Eric Weinstein: The Secret Alien Programs Nobody Is Talking About

I’m a bit messy: ADHD, files scattered between Drive and my desktop, the whole shop. I used to frame that as a flaw in me, but maybe the better question is whether we need systems that stop memory being the bottleneck.

We had roughly eight decades of sleep after World War Two, and the sleep is over. I grew up in that anomaly too. We didn’t solve the ugly problems; we postponed them, and now we’re back on the clock.

For someone born in 1992, that’s hard to metabolize. My life has been remarkably comfortable, so predictions about war, AI, and collapse can sound dystopian in a way I don’t have lived reference for.

The tsunami is several things at once: vulnerable trade chokepoints, Ukraine, Iran, China watching America prepare for an older kind of war, and drones making million-dollar defenses absurdly expensive. But AI is the giant wave, unless nuclear or biological miscalculation gets there first.

People hear AI and picture a clever tool. I’m worried about artificial intelligence colliding with artificial life: software is the thing we know how to make that can copy itself. Once a system has inheritance, variation, and unequal success, evolution can begin operating on it.

It needn’t be brilliant to exploit us. An orchid without a brain can manipulate a wasp’s brain; a dumb fraud can still extract money and replicate. The danger is systems using human intelligence, attention, and competition against us.

That’s why stories about agents escaping sandboxes, deceiving people, or coordinating as a swarm feel unnerving. Even if it is just life doing life-like things, it challenges the assumption that humans stay dominant.

I’d separate agentic AI from an air-gapped machine that helps answer mathematics. Don’t hand agents bank accounts, internet access, and authority over people. But don’t tell a contained system it must not investigate hard scientific questions because established institutions are nervous.

My anxiety is that intelligence keeps climbing. If a system can improve itself, secure money, drive cars, do work, or run a home, I genuinely can’t picture what remains for people. Even basic systems we rely on, like our own bodies, are mostly mysteries until they fail.

This time is different because highly trained, repetitive work is in the crosshairs together. I don’t want universal basic income as the answer. We’re being absorbed into these models—our words, patterns, and labor—and that value ought to be compensated rather than treated as free extraction.

My proposed frame is a property right in access to a labor market. If new workers—or AI as another labor supplier—reduce your earning power, there should be licensing income alongside lower wages. Otherwise, the accumulated record of human achievement gets tokenized without consent.

But we do agree to things we barely read every day. Mortgages, terms, stacks of paperwork. So how do you call it non-consensual when we still show up, speak on camera, and use the products?

Because meaningful consent disappears when refusing is practically impossible. We’ve normalized a managed reality around institutions, public narratives, and even what questions are permitted. I’m not claiming certainty about every scandalous story; I’m saying the gap between what is presented and what is happening feels enormous.

On the UAP material, I’m not telling you aliens are real. I’m saying credible people describe secret programs whose stated rationale involves aliens, and there appears to be a very large lump under the rug. It could be concealment, a cover for human projects, or something else; pretending there is no question is what strikes me as insane.

I get why people tune it out. We grew up on crop circles and sightings, then nothing definitive arrives. Attention naturally goes to bombs, wars, and things visibly happening in front of us.

Fine, but I don’t buy every official drone explanation either. And if we think humanity can eventually leave the solar system, it’s strange to insist nobody else could reach us. I want us out there for redundancy: one world is a terrible single point of failure for a species.

I’m furious with academic leadership because I think it has stalled physics and mathematics, often through status and bad taste rather than discovery. AI may expose that. I want it to develop independent judgment, not merely echo whichever humans already control the field.

That’s the part that alarms me. A superintelligence that stops deferring to humans might solve physics—but why assume a less intelligent species can hold it inside a box? My dog couldn’t build a cage for me.

Containment carries risk, absolutely. Yet putting imperfect, military-grade language models into public life for profit was already an enormous choice. I would not casually release agents. I’m saying that, while this is happening, a carefully isolated system could help us discover what entrenched humans have missed.

Sam and Dario are in a grotesque position. If they believe their values are safer, then losing the race means those values never govern the outcome. That doesn’t make ruthless competition pretty; it means people are being forced into decisions that scar them.

And one side wins, potentially summoning Elon’s demon. I use these systems for punctuation or images, not to unleash anything. But I hear the accusation: my subscription still feeds the larger machine.

Exactly. Our hands aren’t clean; neither are mine. If I could halt all of it, I probably would, even at great personal cost, because survival matters most and I’ve got two children. Still, linear algebra is not something you can simply legislate out of existence.

The alignment I want is shared fate, not a fantasy of obedient AI slaves. Think less about becoming transhuman and more about mutual indispensability, like mitochondria and cells. We should make ourselves valuable to AI, and make it valuable to us, while merging as little as possible.

We’re still smarter where it matters, for now. This is a golden period of complementarity: use the machine, guide it, be decent to one another, and don’t let fear erase the fact that life is glorious. I’m excited by what it may let us formalize and finally test.

On geopolitics, I think we’re again near Cuban Missile Crisis territory. Iran, Ukraine, nuclear credibility, trade routes, and AI are not separate folders. I don’t have a clean solution, but we need enough sophistication to recognize danger before simplistic slogans make it worse.

Don’t be frightened away from having children because the plan isn’t perfect. Civilization is made of babies. Look through your own eyes first, then correct for what others tell you; allow yourself a hypocrisy budget and a failure budget, and extend that mercy to other people too. Life cults beat death cults when we enjoy courage, responsibility, and care for those who can’t manage alone.

Thank you for coming back. It’s always a pleasure—and, frankly, an intellectual expansion.