0:00:04 It seems to me right now you could do like a double-blind test of the same prompt given to
0:00:11 Grok, Claude, Gemini, Mistral, DeepSeek. I bet most people wouldn’t be able to tell which is which.
0:00:16 Benedict Evans is a technology analyst known for his insightful takes on platform shifts in the
0:00:23 tech industry. He sees AI differently than others. He’s spent decades spotting patterns others miss
0:00:29 and dives into how people really use AI. Why is it that somebody looks at this and gets it and goes
0:00:35 back every week, but only every week? The very high level threat to Google is that you have this
0:00:39 moment of discontinuity in which everybody resets their price that we consider as their defaults.
0:00:43 And so it’s no longer just the default that you go and use Google. There’s this sort of question for
0:00:49 Apple around does this net actually change the experience of what a smartphone is, what the
0:00:54 ecosystem is? Does it end up kind of getting Microsofted in the sense that…
0:01:07 I want to start with your most controversial take on AI.
0:01:13 It’s funny. My, I suppose my take on AI, controversial take on AI, rather like my controversial take on
0:01:21 crypto is being a centrist in that it seems to me very clear this is like the biggest thing since the
0:01:28 iPhone. But I also think it’s only the biggest thing since the iPhone. And there’s a bunch of people who
0:01:35 think, no, it’s much more than that. It’s a minimum. It’s more like computing. And then you’ve got people
0:01:39 going around saying, no, this is more like, you know, the electricity or the industrial revolution or,
0:01:45 you know, transhumanists or something. My sort of base case is to say, this is kind of another platform
0:01:50 shift and all the new stuff will be built around this for the next 10 or 15 years. And then there’ll be
0:01:56 something else. And so the impact on employment will be kind of like the impact on employment from the
0:02:01 other platform shifts and the impact on the economy and productivity and intellectual property. And
0:02:06 there’ll be, there’ll be a whole bunch of different weird new questions, just like there were a bunch of
0:02:11 different weird new questions before. And then in 10 years time, it’ll just be software.
0:02:17 Put this in historical context for us with other platform shifts. Everybody’s saying this time is
0:02:21 different, which everybody does at each platform shift, I would imagine. What’s the same?
0:02:28 Well, that’s the, there’s a, there’s a famous book about financial bubbles called This Time is
0:02:33 Different, because people always say this time is different. And it always is like the dotcom bubble was
0:02:39 different to like the late 80s. And the Japanese financial bubble was different to, you know, pick any other
0:02:45 bubble you want. They’re always different. But that doesn’t mean they’re not a bubble. And the same
0:02:53 thing here, I have a diagram I use a lot from 1995. This research firm made a diagram of something called,
0:02:59 they called cyberspace. Because it wasn’t clear it was just going to be the internet. It was clear that
0:03:05 everyone was going to have some kind of computer thing connected to some kind of network. But remember
0:03:10 the phrase information superhighway? Yeah. Which sort of conveys that it would be centralized and
0:03:14 controlled by cable companies and phone companies and media companies, which is sort of how everything
0:03:18 had always previously worked. It wasn’t clear, no, it was going to be the internet. It wasn’t clear the
0:03:21 internet was going to be kind of radically decentralized and permissionless and anyone could do what they
0:03:25 wanted. It wasn’t clear the internet was going to be the web. And only the web because there were all
0:03:30 these other things going on. If you look at Mary Meeker’s first big public internet report from 1995,
0:03:35 she has a separate forecast for web users and email users. And she thought email users would
0:03:40 be way bigger. It wasn’t clear like that was all one thing. And then it wasn’t clear that it was about
0:03:45 the browser. It wasn’t clear that the browser wasn’t where the value capture was because Microsoft
0:03:49 craybarred its way into dominance in browsers, but that turned out not to matter. And then all the value
0:03:55 was in site advertising and social, which were five years later and 10 years later. And so like,
0:03:59 you can be very, very clear that this is the thing and then still be completely unclear how it’s going to
0:04:03 work. The same thing with mobile internet. Just funny, mobile internet now, it’s kind of like
0:04:08 saying black and white television on color television, desktop internet, mobile internet,
0:04:13 black and white TV, color TV. No one really says mobile internet anymore. It’s like talking about
0:04:18 e-commerce. You’re starting to have people talk about physical retail and retail. And, but it wasn’t
0:04:23 clear, you know, I was a telecoms analyst in 2000 and it was very clear mobile internet was going to be a
0:04:29 thing. It was not clear that there would be basically small PCs. Like that was the fundamental shift
0:04:33 to the iPhone is it’s a small Mac. It’s not a phone with better UI. It’s a small Mac.
0:04:40 And it wasn’t clear that the telcos would get no value. It wasn’t clear Microsoft and Nokia would get
0:04:46 no value. It wasn’t clear it would take 10 years before it took off. And it wasn’t clear it would
0:04:50 replace the PC as the center of the tech industry. I mean, everyone was talking about, well, what’s a
0:04:55 mobile use case? What would you do? You’ll do some things on your mobile phone, but what? But obviously
0:04:58 your PC will be how you use the internet. And of course, that’s not how it worked.
