Transcript:

Mike Slater: Want to go to a very smart person, Dr. Brian Keating. He is an experimental, I want to say cosmetologist, but that’s because I’m not a smart person. He’s a cosmologist at UC San Diego and he’s a host of the show Into the Impossible with Brian Keating. Dr. Keating, how do you do, sir?

Brian Keating: Very well. It’s good to be with you. And you could say cosmetology. You look very handsome and you have good hair. And that’s what cosmetologists do. But actually, yeah, they do have [inaudible]

Mike Slater: I was going to say, what percentage of people call you a cosmetologist?

Brian Keating: More than call me an astrologer, which I’m also not. [laughter]

AI is like the Tower of Babel

Mike Slater: Okay, fair enough. Well, let’s chat about AI. You made a reference earlier that AI is like the Tower of Babel. Say more. What are their similarities?

Brian Keating: Well, you know what’s interesting? In the Torah and the Old Testament, in the Bible, the Tower of Babel is a very famous story, but it’s always mentioned without mentioning that in the actual Hebrew and the original, it’s actually always the city and the tower of Babel. In other words, there’s an association between the technology and the inhabitants and the population, both the population that creates it, that desires it, that seeks to actually, as it says, make war with God, to be as one with one voice, etc. And the city, the inhabitants, and the Bible is quite skeptical of cities. There are good things that come from cities, but there are a lot of bad things that come from cities. And I think one of them is this sensation that you can sort of divorce yourself from the rest of the world, the real people in sort of flyover country. I always hate that term, but living here in the coast in San Diego, we have a sort of inflated ego about our importance.

And I think that’s sort of what’s happening. We see that with the creation of artificial intelligence, which is both, as also as the Bible says, a double-edged sword. So I often joke, where are the single-edged swords that only do good? But in this case, it’s a double-edged sword. It can do good. It can do wonderful things. It can discover drugs. It can help your kids with their homework. But it can do awful things, too. It can create superviruses. It can create incredible weapons and technology that could be used for destruction. And it could teach your kids to fall in love with an AI avatar that doesn’t exist, in some cases allegedly be contributing to their propensity for suicide. So how do we monitor this as parents, as members of cities, of society in America, when we can’t divorce ourselves from the other countries that are also striving to create this technology?

They’ve become like gods to us

Brian Keating: And what I see now is a very terrifying juxtaposition of these two forces, the two sides of the sword, the two edges, in that you have the creators, the progenitors, the ones that are responsible for that. And by the way, they have these interesting names like Sam Altman, like an alternative man, and then Dario Amodei, which means lover of God or love of God, right? Amo, like amateur. Dei means God in Latin. And then Elon Musk, I don’t know, maybe he’s got a nice smell or something like that. But there are these names that are associated, and they’ve become like gods to us, right? They can do superhuman things. It used to be if you had a question, who do you go to? You go to your priest, you go to your rabbi, you go to your minister, you go to your preacher. Now you tell things that you would never tell to them to the AIs, to get answers, to get the hive mind, the brain, the global brain working, coordinating with the sum total of human knowledge, which is both beautiful and it’s terrifying.

And these models have gotten incredibly powerful, and there’s no slowing them down. And the leaders, the Amodeis, the Altmans, they’re now saying that we need to slow down our own technology because it’s so dangerous, and you need to regulate us, and by the way, you may need to take us over. And I think that is the ultimate goal. They want the government to subsidize them. They want the government to be responsible to shepherd them as they play God themselves. And I think that’s quite dangerous, unprecedented. And I’m quite excited, but I’m also worried as somebody who uses this every day for his teaching, for his own kids. I use it. I thrive on it. But you have to be very careful with it. It’s a very powerful technology.

I’m an accelerationist

Mike Slater: I agree with everything you said. Now to go back to our Bible, the point we made earlier was that they made bricks. So this was a technological advancement, like, wow, we’re now able to make these bricks and this bitumen and stuff. We’re able to build taller than ever before. I love your city point as well. But it’s like anything with these double-edged swords, right? Who was it who said that fire can be a good master but a terrible slave, or however that goes, right? And you can use a knife to cut your food and you can use a knife to murder someone. So, as a scientist, how do you think about technological progress? How do you think about a scientific advancement and weigh the good and the bad? From a scientific mind, how do you do that?

Brian Keating: Yeah, purely scientifically, I’m an accelerationist. I would love to see nothing more than the solutions to age-old problems that I’ve wondered about, simply because it would allow me to do something that’s uniquely human, which is to have my curiosity stoked to investigate the questions that then I prompt the artificial intelligence with. I think that’s the healthiest, the safest way to use it. Use it as a curiosity amplifier.

I have graduate students. They’re made of flesh and bone. They’re the smartest people I’ll ever know. And I’m entrusted with them by their parents, by the university, etc. And we have deep relationships. I go to their weddings, I counsel them, and they’re wonderful. They’re almost like children, but of course it’s separate from my actual children. But I feel responsible for them. I get to interact with them and stimulate their young minds to become the future leaders of science. I worry about them in the age of AI, but currently I’m trying to make them AI native.

