Who Are the AI Champions?

Speaker

[music]

Andrew Keen

Hello, everybody. It's Saturday, the 22nd of August, 2026. We're going towards the end of the summer. And there are two themes on the tech front. On the one hand, there's... Two of my VC friends from Silicon Valley suggested in an earlier show this week, Dave McClure and our man Virgie, were on the brink of a $7 trillion future when OpenAI and Anthropic go public. They're both forecasting that OpenAI, Anthropic, and SpaceX will be collectively worth about $7 trillion. So that's the good news, I guess, if you want to think in good and bad news. The bad news, at least in terms of Silicon Valley, is that... AI isn't popular. Ed Luce had a piece in the Financial Times this week suggesting that AI phobia has become America's new consensus. And America's best historian, I think, Jill Lepore, had a weekend essay in the FT suggesting that AI versus the people will define the politics not just of 2026, but of 2028. So this division between the economic locomotive of AI and the resistance, the unpopularity of it, is the theme of Keith Teare's We Are The Champions, editorial for That Was The Week. And Keith, as always, is joining me from the center of innovation, perhaps the most unpopular place in America, Silicon Valley. Keith, would you agree that these two realities are kind of jarring? On the one hand, it's clear that the AI economy is... Remarkably successful and profitable. On the other hand, most Americans are increasingly suspicious of AI.

Keith Teare

I think I don't agree, but I do agree that that's the zeitgeist. I think the zeitgeist exaggerates the negativity. I think, in fact, most intelligent people are now using AI from very young. High School onwards. And just don't talk about it. So the loudest voices are negative. And I fear for the Democrats that they're going to embrace that, assuming it's a populist rejection of AI and end up handing the election to the Republicans.

Andrew Keen

But then what's the point of your editorial? You said, who are the, and this is about champions, AI champions, and we'll come to those.

Andrew Keen

You say in the beginning of your editorial that you're concerned that the AI argument, I'm quoting you here, it's not the best English, the AI argument keeps arriving in negative form. So isn't that the problem that people aren't making the best argument for AI? That seems to be what you're saying in your editorial.

Keith Teare

Well, I am saying that, but if you weigh on a weighing machine, The people using it to good effect and the people speaking against it, I suspect the users who are happy with it massively dominate. But the voices are dominated by the negative. So in fact, it's a complex judgment call whether we're really on this anti-AI binge or whether it's really fake because the people that are Happy with AI don't feel the need to evangelize it. They just use it.

Andrew Keen

But then what's the point of all your hysteria? It's almost like a weekly hysteria you have about everyone who's critical of AI. It doesn't really matter. I mean, before we went live, we were talking about Gary Marcus's arguments, and he's one of the leading critics of AI. You suggested that he's out of control. He's an old friend, been on the show many times. Maybe he is out of control, but you just don't have to read him. So what?

Keith Teare

I think you can conclude that. You could nail me with why you're talking about champions when you say there isn't really a problem. I think that there is a different problem. That's why. The problem isn't negativity. The problem is lack of vocal connection. Between AI and people's futures, vocal connection. So AI is missing an articulate group of people who together are reinforcing an understanding of the future that will make normal people happy that it's happening. I think it's missing that. I think anyone who examines it does draw those conclusions. But of course, in normal life, most people don't examine these things. They just live through them. So the editorial is trying to stake a claim of saying, okay, well, there are champions. They don't speak coherently altogether, but here they are. Here's what they're all doing. Let's try to fill that gap, the gap that is the I'd say silence of the majority.

Keith Teare

So that the minority who are politicizing negativity can be exposed as a minority.

Andrew Keen

So you're reintroducing Nixon's silent majority. So this idea of champions, what does that mean? I mean, anyone who works for an AI company is, by definition, an AI champion. You don't include Amidai in this. You put Sam Altman. I don't want to get back in the... Altman versus Amadai, OpenAI versus Anthropic argument. But isn't by definition anyone who works in the sector, and it's increasingly the sector, everything in Silicon Valley, everything in innovation is built around AI. They're championing AI, aren't they?

