Eric Weinstein

We had things like, what is it, Operation Starfish, where we exploded a weapon in space and knocked out the grid in Hawaii.

We retarded ourselves a great deal.

And then we signed the bioweapons conventions.

uh in the in the 70s which had a big effect on how we do research we have scared ourselves before into some kind of cooperation that does something but isn't the challenge with that and i think it's a very good example but nuclear weapons are mainly

Well, first of all, Russia doesn't agree with you.

Russia used nuclear weapons for engineering purposes.

They would use them to dig holes or put out fires or do all sorts of crazy things.

So even that is not quite true.

And then with respect to AI, things don't always spread just because they're known.

I keep giving this example, like if I ask people, do you know the formula for black powder?

No.

Do you?

75% saltpeter, 10% sulfur, 15% charcoal.

Okay, now everybody at home knows, but they'll forget it.

We in general don't see people taking advantage of these huge leverage opportunities.

Right now, we're in some race because we don't know what we're racing for.

We just know that if the other guy had something that we didn't have, that feels vulnerable.

But the problem is the other guy may become the AI itself.

So the US and China may get into a race to get this thing, and then the thing says, you dear sweet children.

Nobody even knows what we're talking about.