There’s a war in the university

Brian Keating: Right now, there’s a battle in the university. We start teaching today, classes open at UC San Diego, and there’s a war between the professors that know how to use AI, that want to use AI, and the students who want to use it, who use it for good. And then there’s also the double-edged sword. Again, you cannot escape this analogy. The double-edged sword is that they use it to do their homework, to solve their problems, to bypass the thought process themselves.

When I was a kid in the ’90s in college, we had this cartoon in our student newspaper, and it was first day of class, professor standing up on stage, a room full of students staring off at the professor in wide-eyed amazement. And then the next day, the students get replaced by tape recorders. Remember those? I don’t know. You might be young. You might not remember tape recorders. Tape recorders recording the professor speaking. And then the last scene is now the professor is replaced by tape recorder playing the lecture to the students, who are listening on their tape recorder. It’s crazy. Now we have exactly that. My students don’t use Instagram in class anymore. They’re using AI.

It could be great. It could enhance their understanding. I’m not the best educator. I often want to take classes with Albert Einstein. I’m too young. I can’t do that, unfortunately. But I have all of his collected works. I can create [inaudible], an artificial Einstein. That’s the good news. The bad news is that it can completely shortcut the process that not only are these kids dedicating four hard years of their lives to, but they’re also, their parents can be spending at a private school. I teach at a public university, but a private school is quarter million, half a million dollars. That to me is insane, and it’s not what this technology is best used for. So I caution people to use it as a thinking partner, as a sounding board. It’s a wonderful editor, but it cannot be used for primary creation.

And recently, I had an article that I wrote in my newsletter, my website, briankeating.com, and it has to do with the double-edged feature of nuclear power. Nuclear, as you said, fire was a powerful weapon. It could be used, it could be harmful, could be helpful. Nuclear technology is even better analogy. Nuclear technology is used to end World War II, but now it’s threatening America as well. I worry about this cognitive offloading if we do that for a generation or so. Will we even know what to think or how to think, or can we onload that cognitive thinking process again, or have we just totally uploaded it to the cloud and now it’s gone forever? I worry about that.

How many phone numbers do you remember?

Mike Slater: Yeah. How many phone numbers do you remember? I guess it’s my last [inaudible]

Brian Keating: Yeah, please go ahead.

Mike Slater: No, I was going to say, as an illustration of the danger that you just mentioned, how many phone numbers do you remember?

Brian Keating: I mean, I know my mother-in-law, my mother, my wife, in case they’re listening. [laughter] But besides that, I don’t even know some of my kids’ phone numbers. I can’t remember them. And that’s what we’ve done. We cognitively offloaded, offshored, just like we’re doing with jobs, with manufacturing. That has nothing but at least caution signs that should be flashing in your mind.

What is their world view

Mike Slater: I can remember my high school friends’ numbers. How about that? I can remember phone numbers from when I was 13 and I wanted to go to Ben’s house, but I don’t know anyone else’s cell phone now. I love that analogy. What do you say about the people who are running this? You mentioned their names, and that’s very interesting. What is their world view that maybe we should be concerned about as the people who are going to be controlling this technology?

Brian Keating: I’m very dismayed by them, to be honest with you. Some of them are one-degree-of-separation friends of mine, friends of friends of mine. Very scary when you hear things like Sam Altman. And when asked in front of Congress about these effects, and then considering, and the comparison that he made was to, well, don’t look at how much energy and data center usage my models use to train, to do these queries and responses. Look at how much it costs to train your child till he’s 18. And I thought, a child is a precious gift of God. And an AI is a piece of silicon, electrified and processed in data centers around the world, using tremendous amounts of resources that will never be able to capture back.

And so I worry about this. I worry what they do with your data. I don’t trust it with personal stuff. There’s ways to anonymize it, and I’m going to be writing about that on my newsletter. How do you do safe searching? How do you extract the most out of the technology but keep your private data? It’s the most valuable commodity that you have, because it’s the one thing like that they cannot do. They have to train on what humans have done.

It needs us

Brian Keating: Which is, by the way, one of the reasons I’m not as terrified as many of my doomer friends are that it’s going to create a hyper advanced species of Terminator 9000s, because it has to train on what humans do. It is not generative ab initio. It cannot start from scratch and come up with the Brandenburg Concerto or come up with Shakespeare or anything, God forbid, anything like the Torah or the Old Testament or the New Testament. It’s lame. It’s impotent. It needs us. It needs the prophets. It needs the minds. It needs human intuition. It can’t have that. So it won’t be what we’re worried about, but it can do a lot of damage in the meantime, and it can do a lot of good.

And I urge people to pay attention, to learn about this, to play around with it. Don’t expose private secrets, your medical stuff, your accounting, financial stuff. Make sure you trust what you’re doing. There’s types of AI called local models, which I write about on my website. How do you run on your own computer so it never goes to Sam Altman’s house and used by that alternative man named Sam?

Mike Slater: Dr. Brian Keating, experimental cosmologist, UC San Diego. Go watch Into the Impossible with Brian Keating. And then you said the newsletter, briankeating.com. Brian, great to talk to you, sir.

Brian Keating: Thank you. Great.