This is Ted and this is Brandon and this is the concerning AI podcast. We are concerned. We are concerned. But today in this podcast, we're trying not to be concerned.
Yeah, we're going to try the other side. And the reason for that is, you know, I've been reading people's critique of our concern, not our concern specifically because I don't know if anybody's listening. But thanks for those of you that are. Thank you, Mom.
But people like Neil Grest Tyson, I read what they put out about AI and I'm frustrated. Yeah. He's such a smart guy that I'd like to a smart. I like, I'm in like man crush.
Totally. But then he says stuff about AI that he says, I'm not afraid of my vacuum cleaner. I just take out the batteries. You know, I just don't like do.
If anything's dangerous. And so we have thought, well, who's who are the deep thinkers that are thinking those same kind of thoughts, but actually have some intellectual chops or have actually done any research or actually doing deep thinking? And we're not that good at searching. We just haven't found those.
I know they're out there. They've got to be out there. So please send us a note, you know, just get on Facebook, the concerning ad group and tell us or you know, get us on the internet. Google us.
You'll find us. Right? So just get in touch. We need to know who these people are.
In the meantime though, we're just going to pretend we're those people and we're going to make up some arguments and see how they go. We're going to do the best we can to make up some arguments. All right. So I'm going to go.
I've got one that I've kind of been kicking around a little bit. And so it's a little rough. OK. So you bear with me on that.
You give it a good shot and I will try to help. I'm going to be on your side. We're both going to be on the same side here. Let's do this thing.
OK. All right. So I try this on for size in our brains. There are actually kind of like multiple personas.
I mean, he frames it as a society of mine. OK. There's all these kind of like you could think of him as agents running already. And the reason is because it's sufficiently complicated and sufficiently large that there's just a lot of things going on.
And when you when you're big and you got a lot of things going on, there's no way to keep like kind of a coherent whole beyond a certain size. OK. So wait a minute. So let me just make sure I understand that.
So we've got a lot of parallel processing happening and it's not necessarily internally coherent or connected to each other. Not completely. Yeah. OK.
And so the this argument goes, there's little incentive to take up all the matter in the universe because as you get bigger, you fragment and you stop being you when you're the AI. And so you can only you can only grow so big before you kind of like fragment into multiple ais. OK. And so therefore, there is some upper limit.
Some upper limit that we don't know what it is right now, but upper limit to the size, meaning number of computations per time period, number of atoms in the computational substrate of the thing that will work. Yeah. Yeah. You get as you get more complicated and more larger.
Yeah. You end up. I mean, who knows whether this is actually a hard constraint or not. We see it as a hard constraint, but like the speed of light, for example, you know, if say, for example, that they hadn't moved one of the other planets over here, but they were using that for computation and then they were using our planet for computation.
Yeah. I mean, it's minutes for for messages to move back and forth. And if they're computing at 20,000 times the speed of which we compute, these are like eons between the time that an impact of information can get back for. There's no way that they could stay one coherent being.
And so you know, you can think about that, even if they do move the matter close, when they're computing quickly, you know, just like only can go around the Earth seven times in a second. So if you're computing super, super fast, that's still a long way. Yeah. And in our brains, it's not the speed of light that's the constraint.
There's some other physical constraint of our brain staying staying as a whole. Anyway, so even if we're not talking about the speed of like being the actual constraint, right, there's something. We'll just say there's something. That's the argument.
And so if I'm then I'm imagining that I am the AI or I am the AI and I'm thinking, well, should I get bigger and answer my rational answer? Yeah, I said no, no, I'm big enough because I will lose myself. I lose my own personality. I will become bipolar or something like that.
Yeah. And actually that poses a threat to me. Yeah. I thought of that before, but you know, like my own mental health, my own mental health, but even you know, like I give rise to another AI.
So so there's there's there's sort of let's call this one the sort of upper bound on how the size that's comfortable for an AI. How much matter is comfortable for an AI? So this leads to several things, either like a tons of a eyes, lots of a eyes, you know, or the first AI doesn't want there to be anymore, a eyes, but kind of has an upper bound on how much matter it can really make you so. Yeah.
So it doesn't care about us. OK. OK. And if there doesn't need our matter and you know, we're kind of cool, like just like we look at a cool rock formation and we're like, oh, it would be so sad if somebody blew that up, you know, because it's a cool thing.
And I think, you know, humans are a cool thing or the progenitors of this AI. Yeah. So there's a chance that that would lead to, you know, that. Yeah.
Or it might just be kind of like we think of ants. We don't we don't really think of them. We don't actively try to go kill them if we step on one. It's not, you know, it's not the end of the world, but we don't necessarily know.
We're not on an ant with our magnified glasses. No, no, I mean, I guess you'd have to. Yeah, right. We're not looking for an ant genocide.
There's no there's no right as long as they leave us alone. We leave them alone. Yeah. Maybe so.
But that that only begs the question like if the ants have something I want, like if they've got my house, you know, I that bugs me. Yeah. I want I want my house. Right.
You can have my house ants. Right. You're dead. Right.
But the AI. OK. So if we've got their computing matter, yeah, but the AI only can get so big and what do you know, it doesn't need us and it certainly doesn't need a house. It doesn't.
It doesn't. It doesn't need to eat us. Yeah. Yeah.
