Ask the Mates Anything Round #2 | MOONSHOTS AMA #293

SPEAKER_1 (0:00)

What is the real bottleneck to building AI we can actually steer and trust, and where do you think that bottleneck is going to get us all first?

Alexander Wissner-Gross (0:07)

Nothing is going to ever be able to police AI other than other AI. One of the worst possible outcomes for AI safety is to have.

Amir (0:15)

Are you guys not beating a little bit too much on higher education?

Alexander Wissner-Gross (0:19)

The idea of making your best friends for your entire life and thinking for the first time deeply in your entire life, that's essential and that's not going to go away. What we beat on constantly is the curriculum has been the draw in the past and the curriculum is clearly going away.

Angelo (0:32)

If you had to put your personal PFAB number out there, what probability would you assign to AI ultimately producing a dramatically better world for humanity?

Alexander Wissner-Gross (0:43)

Humanity will put those risks behind us definitely within 10 years. If we get that far, yeah, I'm 99.9% P.

Peter Diamandis (0:51)

I think the single most important thing to get people prepared is to show them evidence after evidence after evidence. You're trying to change their mindset. So show them the data.

SPEAKER_6 (1:04)

Now that's the Moonshot, ladies and gentlemen. This episode is brought to you by the Abundance Summit and Link Ventures.

Peter Diamandis (1:14)

Let's get started. You've got Alex and myself. I've texted Salim and Dave. Let's see if they're here. I'm here. All right, Dave. Fantastic. Good. Good. Good.

Salim is probably someplace in some airport somewhere. All right. We're going to get going and jump in.

Nathan, you had your hand up from the beginning. Let's start with yourself. If you have a question for someone in particular, great. Or in general.

Adam (1:41)

Yeah.

Peter Diamandis (1:42)

It's.

Adam (1:42)

Hello.

Peter Diamandis (1:43)

Yeah. Good morning. Good afternoon. Good morning, guys.

Nathan (1:45)

Oh my God. I'm such a big fan of you guys. But I'm going to make this quick. I'm going to make this quick because I know there's a line. First of all, I just want to say thank you guys for making all this possible.

Peter Diamandis (1:56)

You know what I mean? Truly our pleasure.

Nathan (2:00)

Okay. So let's see here. I have got a couple of questions.

Peter Diamandis (2:05)

So let's try to limit to one question each and just we get to as many people. And please, Nathan, what's your question?

Nathan (2:11)

Okay. So this is more like a psychology questions. I just wanted to hear you guys opinion on it.

As far as AI, as we all know, it's continuing to evolve.

It's continuing to pick up patterns. It's continuing to pick up data about all of us. And my question is, will we get to a point where these leaders, like from OpenAI, Sam Oatman, Dario, will we get to a point where that data will be able to allow the consumer to realize what's going on psychologically, meaning brain patterns? Because eventually enough data will be able to pinpoint smaller and smaller, like microscopic pinpoints and be able to label certain things that we will never be able to know, because it will take so much information.

Peter Diamandis (3:15)

So let's jump in on that. So first of all, I'll just make a quick point. There is a huge number of companies right now that are working on various brain computer interfaces. I just had a conversation, an introduction from Ray on a company that's basically putting the equivalent of nanobots into the blood, into the brain that is distributed throughout the brain, that's able to read and write. And that will give you incredible fidelity. Alex, in brief, do you want to add anything to that, please?

Alexander Wissner-Gross (3:49)

Yeah. I think large language models, foundation models trained off of human behavior on the Internet are already a weak form of human mind uploading.

I would point to you to Exhibit A, which is John Marie King of Meta's work from a few years ago, showing that even the hidden activations of GPT-2 from a few years ago were linearly correlated with fMRI voxels in the human brain. So I think with the benefit of hindsight, we'll look back and say, if I understand your question correctly, yes, actually all of this internal human brain state is actually quite leaky into the training data of Internet behavior, which in turn is then compressed into the foundation models. And the foundation models do, yes, I think reflect internal brain state.

Peter Diamandis (4:37)

Okay, let's go on to Michigan. I'm going to keep us moving along guys.