0:05:04 And so we kind of forget because now we don’t see it because now it just kind of became part of the
0:05:08 air we breathe. How weird and strange and different all these things are. There’s something I love
0:05:16 talking about, which is, is the rise of automatic elevators. So until the fifties elevators were
0:05:20 manually operated, they were basically vertical streetcars. They were trams, they were pub trains.
0:05:25 And you have a driver who has a lever with an accelerator and a brake. If you’ve been into a New York co-op,
0:05:30 you may have seen one of these, but they call it an attended elevator. There’s a lever. You push it
0:05:34 that way to go down, middle to stop, that way to go up. And then in the fifties, Otis creates the
0:05:39 autotronic, I think it’s called the autotronic elevator, which had electronic politeness, which
0:05:43 basically meant the infrared thing that stops the door closing. But if you get into an elevator now,
0:05:48 you don’t say, oh, I’m going to use an automatic elevator with electronic politeness. It’s just lift.
0:05:53 We kind of forget how weird and different all the other things were. And yes, this is new and weird
0:05:58 and different in a bunch of kind of strange, confusing, confounding ways we can probably talk
0:06:02 about. But we sort of forget that other things were weird and strange and different too.
0:06:09 Is this the first major platform shift where the incumbents have an advantage because they have the
0:06:09 data?
0:06:17 I’m pretty sure people thought Microsoft had an advantage on the internet and Google and meta had
0:06:24 an advantage on mobile and everyone thought IBM was going to win PCs. Once IBM made a PC, that was it.
0:06:30 It’s all over now. And we kind of forget that there were PCs before and then IBM made one and that kind
0:06:32 of became the standard, but then IBM lost it.
0:06:38 So what happens with the incumbents? Do they grab on to using the technology instead of adopting it
0:06:43 because adopting it would mean killing the golden goose? Like what happens in a platform shift with
0:06:43 incumbents?
0:06:48 The master of my college at Cambridge said that history teaches us nothing except that something
0:06:55 will happen. And, you know, there’s always the example and the counter example. So with any new kind
0:07:02 new platform shift and a platform, the term platform shift itself is, you know, it’s a useful term, but you have to be
0:07:07 careful not to be trapped by your terminology and get into this sort of arguments about, well, is it a platform shift
0:07:14 or is it not a platform shift? And how do you define a platform? Shut up. Like, you know, the thing is when any
0:07:19 with any of these sort of fundamental technology changes, the incumbents always try and make it a feature and they try and
0:07:25 absorb it. And the same thing outside of technology, existing companies try and absorb it and they use it to
0:07:31 automate the stuff they’re already doing. And then over time you get new stuff, you unbundle both the
0:07:36 incumbents intake and you unbundle existing companies because of something that’s possible because of this new
0:07:41 technology. So you can always kind of jump into the new thing. And sometimes the new thing kind of really is just a
0:07:49 feature. And sometimes it’s no, it’s a fundamental change in how everything works. And sometimes that sort of
0:07:54 contingent, you know, there’s this whole sort of parlor game, like drinking game that historians play
0:07:58 about kind of historical inevitability, you know, well, what would have happened if that battle had
0:08:04 been lost or if that politician had been assassinated or not assassinated? And it depends. Sometimes the
0:08:09 answer is, well, no, nothing, then everything would have been completely different. And sometimes the
0:08:14 answer is, well, no, then, you know, what if Napoleon had won at Waterloo? Well, then he’d have lost
0:08:17 another battle six months later, like nothing would have changed. The whole environment had changed.
0:08:22 What if, you know, the revolution hadn’t happened in spring of 1917? Then it would have happened in
0:08:27 the summer or the autumn. Sometimes it’s like really clear. I mean, I always think the Kodak
0:08:30 example here is kind of interesting.
0:08:31 Tell me about it.
0:08:35 Because, you know, like, it’s like the cliche that people say, oh, Kodak had digital cameras and they
0:08:39 didn’t get it or they ignored it or they didn’t want to do it because it would destroy their business. But
0:08:43 then you go and look at it and that’s like, well, that was 1975. And the thing they had was the size
0:08:48 of a refrigerator. I know that was not a consumer product. And it took until the late 90s before the
0:08:53 technology was actually viable as a consumer product. So of course, like they didn’t do it in the 70s
0:08:59 because you couldn’t. What actually happens is once it starts happening, Kodak go all in on digital
0:09:04 cameras. At one point, they were the best selling digital camera vendor in the USA. And if you look at their
0:09:08 annual ports at the time, they think this is going to be great because they’re going to sell way more photo
0:09:13 printers. So they’re selling these inkjet photo printers. People are going to produce way more photos. So they’re
0:09:16 going to take way more, they’re going to print them all. Two things that screw Kodak. One of them is
0:09:22 smartphones. And you’ve got to argue that what actually screws Kodak is not the camera. It’s the social
0:09:28 media and it’s not printing anymore. Because that’s what killed. That’s one side. The other side
0:09:35 of it is that film was this high margin product where they had a bunch of unique intellectual
0:09:41 property. And digital cameras are a low margin commodity where they were competing with the
0:09:46 entire consumer electronics industry with no differentiation. And so even if they go, even at
0:09:51 the point being, even if you go all in into that market, it’s still a crappy market where you’ve got
0:09:56 no differentiation. So you can kind of, you know, you can, you can kind of put all of these things on the
0:10:00 table and shuffle them around and say, well, in hindsight, obviously BlackBrew was screwed.
