Do We Really Want a No-Hands Job From Silicon Valley? Who Holds the Power in the Age of AGI

00:00:31 Andrew Keen: Hello, everybody. It's Saturday, May 2, 2026. As always, I'm talking to you from San Francisco on the edge of Silicon Valley. It's been quite a week for tech. All the big companies, Amazon, Google, Microsoft, and Meta reported their Q1 numbers, had a huge reaction in the market. Google was up 10%. Meta down almost 10%. A lot of it had to do with what financial people call CapEx, capital investment, which is blooming, blossoming, exploding, bloating, depending what adjective you want to use. And those are defining, in many ways, the direction of the big tech stocks. Meta's down because Wall Street feared Meta was overspending, doing too much capex, whereas Google got away with it because their numbers are so good. So I was expecting in That Was the Week, Keith Teare, the publisher of That Was the Week newsletter — a summary of what's been happening in tech — to lead with news of all the financial doings and undoings in Silicon Valley. But instead, I found his newsletter leading with a video of a rather attractive woman with the words handjob question mark. And I wonder, Keith, whether your That Was the Week has been hijacked. Are you into porn now? What's going on? Hand job? I didn't know that we did hand jobs these days in tech.

00:02:15 Keith Teare: I think I've got a career ahead of me as a headline writer, Andrew. If BuzzFeed needs somebody, they can — I don't know.

00:02:23 Andrew Keen: I don't think BuzzFeed is still in existence. I think they went out of business.

00:02:28 Keith Teare: Oh, we lost Andrew.

00:02:30 Andrew Keen: Didn't, BuzzFeed go out of business several years ago?

00:02:34 Keith Teare: They may well have done.

00:02:35 Andrew Keen: Probably because of headlines like hand job question mark with an attractive looking woman. But anyway, why are you leading with hand jobs this week, Keith?

00:02:46 Keith Teare: It's interesting because you didn't play the video to the end where the subtitle comes up. You purposely left it in the most provocative

00:02:53 Andrew Keen: Would I ever do anything like that?

00:02:56 Keith Teare: But, but firstly, you're right. The final third of my editorial this week is all about the earnings results. So, you're not wrong. But I led with the change in the user interface to computing. It's really the redefinition of the word software, which was, very prominently coined this week in an interview that Andrej Karpathy did with Sequoia Capital at their annual AI Ascent.

00:03:27 Andrew Keen: And that interview is in That Was the Week newsletter for this week.

00:03:31 Keith Teare: It is in there. And he made the point that, you know, he really doesn't use his hands anymore to achieve his work tasks. He's speaking to the computer and it's doing things, and then he's getting things back. And he made the point that, you know, that era where you load a web page, type something in, possibly click send, and that would include emails and text and everything is not quite in the rearview mirror, but the end is nigh. And that led me to conceptualize — you know, about two years ago we did a That Was the Week called "eyes, hands, ears, and mouth." And it was all about the inclusion of the other human attributes beyond hands. Two years later, it's not unreasonable to talk about the end of the hand driven computer era.

00:04:30 Andrew Keen: Yeah. I mean, some people are gonna be listening to this thing, Keith. Maybe there's a Freudian element here. You're talking about the clothing coming off. I'm quoting you here. That clothing is coming off. Whether that's wishful thinking on your part is another matter, but it's worth going back to all this CapEx stuff. I mean, we're talking about doubling down.

00:04:53 Keith Teare: Oh, you don't like my theme then.

00:04:56 Andrew Keen: Well, I just think — I mean, maybe I don't know — but you're, so to speak, maybe burying the lead. I mean, the big story is this CapEx investment. I mean, the overshoot reports that we're talking about double the amount of CapEx investment in 2026 from last year. And it's 2% of all the goods and services produced in the United States. So we're talking about very significant numbers. Our post of the week talks about this — $700 billion on AI infrastructure. I'm quoting the post from @Signal. This kind of spending usually happens via governments or wars, whereas this time, it's four companies racing to build the foundational mechanics of AGI. Those four companies, of course, are Microsoft, Amazon, Google, and Facebook. You argue in your editorial, Keith — and you do come to this, as you know, you're not ignoring it entirely — that this isn't that unusual. That great change is always created by tech companies. So perhaps you might explain that. Why isn't it that concerning? Why don't you agree with @Signal that we're in a new historical era when it comes to private companies investing in infrastructure?