Keith Teare

Well, I think the answer is not necessarily. They're very circumspect. You could call it nuance, but I don't think it's nuance. I think it's defensiveness. There's a circumspection and a defensiveness due to the belief that the negative ideas are strong and dominant. And people don't want to get on a bandwagon and go in the other direction fully. So they nuance everything to the point where they're really not making a point. Yes. On the margin, by the way, I left out Elon as well as Amadai, so I was fairly balanced with who I left out. I tried to pick the middle, the mainstream, if you will. And I do believe that if you are tactically running an AI company today, you've got this little voice in the back of your head saying, don't upset anybody and don't pretend it's safe. Even if it is, because people don't believe it is. So it's almost like they're being trained as politicians as opposed to business leaders. Meanwhile, of course, they are running a business and doing very well, as you acknowledged. But that isn't ideal, is it? I mean, can you imagine any other major breakthrough in human history where people would be shy to champion it? It's kind of a strange feeling to me.

Andrew Keen

You, in the newsletter this week, you refer to a tweet by David Sachs, the former AI czar, Trump's AI czar, now kind of a conservative commentator, very critical of Amidai and Anthropik. I'm quoting the tweet. Dario believes frontier AI is too powerful to distribute. We believe it is too powerful to centralize. Dario appears to believe sincerely that safety and progress are best served by centralizing authority in a marriage of corporate and state power. The weight of human history gives us reason to fear that outcome. Now, I'm in Dario's camp. I don't want to have this argument again with you. I know you're not. But couldn't you at least acknowledge that you may not agree with Dario's conclusion, but he's championing, trying to champion something?

Keith Teare

He's championing handcuffs on AI. He is championing, but it's handcuffs. The state power How could you interpret that other than handcuffs? And so he is basically saying, don't trust us, but if you are going to trust us, just trust me because I'm big and centralized. And that, if you look at Martin Varsasky's Post of the Week this week, which is all about what you can do without centralized AI, and it's a lot, there are really two trends. You could think of it as a, Purely capitalist libertarian trend that doesn't want the state involved and believes people will get control over distributed intelligence through their own hardware at the edge. Varsesky even says it could be done without the electricity grid. So imagine some post-nuclear situation. As long as you've got a box and some solar power, you could have the world's intelligence in that box. It wouldn't have been destroyed and it's there forever. So there's that libertarian trend which Varsavsky really embodies. And then there's the corporate state power trend which Dario embodies. I think most people are somewhere in the middle. I'm in the middle. I don't mind

Keith Teare

I'm fine with that. You know what they

Andrew Keen

said about the middle, Keith, in Texas? The only thing in the middle of the road are squashed armadillos. So you really want to be squashed? You've got to take a position. Either you're in the Saks camp or you're in the Amadai camp. My sense with you is, for all this fiddling around, you're in the Saks camp. If you had to choose between the two.

Keith Teare

I agree with Sachs, I'd say 90%. Yeah. Yeah.

Andrew Keen

So you're in his camp. You're not in the middle. Why even talk about being in the middle? You're a Sachs guy.

Keith Teare

I don't think he has a camp.

Andrew Keen

I think my book- What do you mean he doesn't have a camp? Of course he does. I mean, for better or worse, he has a position.

Keith Teare

No, no, he's not. He's just an individual. He has no followers. He has no movement. He doesn't represent- Well, I don't mean his camp. I don't mean that Sachs

Andrew Keen

has a camp itself, but that position, that general position against the idea of centralizing authority and what he calls the marriage of corporate and state.

Keith Teare

I don't agree with it. That's the 10% I don't agree with. I'm completely happy, let me say it again, with centralized, large AI companies like Anthropic and OpenAI. I'm totally happy. So that differentiates me from SACS. I'm also totally happy that the edge is growing, which also differentiates me from Amidai. And I believe that the human outcome or the dividend we'll get from this is free intelligence that needs investment. And I don't believe that investment can come without large centralized corporations.