0:10:04 And in hindsight, obviously Google is going to be able to make the jump. And in hindsight,
0:10:10 in hindsight, and yeah, maybe. Is there a parallel between the second point you made about Kodak and
0:10:16 Google today where, you know, they have a high margin in search business and a low margin AI business?
0:10:23 So I think it’s, I’d be nervous about knowing what the margins are in AI because we’ve had,
0:10:27 you know, depending on who you ask, like the price to get a given result has probably come down by two
0:10:33 orders of magnitude. But then that’s, that was the state of the art two years ago. And now there’s a
0:10:38 new thing, which is more extensive. And so there’s a little awful lot of kind of shifting planes and
0:10:42 shifting, you know, there’s, there’s a lot of, there’s a lot of algebra and all the variables and the
0:10:46 algebra are all changing at the moment. So it’s kind of hard to quite to know what that is. I think
0:10:58 the, the, the, the, the obvious Google threat right now is that Google shows you a bunch of links
0:11:05 and results and ideas, and those could now be solved in a different way. Well, let me kind of go back a
0:11:12 second. The very high level threat to Google is that you have this moment of discontinuity in which
0:11:18 everybody resets their prize and reconsides their defaults. And so it’s no longer just the default
0:11:23 that you go and use Google. And for this search or that search, like maybe Bing is 10% better on that
0:11:28 search. In fact, as we saw from the, the, the, the Google TAC trial last year, actually Bing is,
0:11:32 Google is still the best search engine by quite a long margin relative to the other traditional
0:11:39 search engines. But what we have now is like a reset of the playing field. And Google has a whole
0:11:45 bunch of advantages as to why they might win in that playing field. But there’s a reset both of
0:11:50 what the product is and, and, and how you sell it and your org structure around selling it. And do you
0:11:53 have the right politics and the right org structure to build that and the right incentives and internal
0:11:56 conflicts? And then the consumer behavior kind of gets reset as well.
0:12:02 My understanding of AI is that so much is data-driven. Like you have proprietary data sources, you have
0:12:07 better data as you can train better models. That gives you one of the key inputs.
0:12:12 So I’d actually, I think it’s actually the opposite, which is that everyone’s kind of using the same data,
0:12:20 which is you need such an enormous amount of generalized text that the amount that Google
0:12:26 has or that Meta has is not actually enough to move, to be a kind of a fundamental difference
0:12:27 in what you can train with.
0:12:34 So you don’t think like YouTube as a repository is an advantage for like a significant.
0:12:36 So it depends.
0:12:37 Push back, yeah.
0:12:42 So it depends. So the models that we’re training now, we’re training on text.
0:12:48 So that’s not really being trained on YouTube. We saw this lawsuit around book copyright with Meta
0:12:53 that they downloaded a torrent of pirated books. Because guess what? That’s, they don’t have
0:12:59 enough stark text and it’s not the right kind of text. They don’t have lots of prose. They’ve got
0:13:04 lots of short snippets of text. So I think the generality of LLMs is you just
0:13:08 need such an enormous amount of data that everyone kind of needs all the text that there
0:13:12 is. And all the text that there is, is kind of equally available to anyone.
0:13:16 So we’ve kind of like the data is a level playing field effectively.
0:13:22 Yes, because you need so much more. And it’s also not necessarily the kind of data that you
0:13:27 have. So obviously Google has, you know, an enormous repository of scrape data because
0:13:31 they read the web all the time. But anyone else with a billion dollars can go out and do
0:13:31 that.
0:13:33 Right.
0:13:35 Or you can go and download the common crawl from AWS.
0:13:41 How far away do you think we are from autonomous sort of AI making AI better?
0:13:49 So no human intervention, but AI sort of going out in the real world, getting feedback, adapting
0:13:51 itself and making itself better.
0:13:56 There is this sort of like people parody of like the foomph, like suddenly, magically, this
0:13:58 thing grows and becomes amazing and learns everything.
0:14:05 I don’t think we’re at that stage now. I don’t think anyone really knows when it would happen.
0:14:07 So it’s very sort of impressionistic.
0:14:14 I think another answer might be you kind of have to be very careful looking at headlines
0:14:18 and thinking like, what exactly is that telling me?
