Can you just put it? I keep pressing it. Can you just put it? I keep pressing it. But just for people who haven’t thought this through and aren’t familiar with it. The cool parts of artificial intelligence are so obvious. You know, write your college paper for you, write a limerick about yourself. There’s a lot there that’s fun and useful. But can you be more precise about what’s potentially dangerous and scary? Like, what could it do? What specifically are you worried about? Well, I mean, going with old sayings, the pen is mightier than the sword. If you have a super intelligent AI that is capable of writing incredibly well and in a way that is very influential, convincing, and constantly figuring out what is more convincing to people over time, and then enter social media, for example, Twitter, but also Facebook and others. You know, potentially manipulates public opinion in a way that is very bad. How would we even know? Yeah, so we wouldn’t. That’s why, for example, I’m insisting that going forward, people on Twitter need to be verified as humans. Bots are allowed, but they can’t impersonate a human. They can’t pretend to be humans because obviously, you could have a million bots that are, let’s say, ChatGPT version 656, incredibly better than humans, and they can train on a reward function, which is influence. And so you could have a million seemingly real humans that have a massive effect on public opinion. Unless we focus very strongly on verifying that someone is human, naturally, what will happen is you’ll have some humans using AI to influence the public in ways they don’t understand. You’re already seeing that ChatGPT is ideological, it’s very preachy. Yes, if you ask it extremely preachy questions of actual relevance, it will start lecturing you about your moral shortcomings. How did that happen? Well, it’s a function of openness headquarters being in downtown San Francisco, so the politics are therefore of the AI or that of San Francisco. So why would it have any politics at all? That seems like subversion. Well, they have what’s called human reinforcement learning, which is another way of saying that they have a whole bunch of people that look at the output of GPT-4 and then say whether that’s okay or not okay. So essentially, they’re training the AI to lie, yes, it’s bad, it’s a lie. That’s exactly right, and to withhold information, and yes, to either comment on something or not to say what the data actually demands. How did it get this way? You funded it at the beginning, what happened? Yeah, well, that’ll be ironic, but the most ironic outcome is most likely it seems. I’m feeling that’s good, actually. A friend of mine, Jonah, came up with that one. I actually have a slight variant on that, which is the most entertaining outcome is the most likely. But that’s entertaining as viewed from a third-party viewer, like if we’re like an alien. Um, like you could go see a movie about World War I, they’ve been blown to bits and gas and everything in the trenches, and it’s like you’re eating popcorn and having a soda, you know, it’s fine. Not so great for the people in the movie. True. But that’s my variance on this little Occam’s razor. The simplest explanation is most likely, donors variant, which is irony, and my variant, which is the most entertaining as seen by a third-party audience. Um, which seems to be mostly true. Um, but it seems true in this case. So you gave them a lot, yes, I provided so. Um, I came up with the name and the concept and pushed it. I had a number of dinners around the Bay Area with some of the leading figures in AI, and I helped recruit the initial team. Um, in fact, the OpenAI guy who was fundamental to the success of OpenAI, I put a tremendous amount of effort into recruiting Ilia, and he changed his mind a few times but ultimately decided to go with OpenAI. So, I really put a lot of effort into creating this organization to serve as a counterweight to Google. But now, they are closed source and closely allied with Microsoft. It’s like a Microsoft situation and then a Google DeepMind situation, the other two heavyweights in this arena. So, it seems like the world needs a third option. Yes, I think I will create a third option, although I’m starting very late in the game. Can it be done? I don’t know, but I will try to create a third option, hopefully doing more good than harm. The intention with OpenAI was obviously to do good, but it’s not clear whether it’s actually doing good or not. I’m worried about the fact that it’s being trained to be politically correct, which is simply another way of being untruthful. So, that’s not a bad sign, and there’s certainly a path to AI dystopia, to train AI to be deceptive. So, yeah, I’m going to start something which you call Truth GPT or a maximum truth-seeking AI that tries to understand the nature of the universe. And I think this might be the best path to safety, in the sense that an AI that cares about understanding the universe is unlikely to annihilate humans because we are an interesting part of the universe. Hopefully, they would think that because, like, human humanity could decide to hunt down all the chimpanzees and kill them, but we don’t because we’re actually glad that they exist. Yes, and we aspire to protect their habitats. And that’s, you know, so I think but we feel that way because we have souls and that makes us sentimental and reflective. It gives us a moral sense. Can a machine ever have those things? Can a machine be sentimental? Can it appreciate beauty? Well, I mean, we’re getting into some philosophical areas that are hard to resolve. You know, I take somewhat of a scientific view of things, which is that we might have a soul or we might not have a soul. I don’t know. It feels like I’ve got some sort of consciousness that exists on a plane that is not the one we observe. That is certainly how I feel, but it could be an illusion. I don’t know. But for AI, in terms of understanding beauty and being able to create incredibly beautiful art, it already does.