00:06:38 Keith Teare: Well, firstly, it's not a new historical era. I mean, every era, you know, the American railways were not built by the government, and neither was the Internet.

00:06:48 Andrew Keen: Hold on, mate. But here's where I'm not sure you're right. So you say — and I'm quoting you here — "innovation waves have always been privately financed." And you say the steam engine, the railway network, the telephone, the electricity grid, the car, commercial aviation, the personal computer, the commercial Internet. But explain what you mean by the commercial Internet. How was that funded by private companies?

00:07:16 Keith Teare: Well, if you measure the Internet by traffic or by revenue, none of it's coming from or going to the government. What the government did is fund a research project at DARPA that led Vint Cerf and Bob Metcalfe to invent TCP/IP. And —

00:07:35 Andrew Keen: Hold on. I mean, you're not giving it full credit. This was ARPANET, which was a massive government investment, which created the very infrastructure for the Internet. I mean, this is a—

00:07:44 Keith Teare: But the investment didn't create the infrastructure — it created the protocols. ARPANET was a pathetic university-based bulletin board style Internet that I used to use in the nineteen eighties. It wasn't nothing — it was the precedent that needed to be in place for the Internet to exist. So, yeah.

00:08:08 Andrew Keen: I'm not sure everyone would call the ARPANET pathetic.

00:08:11 Keith Teare: That may be your — by modern standards, it was, yeah.

00:08:14 Andrew Keen: By modern standards — but that's not really the point. Our interview of the week is with John Steele Gordon, a distinguished economic historian and contributor to a Wall Street Journal series on the most impactful US inventions, in which the Internet comes up number one. But as the Wall Street Journal suggests, Keith — this, for better or worse, was a Cold War project that came out of ARPANET, funded by the Defense Department. So this wasn't private. I mean, the reality is that

00:08:53 Keith Teare: No. But you shouldn't flatten the argument. Don't flatten the argument.

00:08:57 Andrew Keen: How am I flattening the argument?

00:08:59 Keith Teare: Well, you know, you're making the case that research funded by government is important to innovation and I'm not gonna disagree with that and that's true of AI as well.

00:09:08 Andrew Keen: No. I'm not I'm not making that. I'm simply saying that you said that what you call, or what you mean by, the commercial Internet was funded by private companies. But the reality is that the Internet was built by the government —

00:09:25 Keith Teare: No. No. It wasn't built. You're using the wrong word. It was research that was funded, which led to protocols. That's the truth. And that those protocols took on a form, which, it starts, as you were just about to say, with the National Science Foundation and then DARPA and ARPANET. But these were primitive research projects, not very heavily funded, actually, that led to the first Internet service providers like EarthLink here in The US or my company, EasyNet, in Europe. And we were the ones who raised money to build the commercialization of a protocol.

00:10:07 Andrew Keen: Although, maybe that's true, but this didn't happen till about 1995 because of the

00:10:12 Keith Teare: true '94 actually

00:10:14 Andrew Keen: Well, '94 — because only then was the law changed so you could have, quote unquote, a commercial Internet. I'm just trying to defend this guy. I don't even know whose Signal is — I found this; it was an interesting post. But so are you suggesting that what Google and Meta and Amazon are doing is like the commercial Internet 1995?

00:10:45 Keith Teare: Well, look. Let's start with this guy. He's not wrong that this scale of investment is normally associated with governments and wars. And I think both words are important, governments and wars. And so the scale of investment is very large, and it's private. I do think that, you know, the Internet bubble in the late '99/2000 era was very small compared to this, but it was the second biggest. It was also huge. It was also people like Amazon, Yahoo, Netscape, Microsoft, who were who funded the global growth of the Internet, including lots of, non American companies in China in particular. So the guy isn't wrong about the scale of investment and the fact that it's private. And I wasn't really criticizing him when I made the point that it's normal for the commercialization of any technology to be mainly private. That is normal. I can't think of a non private commercialization of any scale. Even in China, I make the point in the Internet. It's private companies, that, you know, both raise money, then build systems, then get revenue that fuels their building even more if they're successful. Let me

00:12:08 Andrew Keen: Let me ask — maybe this is a bit of a dumb question — but a lot of the money, the CapEx investment is going towards data centers. And this is an increasingly controversial issue. Lots of local communities, for one reason or other, for better or worse, don't want data centers. New York Times describes opposition to data centers as the most bipartisan issue since beer. In other words, everybody agrees on the left and the right that data centers are, for better or worse, wrong. There are fights all over the country — 37 data center fights just in Virginia alone, going to the Virginia Supreme Court. These companies are investing more and more in data centers. If you want to extend the metaphor to the Internet from the nineties, Keith — how would you describe the equivalent of a data center, or this investment in data centers, in the nineties? Wasn't the Internet built out by them pretty much? Maybe it was a bit pathetic, as you note. But isn't this private investment in data centers — which is the core of CapEx, what all these companies are spending their money on — going to mean that mean that these companies essentially own the infrastructure for AGI? And in that sense, isn't @Signal correct?