0:14:26 So Anthropic has done a bunch of things where they say like, the AI was threatening to blackmail
0:14:26 me.
0:14:28 Yeah, I saw that.
0:14:33 And you read the story and you think, okay, you asked what’s basically a story generating
0:14:41 machine. Please tell me a story of what you would do if in this situation where most people
0:14:47 would probably say X and the machine says probably X. And you say, my God, it said it would do X.
0:14:51 It’s like, well, yeah, it would blackmail.
0:14:52 It’s based on human behavior.
0:14:55 It would blackmail you. Okay, how would it do that?
0:14:55 Yeah.
0:14:59 What do you mean it would blackmail you? It’s kind of like the reductio ad absurdum of this
0:15:05 is you write, this is a point somebody else made, that like you write murder is good on a piece
0:15:10 of paper and you put it in a photocopio and you press go and you say, my God, the machine
0:15:15 says murder is good. Well, no, you told the machine to say that. And that’s what these Anthropic
0:15:20 studies are. They’re basically, you tell the machine to say a thing and then it says it.
0:15:23 Like, well, you haven’t proved anything.
0:15:29 You know, people talk a lot about product market fit, sales tactics or pricing strategy.
0:15:34 The truth is success in selling often comes down to something much simpler, the system behind
0:15:40 the sale. That’s why I use and love ShopPay because nobody does selling better than Shopify.
0:15:47 They built the number one checkout on the planet. And with ShopPay, businesses see up to 50% higher
0:15:53 conversions. That’s not a rounding error. That’s a game changer. Attention is scarce. Shopify helps you
0:15:58 capture it and convert it. If you’re building a serious business, your commerce platform needs to
0:16:05 meet your customers wherever they are on your site, in store, in their feed or right inside their inbox.
0:16:11 The less they think, the more they buy. Businesses that sell more sell on Shopify. If you’re serious
0:16:16 about selling, the tech behind the scenes matters as much as the product. Upgrade your business and get
0:16:23 the same checkout that I use. Sign up for your $1 per month trial period at shopify.com slash knowledge
0:16:30 project. All lowercase. Go to shopify.com slash knowledge project. Upgrade your selling today.
0:16:33 Shopify.com slash knowledge project.
0:16:39 Do you ever struggle to stay focused? There’s a reason I reach for my remarkable paper pro when I need
0:16:44 to think clearly. If you’re looking for something that can help you really hone in on your work without
0:16:49 all the distractions. Remarkable, the paper tablet might just be what you’re looking for.
0:16:55 Remarkable just released their third generation paper tablet, Remarkable Paper Pro. It’s thin,
0:17:01 minimalist, and feels just like writing on paper, but comes with powerful digital features such as
0:17:07 handwriting conversion, built-in reading light, productivity templates, and more. It’s not just another
0:17:15 device. It’s a subtractive tool. No notifications, no inbox, no apps, just you, your ideas, and a blank
0:17:20 page that feels like paper, but smarter. Whether you’re in a meeting or deep in a creative session,
0:17:25 this device is built to keep you in the zone. In a world built to steal your focus, this tablet gives
0:17:31 it back. It’s become the place I do my deepest work, and it travels light, slim enough for any bag,
0:17:37 powerful enough for any boardroom. Not sure if it’s for you? No worries. You can try Remarkable
0:17:42 Paper Pro for up to 100 days with a satisfaction guarantee. If it’s not the game changer you were
0:17:48 hoping for, you’ll get your money back. Get your paper tablet at Remarkable.com today.
0:17:51 Where do you stand on regulation of AI?
0:17:58 So I think regulation of AI is sort of the wrong level of abstraction. Talking about regulating AI
0:18:03 as AI is the wrong level of abstraction. It’s like saying we’re going to regulate databases,
0:18:09 regulate spreadsheets, or regulate cars. Well, we do, but not like that. When you regulate stuff,
0:18:14 there are trade-offs. You learn about this in your first year in economics class, like regulation has
0:18:21 costs and consequences, and it’s not necessarily, you know, there’s always a trade-off. And often you’re
0:18:26 making product decisions or engineering decisions that do actually have trade-offs. There’s like a
0:18:29 three-way trade-off of like what’s good for the product, what’s good for the consumer,
0:18:33 what’s good for competition, what’s good for the company, what’s good for the consumer.
0:18:39 I think the regulatory stuff is interesting in the framework of multiple countries sort of competing
0:18:45 for superintelligence. How would you advise a country to prepare for AI? If I’m the president
0:18:52 of the United States and I call you and I say, Benedict, you have five minutes, I need, what do I need to
0:18:55 prepare for? What can I do to put our country in the best position possible?
0:19:01 Well, what’s your objective? Is your objective to have a nice press release?
0:19:03 No, it’s to dominate AI.
0:19:08 A long time ago, I used to get these questions about like, how can we replicate Silicon Valley?
