Things Daron Says on Podcasts

I don't know. I think that's why there are, I say, there are many gray areas that liberal ideas, and even at its best,

liberal philosophy will not have an answer, but we have to navigate them, trying to reach compromise.

Yeah, it used to be their call sign, you know, move fast and break things. And now it's, you know, be careful, shareholder value. We have an IPO coming up, and we'll talk about open AI and Hugging Face and the Navier-Stokes controversy in just a minute. But you talk about paradigms, you mentioned Thomas Kuhn, and I'm recalling your fellow Nobel laureate Max Planck, who said that science advances one funeral at a time. My question is: how much self-correction, how much introspection is good for institutions? I mean, you and I, as I said in my invitation letter to you, I mean, we are the, you know, we're basically scratching on a rock, you know, a blackboard with another piece of rock like Gal like people did in the University of Bologna in the year 1000. Very little has changed, despite, you know, it seems like academia is very resilient. What will it take for academia to fully change? So it would be unrecognizable to Galileo should he come back, you know, from the dead.

New and exclusive holiday decor just launched online at the Home Depot. Check out the viral Grand Duchess collection with smart holiday lights you can control from your phone. Easily customize the twinkle across trees, wreaths, and garlands to match your holiday vibe. Create stunning light effects, then switch it up with the tap of your fingers. Even the man in red will be impressed. Shop the holiday decor you want online only at home depot.com.

Well, look, conditions are different. So I don't know to what extent some of these adaptations we've made have been good adaptations for our times versus distortions. But I wouldn't say our real problem right now is that science has slowed down. It's more like we are not getting the fruits of science in productivity. And we have also sacrificed freedom of speech and diversity of opinion in many areas outside of science, sometimes even touching on science, but outside of universities, in schools, in organizations, sometimes we don't discuss and compromise as we need to in a pluralistic society based on liberal ideas.

I think he was so far-sighted in his understanding of some of these forces.

He really was. But in his mind, it was really the government doing this. And it seems to me, you and I, and maybe Max Tegmark, you know, he's just got off the phone with Bernie Sanders, you know, had this conception that the government's maybe not doing enough. And so, again, I keep asking you because you're an economist, but how do you balance this minimum, you know, this optimization? Because I think that's what economists do is as far as the ones that I've talked to on this podcast. So how do you balance that between the government and these companies? They're very driven by different things. And yes, Meta just got regulated and they had to pay a whopping, you know, one-tenth of a percent of their total market cap. So how do you balance, you know, it seems to me you should have punishment, but then you have the government with the Orwellian panopticon. How do you balance that?

Yeah, and then unfortunately, yeah, they won't be able to sell as many Viagra ads. And that will hurt my friend and fellow laureate of yours, Dr. Lou Ignaro, who is a brilliant man. Also worked on dynamite like Alfred Nobel himself. But you mix it up with Elon Musk, I've noticed. You're not afraid to mix it up. What is your take on Elon? Is it sort of a net good what he's doing overall? I don't like to ad hominem at all. So I'm not asking you to say anything you wouldn't say online or haven't said. But what's your take on economically, motivationally? What drives him? And where do you go to the public? Look, all

entrepreneurs are very driven. We like that about them. They're optimistic, they're self-confident, they're driven. They want often to build empires. They want to be larger than life. But that has to be within laws, within rules, within norms. And the problem is that we have created a society in which courts are weak, the government is weak. People have built tremendous, tremendous fortunes that are just unimaginable to people that would have been unimaginable to people like 20 years ago. And we have no guards rails against that. They can take over newspapers and social media companies and use them as their mouthpiece. That's where the problem lies. And the more they do that, the more they become convinced of their own power and the more they are surrounded by yes men, which creates pernicious dynamics. You know, Elon Musk shouldn't have this type of wealth. Part of it is because we have not taxed income. A lot of his wealth is in capital income, which gets taxed very lightly. If you look at companies that have made a huge splash in terms of Valuation, they often have often taken over their merge with their competitors. That should not have been allowed according to our own antitrust laws. So it's a failure of our system that these companies and these individuals have become so wealthy. We haven't taxed them. And now their power is not just economic, it's social and political. And that's where the problem starts.

You pushed on him and challenged him to actually be taxed on his unrealized gains. And I think he pushed back, or I saw some controversy about that. What do you make of that claim? That, look, Daron, you're brilliant, but these gains are unrealized. And the day he was a trillionaire, like two weeks later, he was not a trillionaire. So how do you balance that? I mean, yes, how many yachts does he need to water ski behind? That's true. But in reality, how do you deal with that question of unrealized taxation on unrealized gains for ordinary, maybe millionaires? I mean, not billionaires, trillionaires.

It should be.

Okay, so explain that argument. Yeah. How could it be there? What example? Use teaching. Use teaching as an

Today you tweeted about the alignment problem, which I've talked a lot about with people from Jan Lacun to Max Tegmark to Nobel laureates as well. But the question that you raise is that the problem that people are mistaking is not the capabilities, as you say. The capabilities are real. But more than the capabilities is, I quote, some imperfect quantitative metrics, user approval, user engagement, simple completion metrics. This is the process that leads to distorted behaviors in the form of gaming the evaluation of completion metrics, cheating, overconfidence and wrong answers, sycophancy. Although I love sycophancy, Durham, because I asked AI once, I said, what books has Brian Keating written? And he said, Losing the Nobel Prize into the Impossible and A Brief History of Time. I thought that was great. If I can get 1% of Stephen Hawking's book sale. Okay. And then you say an analogy may help. It's not that we have a supercar that has its own mind and wants to take control of driving because it has superiority to the driver. It's more that we have a car where the steering and brake systems don't work. It has many capabilities of very good cars, and its engine accelerations and GPU may be very impressive, but you cannot steer it properly. What good is it? Perhaps we shouldn't drive it until it's fixed. How would you fix it? You have control. What do you do to align it the way that you would best see fit?

Look, I refuse that control. I'm not an expert on AI. But my assessment is. That there are so many good things we can do with domain-specific models. You want to use AI in the production process, but better quality control, better design, better training, make the workers more productive. All of these are domain-specific problems. AlphaFold is a domain-specific AI model. Those models don't behave like cars without steering because they are designed for something very simple, very specific. Because they are designed for something specific, their metrics are much more aligned with what we want. It's these general models that are trained on all the literature, all the crazy stuff that's ever said on Reddit and Twitter. And then we want them to have as much agency and decision rights as humans. That's where the problem lies. So I think if we train these models much more for specific purposes and deploy them for specific purposes, we're going to do so much better.

All right, Duron, I know you got to go, but I have one more question for you. Please. I want to leave you with a line that I think is quite beautiful. I'm sure you've known it your whole life because it comes from Ataturk. He said, For civilization, for life and for success, the truest guide is science. You spent your whole career studying institutions, working at, arguably, outside of UC San Diego, the best scientific institution in America.