00:13:37 Keith Teare: Let me take you back to 1994 when I built a data center in Brick Lane in London for EasyNet. At that time, pretty much every London university had had an Internet node for some years. There was no such thing as a data center in any of those universities. There were a few modems hanging off a wall and a router. My data center in Brick Lane lived alongside about five other data centers in London that were all owned by telephone companies. And so 1994 was an entirely privately funded Internet at any scale.

00:14:19 Andrew Keen: But yours was a start up, and the telephone companies weren't the smartest. They didn't really understand what was happening.

00:14:27 Keith Teare: Well, they understood enough to build, TCP/IP, facilities. British Telecom in 1994 came and did a deal with me to train

00:14:37 Andrew Keen: [unclear interjection]

00:14:39 Keith Teare: No. No. No. No. Well, I'm I'm just trying to understand history. The history of the Internet is not a history of government success. It is also a history of government success, but that is such a small part of the story.

00:14:54 Andrew Keen: Okay. But I don't think that's the argument. We're not talking about whether — what are you saying?

00:14:58 Keith Teare: Why are you focused on this?

00:15:02 Andrew Keen: Because it's such an important issue, I think, for the future of the AI economy. Yeah. Well, let's go back to your example then of, let's go back to your example of British Telecom. Whether or not they invested in your data center in Brick Lane, it didn't mean that, quote, unquote, they owned the Internet, did it? What did they get for that investment? What

00:15:30 Keith Teare: Well, they I'll tell you how pathetic it was. This is why startups are so important. They, gave me three years of exclusivity to dial up Internet in The UK. They even manned an 800 number to field consumer calls to get people to sign up for EasyNet. And all I had to do in return was buy a T1 line from them, into our headquarters, which was also Cyberia Cafe, which cost about £12,000 a year. So for £12,000 a year, they gave me three years of exclusivity to the British Internet for consumers. That's how little they understood about what was going on.

00:16:13 Andrew Keen: Right? As I said, they weren't very smart, but we're talking about entirely different, stratospheric numbers now. Goldman, I think, estimates that when we talk about almost $800 billion in 2026, this CapEx investment is gonna go up to about $1.6 trillion in 2031. And most of that is gonna come — this is not a startup economy. What we're talking about is

00:16:38 Keith Teare: Andrew, what is your point about that? Are you saying it's too big?

00:16:43 Andrew Keen: I'm simply suggesting that it's a different historical moment when private companies, and we're not talking about startups. We're not

00:16:51 Keith Teare: I agree. It's a different historical moment, but I think it's

00:16:55 Andrew Keen: To requote @Signal, this kind of spending usually happens via governments or wars. Whereas this time, it's four companies racing to build the foundational mechanics of AGI. Now these aren't startups. Hold on, let me be clear. We're not talking about OpenAI or Anthropic or any of the other many AI startups. We're talking about four companies — the four biggest tech companies. So this is —

00:17:23 Keith Teare: But what are you suggesting —

00:17:24 Andrew Keen: I'm simply suggesting that this guy is right, that this kind of spending usually happens via governments or wars, whereas this time

00:17:34 Andrew Keen: It's racing to build the foundational mechanics of AGI.

00:17:38 Keith Teare: Yeah. I mean, I obviously made him the post of the week for a reason. I think he's right as well. So I'm not sure what we're debating. He's right. But thank God. Thank God for those four companies. Because without those four companies, we wouldn't be getting this.

00:17:55 Andrew Keen: But the reason — and you know this better than I do — why people don't trust AI, and you cover this in the newsletter, particularly in the US. It's really weird. People trust AI less in the US than they do in China, which — maybe we can come to that later. It's because — not OpenAI and Anthropic — but these large companies are controlling the AI infrastructure. Isn't that the reality?

00:18:31 Keith Teare: What do you mean? You sound like a conspiracy theorist. What do you mean they're controlling our AI thing? What do they control?