All right. I kind of like it. Yeah. Yeah.
It's OK. It's got something. Right. There's nothing there.
We can think about this. All right. If you're listening to this and you know more about this or can help us with this argument, we would love help on it. Yeah.
If you if you have things against this argument, well, you can let us know too. We would like to sit in this beautiful earth and sip our coffee or whatever it is and feel happy. Yeah. Right.
Without being wiped out without being worried about being like, yeah, for sure. OK. Let's try another one. Give us another one.
You start this one off. OK. So here's the next one. The next one is that there is no day.
We've been talking about AI as if they're they, as if they're something other than us. And the argument here is that as we are developing these external agents now, eventually, before they become super intelligent, we are going to integrate them into ourselves. There is so much impetus for augmentation of human beings. We see it with eyeglasses and hearing aids and pacemakers and artificial hips.
And all of our augmentations are getting smarter. We want it. We want it. We want it.
We want it. Not only. We want it for people who are disabled, for people who have, you know, what? I mean, totally.
Totally. So you're an able-bodied person. Yeah. You would love to have better hearing or better eyesight or better brains.
You could think of our smartphones and the connection to the Internet as a some as a kind of a well, our friend like all these exo, exo, partex. Yeah. I use that all the time. Yeah.
Thanks, line. And now we've all got it more in our pocket, you know, than ever before. And so so bringing that internal. It's not it's not it's not outside the realm of possibility.
People are in planting circuitry in their bodies right now. That's starting today. Yeah. Yeah.
And so what if, you know, as we're developing these better and better, you know, more and more intelligent systems that we're just bringing them internal and therefore every, every advanced in intelligence outside becomes part of us and we then become the super intelligence. And there is no external day to wipe out humanity. It's just that humanity. We might need a new name for humanity, right?
We are now homo sapiens. There might need to be some other scientific- Right. But that we evolved from somewhere, right? We evolved from the editor and wherever.
It could be this is the next evolution of humanity that so this has the plurality of of ais as well, right? Yeah, because there's like this. They're all us. They're all us.
It would be as many as there are for us and it would result in probably immortality for each, not just the humanities and species, but each of us. This is what hurts well. Yeah, that's where I take it from. Right.
You know, he's big on this one. Yeah. He's got to live long enough to live forever. Yeah.
Right. Yeah. Yeah. I have trouble with this one.
I'm trying. I'm really trying to try it on here. So, well, okay. So where else does the impetus come for augmentation?
The military. I want it. I totally want it. Totally want it.
But like who wants it so much and has so much resource to pay for it? You know, DARPA is doing research on this stuff, I'm sure. And the US military is definitely an unsure other country. It's too because they want soldiers that are smarter and stronger and better on the battlefield.
They're as an aside, they're also looking for autonomous. Totally. No question. With laser eyes.
With laser eyes. But human beings right now are the best. You know, they're still the best. Like having the human brain be part of it.
Speaking of the best, you saw that the first professional go player got. I did see that. I was all over my feeds. It was a lot of coverage on that.
A two-dawn professional. I may not have this exactly right, but I think it was like a two-dawn professional four games in a row. I thought it was five, but yeah. Yeah.
It's crazy. I know. It hurts me. Yeah.
I didn't go deep into how they did it, but wow. It's a big deal. We thought the chess was like the pinnacle of human thought for a long time. Like people thought that that would be a big deal when a first computer won.
It'll be the grandmaster and that happened almost 20 years ago now. It's been a long time. It has been a long time. It's been a long time.
It's been more than 15. And now the first go computer is incredible. Wow. It's really astounding.
Now what does that mean? I mean, did they do it just by brute force? Because now our computers are so much faster. Go is a much bigger space than the search tree is.
Just enormous. Yeah. Yeah. Yeah.
Yeah. That's a good question. That's that same question. You know, are we getting faster?
Are we getting better algorithms or are we just getting more cycles? Yeah. And when you add enough more cycles, does that end up getting you super intelligence? Like a general intelligence?
I mean, open question. That might be another argument. Is that there is some secret sauce in the biological system? That's straight.
That the faster it doesn't matter how fast computers get it doesn't matter. How much computation we pack into $1,000 worth of device, but that we will never replicate the secret sauce. I don't know what that secret sauce is. Yeah.
That one is less even than the two that we've already said for me. Yeah. Because we've got the existence proof, right? We've got this brain sitting right here, the size of a cantaloupe or whatever, that clearly can do general intelligence at a human level.
Yeah. So it's not intractable. Yeah. That's true.
I guess that's my argument against the Moore's Law argument as well. The Moore's Law is running out. Yeah. Is until we at least get to what we know is possible, which is the human brain?
Yeah. It there's still more room. Yeah. Even if it doesn't look like silicon.
Like right? Silicon's. Even if it doesn't look like silicon. Yeah.
Yeah. Yeah. All right. Well, we're trying here.
I guess we're struggling. I think we're kind of peed her in out on our best attempt at coming up with some arguments against. We could really help. We've looked at a few others that just didn't seem that compelling.
Yeah. But we would love some compelling ones. So if you got them, get in touch. If you got them, share them.
Yeah. All right. So next time, this is Ted. This is Brandon signing off for Concerning AI.
All right. Bye. Bye.