Andrew Keen

Jill Lepore argues that in the past, I mean, she's a historian that the car industry, for example, she writes about in the AI versus the people, she has a new book on this, that that required a degree of regulation. I mean, this is, again, a debate that's ongoing about Whether or not innovative industries need to be regulated, what's your take? Do you agree with Lepore that AI isn't any different from the car industry, for example, that required a degree of regulation about how one can drive, laws for the road, that AI needs the same sort of stuff?

Keith Teare

Yes and no. The no bit is that

Keith Teare

AI is just corporations. And there are already regulations surrounding corporations. And there is nothing new needed to govern corporations. Every legal remedy needed for anyone who does anything illegal is already in place. And so the call for special regulation is not required. I don't think AI itself as a technology needs regulating. I don't think it's dangerous. I think it's safe. I think it's not only safe, it's fantastic for humans to use it and they should use it more and more. And anything that slows that down is anti-progressive. So on the whole, I disagree with her. I do agree that in general, states need to regulate corporate behavior to stop illegality.

Keith Teare

And that's fine. And so, you know, AI companies have to live within the law, just like everybody else. But I think what is happening, and this is especially true in journalism and book publishing, is if you want to sell, you have to create a bandwagon. And the bandwagon around regulating AI is a fairly weak one.

Andrew Keen

So you're suggesting that You're suggesting that Jill Lepore made that argument in order to sell a book or sell to publishers. I don't think she quite needs that. I mean, she's America's leading historian.

Keith Teare

Well, historians usually don't comment on current opinion. So she's obviously not acting as an historian in this piece. She's acting as an opinion piece. And that opinion, AI versus the people, is fundamentally going to damage The left in America by identifying against AI. I honestly think they will lose in 2028 if they stick to that idea and quite heavily.

Andrew Keen

We didn't put it, I didn't think you put it in the newsletter, but there was an interesting, Ezra Klein probably has the most popular podcast show, certainly amongst progressives. He interviewed Helen Toner, who used to be on the OpenAI board. She was one of the people involved in the scandal when Altman was being pushed out. She talks about AIs already being out of control, the hugging face episode. You said they're completely safe. Are you not concerned at all about the hugging face incident or the fact that AIs seem, at least in the language of Toner's conversation with Klein, to be out of control?

Keith Teare

Honestly, no. The hugging face incident was contrived in a lab to try to see how far AI would go in order to achieve a goal. And the only critical part of it could be that they didn't pay attention. So it took them a few days to figure out what was going on. But it was a laboratory experiment. And it wasn't very damaging, frankly. I mean, these AI tools And now this is true even for normal individuals like GrokBot, can control browsers, login sessions to websites, your hard drive and the files on it.

Keith Teare

We are giving them access so they can achieve things for us. Like yesterday, I got a text from GrokBot at two saying, Your calendar's telling you to drop off the car at the airport at 3.30, but your flight's at 4.11, so ignore that. You need to check it in by 2.30. And it knew that because it had access to my calendar, and it knew where I was on my GPS on my phone. So we're opening ourselves up to AI having access to software and tools. And outside of the context of what's called a harness, You may have a fear that it's going to do things, but it doesn't have a consciousness where it has a plan. It only has the goals that you set for it. So if I said to the AI, I want you to delete my calendar,

Andrew Keen

it would. Yeah, you should listen to the toner piece. You clearly haven't. But I mean, you weren't expected to. Listen to what? The toner conversation with Klein. In your... In your editorial...

Keith Teare

Are you telling me I'm ignorant of something?

Andrew Keen

Well, I just... Yeah, I mean, you're dismissive rather than ignorant. Well, I think I'm an expert that uses it, so I think

Keith Teare

they're the ones who are...

Andrew Keen

Well, Tona was on the OpenAI board. Now she runs a foundation. But anyway...

Keith Teare

We all know how bad the OpenAI board was. It wasn't technical, Andrew. You shouldn't elevate people above their station.

Andrew Keen

Yeah.

Keith Teare

Yeah. Why do you do that? Do you think she has actually got technical chops and understands this stuff? Well, she sounds certainly more intelligent than some other people I can think of.