00:18:38 Andrew Keen: Well, they own the data centers. They're investing in the infrastructure and this goes back from

00:18:44 Keith Teare: It's like a telephone company because it owns the telephone in your house, as used to be the case.

00:18:51 Andrew Keen: You know? As you've noted, in the nineties the telephone companies weren't very smart. They didn't understand what they were doing. Whereas these companies are massively powerful and incredibly —

00:19:03 Keith Teare: You sound like a MAGA conspiracist. I mean, there's nothing controlling here. The control, to be honest, is largely in the hands of individual users of AI. What are you using?

00:19:17 Andrew Keen: So you're suggesting that the fact that these

00:19:19 Keith Teare: They're facilitating — that's what they're doing. They're facilitating, and they're gonna make money from it as well, which they should.

00:19:29 Andrew Keen: So you don't think this is particularly significant? I think, in historical terms — there may be four, maybe five, there may be three. Who knows what the future is? Maybe they won't be able to compete with Google or Microsoft or maybe Apple will appear in this play. But isn't this — you don't think it's very significant?

00:19:54 Keith Teare: Well, I'm trying to detect what it is you care about. It sounds like you have a concern, whereas I only have — I want to stand and applaud. So where my applause comes from, you know, and the title of this week's hand job thing —

00:20:11 Andrew Keen: Oh, hand job. The video — they're taking our minds off it with this beautiful woman. Maybe we're all getting a hand job from Big Tech.

00:20:20 Keith Teare: They're funding the next evolution of, you know, human-computer relationships, which, I put in the editorial, are gonna be more like human-human relationships. Like, when you go to a restaurant, you talk to a waiter — they don't bring you a keyboard to type in what you need. Although these days, some of them actually do that. And, you know, our relationship to computing has been through keyboards and mice and touch screens. They are funding the possibility that we will relate to computing the same way we do to humans, speaking to them, getting answers, querying things, that are much more natural and at massive scale. And it's gonna change everything, and it costs a lot of money. But they're also

00:21:11 Andrew Keen: So as you say, the clothing's coming off in every sense. I mean, I'm not suggesting it's necessarily a bad thing. All I'm trying to do is support @Signal. I mean, maybe Signal's an AI. I don't know if Signal needs support.

00:21:22 Keith Teare: I'm supporting him too.

00:21:25 Keith Teare: Yeah. The only word I would disagree with really is the word for. It's, you know, if it was spelled f o r. It isn't for the companies. It's for us by the companies.

00:21:43 Andrew Keen: I'm not gonna make fun of your northern accent, "oos," but —

00:21:48 Keith Teare: "Oos."

00:21:50 Andrew Keen: But the reality — and I'm not sure we want to get into all these arguments that divide us, Keith — is that these companies are not doing it out of some sort of civilizational altruism. They're doing it for better or worse, and I can understand it, to pursue their own interest. So they're not giving away this technology.

00:22:07 Keith Teare: Well, that's an interesting insight, Andrew. I mean, I don't think civilization is a conscious project. I think civilization is what you get through innovation. And innovation, you're quite right, is driven by self interest. I actually think self interest is underestimated as the key driver of civilization. Without self interest, there'd be no impetus. And so self interest is a big part of it. And the more they give us these tools — and one of the parts of the editorial this week is that the devices are gonna change. This is gonna be pushed to the edge. I can't wait for Apple's developer conference in a few weeks' time because I think we'll see the first signs of it there. But OpenAI and Jony Ive are also moving in this direction.

00:23:02 Andrew Keen: They're at the edge. I'm always suspicious of that term — it's a metaphor. But the reality is, if the edge is owned by Apple or OpenAI, then I'm not sure it's really the —

00:23:13 Keith Teare: Edge — the edge is owned by you. The very definition of edge, if you use the word properly, is it's owned by you. You pay a thousand dollars for an iPhone. It has models on it, and it's capable of interacting with you, and it doesn't go through Apple's data centers. So it truly is the edge.

00:23:29 Andrew Keen: But coming back to this — doesn't this mean that the data center, which I always thought was a bit, I won't say trivial, but maybe in some ways ideologically frivolous, is in fact an important issue? I mean, it's gonna become — as the Times suggests — the most bipartisan issue since beer. Opposition to data centers. Not everybody understands, but these data centers are the operating systems for big tech in our AI age. Is that fair? How would you describe it?