0:19:14 And I always feel like the answers to those questions as much as possible is like, you can’t,
0:19:18 I mean, occasionally there are things you can do, like you can create funding structures,
0:19:22 you can make it easy, you can, you know, you can try and jumpstart startup ecosystems,
0:19:26 you can try and jumpstart funding availability. But most of the answer is things like getting
0:19:32 out of the way. I think that the idea of trying to create national champions is very hard. Now,
0:19:37 that almost kind of becomes an economist’s question rather than a technology question. How do you create
0:19:41 national champions? Where does that work? Where does that not work? I’m sure there’s a bunch of
0:19:46 books and papers about, you know, where does industrial policy work? Where does it not work?
0:19:52 From a technology analyst perspective, I think of this in terms of, A, what are you doing that would
0:19:56 make this harder? And B, think of this as just more startups.
0:20:02 What are the things that we’re doing that make it harder to develop that ecosystem without picking
0:20:05 a winner? It’s not about picking a company and backing them.
0:20:12 Well, if you do like this ridiculous law that California had a year or two ago, if you treat
0:20:18 this as like nuclear weapons, and you say this is incredibly dangerous, and we need to have it under
0:20:21 extremely tight control so that nobody does anything bad with it.
0:20:23 Which is basically the EU approach.
0:20:30 To go back to your economics class, policies have trade-offs. To govern is to choose. You’re
0:20:34 making a choice when you do that, and you’re choosing that has costs. Personally, like most
0:20:38 people in tech, I think the idea that this is all going to kind of produce bioweapons and take over
0:20:42 the world and kill us all is just idiotic. Like, I think it’s just a bunch of kind of childish
0:20:46 logical fallacies within that. But you have to be conscious. You say, if we’re going, you know,
0:20:53 the kind of the Biden approach to generative AI very explicitly was to say, this is sort of social
0:20:59 media 2.0. Like, social media to one was terrible and destructive and bad. And I think there’s a,
0:21:04 I don’t agree with that. I think there’s a huge dose of moral panic within that. But be that as it may,
0:21:09 if you make a decision that says we are deliberately and explicitly going to make it really hard to build
0:21:14 models, and really hard to start a company that builds models, and really hard to do anything with
0:21:19 any of this stuff, then guess what? It’s kind of like, you know, the mayoral election in New York
0:21:24 today. Like, if you make it really hard and expensive to build houses, houses will be more
0:21:30 expensive. You’ve made that choice. If you do that, you cannot then complain that houses are more
0:21:35 expensive. You can choose that, but you can’t complain. Why do you think as a society, we don’t
0:21:42 understand that? Part of this is that, like, in most non-emotive fields, we kind of do. You understand,
0:21:47 people understand that, you know, if you make, you know, more employment regulation tends to produce
0:21:51 slower growth, but more protection for employment, and you’re choosing a trade-off. You know, people
0:21:55 kind of, I think everybody on both sides of that equation understands that, that that’s a trade-off,
0:21:59 and you’re choosing one versus the other. The point is, you can have a fully functioning free market,
0:22:06 and you can regulate some of the negative externalities of free markets that anybody in any part of the
0:22:09 economic spectrum understands there are negative externalities to free markets. You can also have,
0:22:13 like, a government-provided alternative. You can have the government do the fire department.
0:22:20 Where you have some of the kind of most obvious gaps between the U.S. and Europe, it seems to me
0:22:26 sometimes, are in places where you kind of have neither. So, the U.S. neither has a government-controlled
0:22:32 healthcare system, nor a free market healthcare system. Do you see what I mean? You have neither
0:22:36 government-controlled housing, which you have in, like, in weird places like Singapore,
0:22:41 nor a free market in housing. So, you kind of break the free market. So, you stop the price
0:22:45 signaling. This is like the great insight of Hayek is that pricing is a signal. Pricing is an
0:22:50 information system. It’s telling people what’s wanted. It’s not just a signal of worth. It’s a
0:22:55 signal of demand. There’s a fascinating book I read a while ago called Red Plenty, which is about
0:23:02 Soviet central planning in the 60s, 70s, 80s. And it’s about sort of what happens when you have a central
0:23:07 planning that just cannot cope with the level of complexity of a sophisticated economy in the 60s
0:23:13 and 70s, as opposed to let’s make grain and tractors and locomotives in the 20s and 30s and steel, which
0:23:17 really kind of works. But once you actually have a sophisticated industrial economy, central planning
0:23:22 can’t handle the complexity. And so, you try and create incentives and structures around that while
0:23:27 not having pricing. And that just doesn’t work. I suppose there’s a sort of a generalized point,
0:23:32 which is like a market economy is a system. And if you pull a lever here, something will move
0:23:37 there. And you can’t just pull a lever here and say, well, I don’t want that to move because it’s a
0:23:43 democracy. It will move anyway. And so, you have to understand how the system works and understand what
0:23:47 consequences you want from that and what your parameters are within this.
0:23:52 One of the things that I admire about you is that you’re sort of known for spotting patterns.
0:23:58 I have a theory on how to learn pattern matching. And I’d love to hear your pushback on this.
