. He is the author of The Last Economy, which states now, less than 1,000 days from now, because of the developments in AI, human cognitive labor is going to become economically irrelevant. This has immense implications on every aspect of life. And so we're going to be exploring that and more, such as what happens to human meaning, consciousness, and much more. Emad Mostaque, thank you for being here.

**SPEAKER_1** (1:55)
Thank you for having me.

**SPEAKER_2** (1:56)
You said you have been in rooms where the mathematics of human obsolescence has been calculated. That's a trippy statement. Can you unpack that a little bit for me?

**SPEAKER_1** (2:08)
Yeah, I think we've all seen the gerative AI wave. So when I wrote the book, it was a thousand days since ChatGPT was released. And that was the end of last year. And now I said, within a thousand days, 800 as of today, we're heading towards a point whereby human cognitive labor is effectively negative. We've just reached a point at the start of 2026, when you actually have actually competent intelligence. Like the first ChatGPT, where it would hallucinate, it would say all sorts of weird things, to now, when you have Claude and Claubot or OpenClau and other things like that, it's taking over the world. And in these rooms, the discussions are usually, well, how can we increase profits? How can we have more access as opposed to what really happens to the people? Like, what really happens to the people is usually a question of, how do we stop them revolting as opposed to how do we uplift them effectively? Because most discussions around AGI in particular, artificial general intelligence, the most the big labs are pushing towards, were around, how can we get the most power and control first to make sure nobody else gets it? So they've been very interesting discussions, but very scary ones, to say the least.

**SPEAKER_2** (3:19)
Yeah, I'm curious to hear how you think, just pure competition is driving these considerations as opposed to some sort of nefarious act. Occam's Razor would refer to the former. I'm curious what, to just back up a little bit, I would love to hear a bit of actually how the diagnosis of your son with autism came into the picture, you getting into AI, setting the stage for a bit of your introduction into this world, and then we can dive into the many topics we will. So yeah, what happened there?

**SPEAKER_1** (3:50)
Yeah, so I was a hedge fund manager at 23, I'm managing giant funds, and then my son was diagnosed with autism 15 years ago. And I was like, oh, there's no treatment, there's no cure, what can we do? So then I built an AI team, analyzed all the literature on autism, tried to figure out how it actually works, because people with autism, ASD, Alzheimer's, all sorts of other conditions, you're always told, well, you're not really sure how it works. And that's not really good enough. You want to always think from first principles. So first principles, what caused it? So we looked at all the literature, analyzed it, and then found 18 different things that express in a very similar way, just like the common cold. And it came down to the balance between GABA and glutamate in the brain. So GABA, when you pop a valium, it chills you out. Glutamate is what's excitatory, like when you can't focus and you're tapping your legs. Children, adults with autism have an imbalance of that that causes their brains to be firing all the time, so they can't focus because there's too much noise. So you then have to figure out how do you actually reduce the noise. You have things like calcium ion channels in the brain and others that you can repurpose drugs for to enable that decrease in noise and then applied behavioral analysis and others that can reconstruct how you talk, because my son wasn't talking, can reconstruct how you think once you've calmed down the brain, just like we have coping mechanisms for when we're anxious. And he ended up going to mainstream school, which was fantastic.