00:24:03 Keith Teare: Well, I think it is as important as you suggest, but for a different reason. It's emblematic of America's backward-looking elements, that this is controversial in any way at all. The reason it isn't controversial in China is there's a general consensus in China that AI will be good for society. The reason it is controversial in America is because more backward-looking — one could say Luddite-like — sentiments, fear, if you will, as opposed to optimism, dominates the political agenda, and the politicians are scared of taking it on. So as you rightly say, that it becomes a bipartisan issue. But it is entirely symbolic or emblematic of America's decline on the world stage, whereas the actual drive to build them is emblematic of America's leadership.

00:24:57 Andrew Keen: Yeah. It's ironic. And one of the pieces that you cover is the Post's report on AI. Americans are down on AI. Americans are amongst the most suspicious, the most skeptical, about AI. They are less excited about AI than almost anyone else. Whereas China's the most excited, I think, more than anybody, more than Indonesia, Thailand. 84% of Chinese people interviewed for this survey are excited, whereas only 38% of Americans. And this is the week where DeepSeek's sequel — DeepSeek being one of the Chinese challengers in AI — failed to impress. So it's kind of ironic that in America, which leads the world in AI, people are very skeptical. Whereas in China, which still lags behind, most people are optimistic. Is that just a paradox of history? How would you explain it?

00:25:55 Keith Teare: Look. This is one of those areas where you've got to applaud the Trump administration. It is gung-ho — you know, China-like — in its support for AI. And, from the point of view of government leadership, it's a bright shining light when it comes to its understanding of how important AI might be to the future and facilitating that happening. You know, it's in stark contrast to Bernie Sanders, who wants to have a moratorium on data centers being built.

00:26:26 Andrew Keen: So hold on. I mean, you don't have to be a Trump fan to believe in AI. I mean, the accelerationists on the center-left

00:26:35 Keith Teare: Yeah.

00:26:35 Andrew Keen: Ezra Klein, for example, are

00:26:37 Keith Teare: I'm not saying that unsupportive.

00:26:38 Andrew Keen: So it's not either Trump or Bernie Sanders?

00:26:42 Keith Teare: No. I'm just painting a picture of the two extremes, and you're quite right. There's lots in the middle.

00:26:48 Andrew Keen: And some of the Trump people are actually deeply hostile to AI too.

00:26:52 Keith Teare: Yeah. But the main point is that, look, America is performing extremely well in AI anyway. Data centers are gonna be built. They are being built. These sums of money, which you rightly say are large, are probably gonna look small five years from now, because there'll be even bigger numbers, and there'll be bigger revenues. So it is a classic case study in how capitalism invests to produce more value, and it is going to produce more value.

00:27:31 Andrew Keen: So let me ask you this question on infrastructure. I mean, maybe it's not the right word. We've already talked about the commercial Internet. We noted that up until 1995, there wasn't a commercial Internet, and then it became commercialized, and that it was funded by ARPANET, which was a defense research project for better or worse. Will these data centers — if they continue, if the Goldman reports are right — let's say by 2031, over $1.5 trillion is invested annually in CapEx, which is mostly — let's say — continues to be data centers. What will that mean in terms of the ownership of AGI? I mean, this comes back to @Signal's point, to build the foundational mechanics of AGI. Will that mean that the big four — or the big five, or the big three, whatever they are by 2031 — will own those mechanics? Isn't that the key question?

00:28:41 Keith Teare: So I think you've got to peel the onion a bit here to understand it. I think the first thing to say is that it's not at all clear that by 2030, LLMs will be the dominant form of AI.

00:28:55 Andrew Keen: Right. So whatever it is. I mean, I don't think that the Goldman report makes any bets, but it's certainly gonna be one kind of AI or another.

00:29:07 Keith Teare: And then the second thing I'd say on the onion being peeled is, we are already in the age of agentic AI, which sits on your computer and interacts with your data, mainly unstructured data, and can carry actions out on your behalf, which is the hands off, part of it. We're already in that phase. Anyone using AI as a chatbot is living in a two years ago kind of age. AI can now do things for you and deliver stuff automatically based on your instructions. And it can even question your instructions and improve them. So we're already in that stage. So once we get to device-independent AIs — which we're very close to already. If you look at China, the main interesting thing happening there is the open-source models that now run on devices like your Mac mini. And so you can't buy a Mac mini anymore. They're sold out because everyone's running their own AIs on them for free, by the way.