0:24:02 My theory on how we learn is, I call it the learning loop. We have an experience.
0:24:07 We reflect on that experience and we create a compression. And that compression becomes our
0:24:13 takeaway. So, we can watch a movie, read a book, and you come away with a compression of it. But you can
0:24:19 work backwards from that compression to the experience. But what we consume most of the time is other
0:24:24 people’s compressions. So, like when people read your newsletter, they’re consuming a compression
0:24:30 of the work that you’ve done, but not the actual raw work. So, in a way, it’s an illusion of knowledge
0:24:33 if you haven’t done the work in that area.
0:24:41 It’s funny. I have a draft thinking about like what LLMs do to web search and publishing and
0:24:48 discovery in e-commerce and like big foot of hand wavy fuzzy, all of that stuff. And I was sort of
0:24:52 thinking about this and there’s a book written by a French academic sort of 20 years ago or something
0:25:02 called How to Talk About Books You Haven’t Read, which sounds very kind of snide. But kind of his
0:25:06 point is that like there’s the book you read when you were 17 and you really didn’t get it.
0:25:11 And if you read it now, you’d get it. And there’s the book that like, he’s got this kind of list of
0:25:14 like, there’s the books that everybody else has read. So, it’s to say you’ve read them. There’s the
0:25:17 books that like you’ve read three other books by that writer. So, you don’t really need to read this
0:25:21 one too. You get it. Like, do you need, you know, do you need to read another Malcolm Gladwell book?
0:25:25 You’ve kind of got the Malcolm Gladwell experience. So, there’s this sort of generalized sense of
0:25:32 pattern and accumulation of what you’ve seen, what you’ve half seen, what you half remember.
0:25:40 There’s also, I think, you know, what your viewers, listeners might notice is I kind of have two modes,
0:25:45 two or three modes. I have a mode that’s sort of discursive and slightly rambling and
0:25:48 free associating and I’ll kind of spiral off in different directions and hopefully come back to
0:25:56 the point. And then there’s a mode where I want to try and pin the thing down and break it apart and
0:26:01 say, what are the two, three, four things that are happening here? Which is what you see in the slides
0:26:08 is no, like, what is it? It’s this and then this and then that. And as that capturing, that’s a way of
0:26:14 trying to understand what this is. I try and work out what I think about this, how I understand it,
0:26:19 how you can break it apart by kind of pinning it down. The thing about data is, and the thing about
0:26:26 the slides and the analysis is, is like, I’m always asking who cares. And I’m always asking,
0:26:32 yes, but what actually matters here? Why are you showing me this slide? Why am I showing you this
0:26:36 chart? And so you kind of have to ask, like, well, what are the actual questions?
0:26:40 What are the questions we’re not asking on AI that we should be asking?
0:26:43 I mean, we’re asking, okay, well, there are some people who are saying, oh, all the value
0:26:48 capture is going to be in the models. There’s this kind of funny split between people who are just
0:26:52 talking about the models getting better and everybody else who’s saying, well, all the value is going to
0:26:58 be in the application layer. And, you know, what are the companies? And let’s fund Cursor and let’s
0:27:02 fund all of this stuff. And why isn’t there a consumer breakout yet? And other people are saying,
0:27:05 what do you mean there isn’t a consumer breakout? Everyone’s using chat GPT, which answers, well,
0:27:10 not exactly, which is my data point. Some people are using chat GPT. Most people look at it and don’t
0:27:15 get it still, which is fascinating. There’s this kind of core where’s the value capture question.
0:27:19 Then there’s like a bunch of questions we could have asked two years ago where we don’t have an
0:27:24 answer. Will the error rate ever be controllable or manageable? Will you ever get to a model that knows
0:27:29 when it’s wrong? Which to me seems like, given the statistical system, seems like a contradiction in terms.
0:27:33 But maybe it was a bunch of kind of, you could make a list of like a dozen questions we could have asked
0:27:39 in early 23. We don’t really have answers to any of those. I mean, there were some people who were asking,
0:27:43 are these things commodities? Will China catch up to a true answer? Even then was obviously yes, of course,
0:27:49 which is what happened, which Deepsea kind of demonstrated. But we don’t have that many new
0:27:55 questions since then. The thing that I puzzle about right now is, first of all, there’s this whole
0:28:02 nexus, as I said, of like, what is SEO for LLMs? You know, we have infinite product, infinite retail,
0:28:06 infinite media. How will you choose what to buy? What happens if I go to an LLM and say, what mattress
0:28:13 should I buy? What life insurance should I get? How does that work? That poses dozens of questions we
0:28:20 don’t know yet. Then there’s a question around like, the differentiation in the LLMs as product.
0:28:26 Like it seems to me right now, you could do like a double blind test of the same prompt given to
0:28:35 Grok, Claude, Gemini, Mistral, Deepseek, do a double blind test. I bet most people wouldn’t be able to tell
0:28:40 which is which. That question of like, is there product differentiation? Can there be product
0:28:46 differentiation around the LLM as consumer product? Because right now the models are commodities, but
0:28:52 ChatGPT is way, way, way, way, way more usage. So ChatGPT is at the top of the outstore rank. Gemini
0:28:58 bubbles between like 50 and 100. None of the others were in the top 100. Same in Google Trends, same in
0:29:02 the usage numbers, same in the revenue. There’s revenue for corporate APIs. Corporate is a whole other
0:29:07 story. But as a consumer thing, it’s like ChatGPT is now like the brand. It’s the default. It’s the
0:29:12 Google. You use it because you’ve heard of it. And none of the others have broken it. Is that where
0:29:17 we are now? But then if you look at the products, the products, not any of the models are all the
0:29:21 same. The underlying model is the same. The product’s all the same. It’s really hard to tell the
0:29:23 difference except like they’ve got different color schemes and different icons.
0:29:24 Different branding.
0:29:28 Different branding. But the products are all the same. And this reminded me of looking at
0:29:33 browsers. In that browsers are all the same. The rendering engine underneath might be
0:29:37 different, just as the LLM might be different. But you’ve got an app, an input box, an output
0:29:42 box. And the output box wonders what the rendering engine gives you. And the only innovation in
0:29:46 browsers in the last 25 years is basically tabs and merging search in the address part.
0:29:47 Yeah.
0:29:51 And there’s like the browser project. There’s people trying to do it now, but it hasn’t worked.
0:29:54 It hasn’t got traction. And is that sort of how LLMs will work in that it’s about the
0:29:59 distribution and the brand is not actually about the product or the model? Or is it maybe more
0:30:06 like social in that, yeah, photo showing is a commodity, but you, there’s a big difference
0:30:11 between Instagram and Flickr and all the other people that try to do photo showing. And so
0:30:12 you have to really…
0:30:16 That would almost be an argument that it’s sort of winner take all, right? It’s very hard
0:30:22 for, like use Claude as an example. It’d be very hard for Claude to compete if they don’t
0:30:27 have enough usage to continuously make the investment.
0:30:33 Well, that’s a slight, this is a slightly different thing. So the winner, there doesn’t
0:30:38 appear to be a sort of self-reinforcing cycle in which more people use it because more people
0:30:42 use it. The product gets better because more people use it. So more people use it, which
0:30:46 is what you have with operating systems because you have more apps, therefore more users, therefore
0:30:50 more apps. So what you have with Google search, that Google has all the feedback from how people
0:30:54 use it that makes the search engine better. You have a network effect in social media that
0:30:56 you’re there because you’re there, because your friends are there, because you’re there.
0:31:02 There’s no apparent equivalent in LLMs right now. There’s no reason why the LLMs get better
0:31:08 because more people use them. Now that may come. You have that, OpenAI and people have been
0:31:13 doing memory where it remembers what else you’ve asked, but that seems more like a switching
0:31:17 cost than a network effect. And also it might be easier for you to just ask it what it knows
0:31:23 about you and then tell Claude or vice versa. So it’s not clear if that. But we are at that
0:31:26 sort of stage where you’re looking at the browser and saying, is there a way that you can create
0:31:30 stickiness here or that you can create a network effect on the browser? Or is it just that the browser
0:31:37 itself is a commodity? Now, capital is not a winner-takes-all effect in that convention. At any
0:31:41 rate, it’s a different kind of winner-takes-all effect. I mean, I wouldn’t conventionally think of
0:31:45 capital as a network effect. Or is it, it’s not a product that’s in, it’s not something that’s
0:31:52 inherent in the product. It’s something else. It may be that, yes, ChatGP, the OpenAI has more money
0:31:53 so they can make their model better.
0:31:58 There’s like six rabbit holes I want to go down before we move on to something new. If OpenAI can,
0:32:03 I liked your point about sort of at the point where AI gets better because people are using it,
0:32:05 then there’s a huge advantage to being OpenAI.
0:32:08 We don’t have any, we don’t have visibility on what that would be yet.
0:32:12 At that point, though, whoever’s in the lead would sort of…
0:32:17 If it did, then you could get kind of a runaway. But we should kind of go back and think about
0:32:24 MySpace. Because you have like, and it would be MySpace, you know, in the early phases of these
0:32:29 things, and you see the same thing with the early PC industry, you’ve got a dozen of them. And there’s
0:32:34 often an early leader that falls away later. And so MySpace was the early leader that fell away
0:32:39 later. Then you get a late stage where the S-curve is kind of flattened out, where all the network
0:32:43 effects have kind of solidified and the product quality has solidified. It was very easy actually
0:32:48 to get people to switch back and forth between MySpace and Facebook and Bebo and Friends Reunited
0:32:54 and whatever the, Orkut and all these other things. In the early days, then you kind of get this
0:33:00 separation out. But then, of course, then you get, then Instagram comes along. And then TikTok
0:33:04 comes along. Right. And so as soon as you have something that’s a different proposition, that
0:33:10 turned out to be extremely easy to pull that away. You know, Google lost to YouTube, they had to buy
0:33:15 YouTube, Facebook lost to Instagram and WhatsApp, and they had to buy them both. So those are quite
0:33:21 fragile and quite narrow when it takes all effects, or at least they appear to be. We don’t know what
0:33:25 that would be or what the modalities would look like. Modalities, sorry, that’s a great meaningless
0:33:29 word. It’s like saying societal. We don’t know what that would look like. And therefore we can’t,
0:33:33 we don’t know how rigid it would be or how it would work because we don’t have it yet.