00:30:13 Andrew Keen: Not everyone, but people who wanna run their own.

00:30:15 Keith Teare: People who want to are. So the historical trend through to 2030 — Goldman Sachs, obviously, 2030.

00:30:21 Andrew Keen: You missed the second.

00:30:23 Keith Teare: 2030. Goldman Sachs is extrapolating from a present that probably is fictional.

00:30:29 Andrew Keen: Okay. But we're talking about a lot. So — you suggest, I mean, whether it's a trillion or a trillion and a half.

00:30:34 Keith Teare: I think it may well be five times greater than that.

00:30:40 Andrew Keen: Okay. Well, even more so than if it's $7 trillion. And if that money is coming from private companies, what exactly does that mean?

00:30:48 Keith Teare: Well, again, you've got to trace what that means, money from private companies. How does that work? You know, this week, Anthropic is being asked to sell shares at a trillion-dollar valuation. Anthropic. If it says yes, it's gonna take money from investors. Where does that investor's money come from? It comes mainly from pension funds or university endowments. So this is the people's money funding the next stage of evolution. And if it works, it will go back to the people in the form of gains.

00:31:25 Andrew Keen: Yeah, but is that true? I mean, would you include Google and Microsoft and Amazon in that? You'd leave Anthropic out of it at the moment.

00:31:35 Keith Teare: Yeah. Meta this week opened up a $40 billion bond that was snapped up and oversubscribed in 24 hours. Where does that money come from to buy that $40 billion bond?

00:31:49 Andrew Keen: So you're suggesting that the stock market is a kind of public investment fund. So it's the equivalent in a sense of the of ARPANET.

00:31:57 Keith Teare: Well, there's no such thing as, you know, in the stock market, government money. All the money in the stock market is private, and it includes everybody who invests in the stock market. And even the institutions that do it institutionally — where do they get their money from? They get it from the source of capital and GDP.

00:32:20 Andrew Keen: So you're suggesting that really when it comes down to it, when we do a hand job on these on the mechanics of, capitalism, private companies are actually public.

00:32:32 Keith Teare: Well, that's the whole nature of a stock. When stocks were invented, it was a revolution. It was the first time individuals could own companies, instead of private feudal-like ownership. So a stock is a democratizing asset.

00:32:50 Andrew Keen: Well, now we have reason to trust all these private companies. What do you make, Keith, of this week — we always talk about the wacky races on this show, because every week one big company or another, or one startup or another, seems to be ahead. I read a couple of pieces — I couldn't find them actually, for this week — but there were a couple of pieces that suggested (and you always find this on big weeks in the market) that something has changed in the long term. Do you think the Google numbers, the Meta numbers, the Amazon numbers — what do they tell us about the long-term trend in the wacky races between these big tech companies?

00:33:32 Keith Teare: I do think there's a kind of differentiation between them. If you take Meta's numbers, they're a sign of its indecision and its attempt to correct course. And so it's spending more for less than the other three, because it's having to re-orientate the company to a certain extent, and it's not really a player. So it got hammered. Microsoft, through the deal they did with OpenAI last week, has released OpenAI to be a more prolific partner across the whole industry. And Microsoft's Azure revenues, which are heavily subsidizing OpenAI, and therefore shrinking Microsoft margins will suffer through that decision, but it was probably inevitable. But they'll get it back because they own about 20–30% of OpenAI. I think Amazon is the big winner because OpenAI can now spend money with Amazon and is doing as Anthropic is as well. And then Google is probably the biggest winner of all because Google is successfully navigating the path from a search engine advertising company to an AI company very, very well whilst preserving its revenues from advertising. Even though the gross margin level, is being impacted, it's doing very well. And then, of course, there's the hidden ones — Anthropic and OpenAI because they're not public, and probably SpaceX should be included in that. So those three are kinda hidden. They're even bigger winners, especially Anthropic and OpenAI. It's just not visible to us yet. It will become visible once they list.

00:35:31 Andrew Keen: We'll see. I tend to agree with you on Anthropic. I'm much less bullish on OpenAI, but we won't relitigate that one, Keith. We often talk about it. You mentioned the advertising industry. There was a very good piece in the Financial Times on what they were calling the end of the road for the Mad Men — "not the Mad Max" — as AI moves into advertising. People always talk about porn being, so to speak, the canary in the coal mine when it comes to the future of tech and economics. But it may be advertising — what's happening in the advertising industry, Keith, as AI moves into it. Google and Facebook, of course, were always the leading players. But will there be any advertising companies in an AI age or just a small collection of high end boutiques?

00:36:24 Keith Teare: Well, you know, I think — the video I put in the editorial this week was produced by me using Claude and a system called Seedance, which is a Chinese video [generation tool]. Yeah.

00:36:35 Andrew Keen: You know, you're going to get banned on YouTube when your title is "hand job" and you have an attractive woman on the cover. You're not gonna —

00:36:43 Keith Teare: There is a subtitle that is meant to offset —

00:36:45 Andrew Keen: Well, the subtitle is very, very small. All I see is "hand job," but go on. Maybe I have a one-track mind.

00:36:55 Keith Teare: Well, I knew that. That was what encouraged me to stick with that title. Sticky title. Right? You know what? When I publish this on Substack, they do a title test, and I put three titles in, And it tells me which is the one people click on the most, and that becomes the title.

00:37:12 Andrew Keen: I didn't know that. So they tell you which is the best — they give you a handjob themselves, or three handjobs.

00:37:18 Keith Teare: Exactly. And now I've forgotten your question.

00:37:23 Andrew Keen: So the question was about the future of advertising.

00:37:28 Keith Teare: Oh, yeah. Advertising. So look — advertising is, after coding, probably the next big impact of AI. Firstly, design is pretty much a solved issue now. Design of anything.

00:37:44 Andrew Keen: Right. And you got into your hand job thing because you were talking about how you produced this semi-pornographic video — or video with a pornographic title — yourself. You didn't need to go to an expensive, video marketing boutique.

00:38:00 Keith Teare: But, you know, an individual idea person can create a design for anything if they can describe it, and get an end result in minutes, if not hours, that would have cost a lot of money to agencies in the past. Yeah.

00:38:19 Andrew Keen: And, of course, Canva is doing very well, and Anthropic threw their hat into the design ring a couple of weeks ago. So the big players are involved as well. Big tech —

00:38:29 Keith Teare: Well, if you take an average advertising agency like WPP — it's a bunch of suits facing off to customers and then filtering through to a bunch of creatives, you know, what the customer needs are. You're now at the point where the customer themselves can articulate their needs and get back material using AI without paying anybody else. And so you really are at the point where the advertising agency is mainly going to be a distribution agency doing the operational side. And that's the first part of my editorial, really — which human types survive this. And my answer is: leadership, creatives, and operators. But specialists, you know, get replaced. Now the specialists can still be working for the companies, but they're gonna use AI to scale themselves. So you can have 100 advertising agencies working for you in AI, you know, for the same price as one. So, you know, it's not a threatened discipline. It's threatened at a level of, does it need humans and how many does it need and what kind of humans does it need?

00:39:56 Andrew Keen: Yeah. And — it's an interesting and, I think, important piece — don't put your daughter or son into advertising. A lot of liberal arts graduates from these expensive colleges in America can't think of anything else to do, so they all want to go into marketing. But I'm not sure what they're gonna do when the marketing jobs dry up, Keith.

00:40:22 Keith Teare: Well, they can use — you know, I think when Sam Altman this week said it's the revenge of the idea guy. And you'll empathize with this, Andrew, because you and me have both been through this. In the past, an idea guy came up with an idea and said, I need some engineers to help me implement it. And that's true in physical or digital creativity. And many of our ideas never saw the light of day because we couldn't afford to get the engineers to help us build it.

00:40:56 Andrew Keen: Now we can.

00:40:57 Keith Teare: Now you can. And so suddenly, the idea guy, instead of being, you know, this frustrated creative, can actually lean in and make things happen. And that's gonna change how many things get made. And, you know, our experience of that as receivers, as consumers, is gonna be an abundance of choice.

00:41:26 Andrew Keen: We're back to abundance. Let's end with some education. People are deeply fearful of the impact of all this tech on our children. We're paranoid about our kids these days. There was a big New Yorker piece by Jessica Winter: "What will it take to get AI out of schools?" I think you and I keep going on, always agreeing on everything. We're both astonished — given that, why would you want to get AI out of schools? Why would you not want to equip kids with the technology that will enable them to be competitive in the workplace? What do you make of these kinds of articles in The New Yorker, which is, you know, read by everyone in New York City and Brooklyn and all the rest of the liberal progressive coastal enclaves?

00:42:11 Keith Teare: I think it's symptomatic of a particular use case of AI that for those people is the only use case. And that is that they see AI as a chatbot that they ask questions of. And their nightmare scenario is that children don't think anymore because they just get the chatbot to tell them the answers to questions.

00:42:36 Andrew Keen: Or to write for them.

00:42:39 Keith Teare: And then to execute on it. So the idea is that you get this passive human, leveraging AI to not have to do anything. And therefore, their brain will deteriorate because they'll never have to use it. Now that is not how AI works anymore. That might have been a little bit true two years ago when AI first came on the scene. But anyone that's properly using AI for any purpose now knows that you tell AI what you want. It will interact with you to improve your idea if you ask it to, or it'll just take your idea if you ask it to do that. It will then give you a plan and carry out the work derived from the plan. And then you'll judge what it gives you and you'll tweak it, and you're really in control. Not only are you not a passive victim, you're an active user with outcomes in mind. Yeah. So if schools understand that, they'll embrace AI.

00:43:44 Andrew Keen: I think what we'll get is a high end of Montessori, Waldorf educated kids of the new rich, of the new tech rich. Growing up without AI, it will create a parallel analog economy, the equivalent of vinyl records these days, but I'm not sure it's gonna subsidize an entire generation. It's weird that Americans don't trust AI and yet they're using it more and more. There was an interesting Wall Street Journal piece about how YouTube took over the American classroom. The increasing commercial success of YouTube is also driving Google's stock price. What's happening, Keith? Why on the one hand does nobody trust tech in America, and yet on the other hand, everyone's using it all the time?

00:44:33 Keith Teare: I think if you bifurcated that by age, what you'd find is that the younger people are completely embracing it. And it's the older people who are, ambiguous, and then there's a spectrum in the middle. I think among young people, and YouTube's a great example of that.

00:44:53 Andrew Keen: I'm a young person, and you're an old person, I think, in this equation.

00:44:57 Keith Teare: Yeah. But, you know, we act as we act.

00:45:01 Andrew Keen: You didn't get that joke.

00:45:02 Keith Teare: I did get that joke. I chose to move on and not react. But it is the case that young Americans are embracing AI more and more. I have three boys. I talked to all of them. I'd say the ones either in work or education, which is two of them, use it a lot. The one who is more creative uses it less at the moment, because it doesn't help him in what he wants to do as much as he would want it to. But they all use it a lot, and none of them have negative views about it.

00:45:38 Andrew Keen: Yeah. I mean, one wonders what this whole generation If these kids are going to school — and if they're growing up without AI, if they're going to their expensive private schools in Sonoma or San Francisco or Brooklyn or Washington DC, and they're not getting AI. They're never gonna forgive their parents. What are they gonna do? They'll sue their parents.

00:46:01 Keith Teare: Yeah. I think this is an enduring theme. You know, we have friends who didn't have TVs at home and wouldn't let their kids do video games. And when those kids come to our house, they actually get addicted more than kids for whom these are just —

00:46:14 Andrew Keen: They probably look at your screen and see "hand job," and they don't get that in their other Palo Alto homes, do they? Parents who put up —

00:46:24 Keith Teare: What's your bet, Andrew? When I do the voting on Substack, does that title survive, or does it get replaced with a different one?

00:46:31 Andrew Keen: What are the alternatives? What are the other two?

00:46:34 Keith Teare: I'll make them up. Maybe one of them will be flipping the subtitle and the title.

00:46:43 Andrew Keen: You mean job hand?

00:46:45 Keith Teare: No. Flipping around. The subtitle is —

00:46:47 Andrew Keen: The subtitle is "the end of typed and touched input." That's a bit boring. Yeah. Well, you'll have to come up with some alternatives. Anyway, this is the best you're gonna get as a hand job — to you That Was the Week and Keen On America viewers. Interesting conversation, Keith. Maybe we'll come back to this in future issues. I'm not sure we really exactly answered @Signal's question about what it means if a tiny group of big tech companies build the foundational mechanics of AGI. You seem to think that's democratic anyway, because the market is democratic, and so it reflects the popular will. I'm not so sure — but we'll come back to that. Good conversation. That was my hand job of the week from Keith Teare down in Palo Alto, the heart of hand job territory. Keith, have a good week, and we'll talk again next week.

00:47:47 Keith Teare: hands

00:47:48 Andrew Keen: Hands — for those people listening, Keith just showed us his beautiful hands. We'll see them again next week. Keith, keep well. Bye.