0:33:37 You can’t, as I’m sure you know, you’re not, they’re not retraining the models all the time
0:33:42 with the data. So you don’t have that kind of a runaway effect as like continuous flow and more
0:33:47 queries produces, you know, better results. So it’s kind of tricky to do that yet.
0:33:52 I want to come back to something you said. You said some people look at ChatTBT and don’t get it.
0:33:56 Yeah, I think this is really important. There’s a whole bunch of survey data on how people,
0:34:00 how many people are using this stuff. You’ve got the numbers from, so, so, so, so Chat OpenAI
0:34:05 pretty say, well, we’ve got this many weekly active users. Funny thing about social is people,
0:34:08 when social happened, people would talk about registered users. You know, in the early days
0:34:13 of the internet, people would talk about hits. Yeah. And then we realized that if your web page
0:34:17 has seven items in the menu bar, that’s seven gifts. So that’s seven hits. So hits was meaningless
0:34:20 and you had to switch to page impressions and then it’s registered users. And then it’s monthly
0:34:23 active users. And on social people said, well, hang on, if you’re using Instagram once a month,
0:34:27 you’re not using it. It’s daily active users or nothing. And weekly active users, we don’t like
0:34:32 either. Now OpenAI is doing weekly active users. And Sam Waltman was a social media startup founder.
0:34:36 He knows this. It’s a bullshit number. You look at survey data, and I did this slide in the last
0:34:40 presentation I did of like five different surveys from the US from late last year, earlier this year.
0:34:47 And it’s all roughly the same. It’s like something around 10%, give or take three or four percent of
0:34:52 people, depending on the survey, they’re using this, say they’re using this every day. Another
0:34:59 sort of 15 to 20% of people say they’re using it every week. And then there’s, say you’ve got like,
0:35:06 say 10% of people using it every day, say 15 or 20% of people using it every week. Another 20 or 30%
0:35:11 of people who say I use it every month or two. And another 20 or 30% of people who said, yeah,
0:35:16 I had a look, I didn’t get it. And then you have this survey where people say 70% of people are using
0:35:20 AI. And I’m like, wait, what do you mean? There’s a whole other rabbit hole, which is, you know,
0:35:25 people say, well, did you use Snapchat’s face filters? Oh, so you’re using AI. What do we mean by AI?
0:35:29 So which is, let’s be specific. Let’s talk about, are you using a consumer facing LLM chatbot?
0:35:33 Like you’re going to chat to UPT or Claude and asking questions. Like that’s the number we want to look at.
0:35:40 Most people don’t think about their inbox as a system, but I do. Email used to be the thing that
0:35:44 helped me run my business. Lately, it felt like the thing getting in the way of it. I’d spend too
0:35:49 much time weeding through low priority messages, trying not to miss the one or two that actually
0:35:55 mattered. And it was draining my focus. Then I started using Notion Mail and everything changed.
0:36:01 Notion Mail is the inbox that thinks like you. It’s automated, personalized, and flexible to finally
0:36:06 work the way that you were. With AI that learns what matters to you, it can organize your inbox,
0:36:12 label messages, draft replies, and even schedule meetings. No manual sorting required. I’ve created
0:36:18 custom views that split my inbox by urgency and topic so I can focus without distraction. And I use
0:36:24 snippets to fire off my most common emails, follow-ups, intros, and scheduling without rewriting
0:36:29 anything. The best part is it works seamlessly with my Notion workspace and is powered by Notion,
0:36:35 the tool trusted by over half of Fortune 500 companies. Notion is known for powerful connectivity,
0:36:41 intuitive functionality, and the ability to supercharge productivity. Get Notion Mail for
0:36:46 free right now at notion.com slash knowledge project and try the inbox that thinks like you.
0:36:51 That’s all lowercase letters, notion.com slash knowledge project to get Notion Mail for free right
0:36:57 now. When you use our link, you’re supporting our show too. Notion.com slash knowledge project.
0:37:09 And to me, there’s a bunch of you could matrixes. So some of this is it’s early. There’s a counterpoint
0:37:13 here, which if people do the chart and they say, oh my God, it’s so fast. It’s like it’s faster than
0:37:17 smartphones. Yes, because you didn’t need to buy a thousand dollar smartphone. Right. It’s faster than
0:37:22 PCs. Yes, because you know what PCs cost in the 80s adjusted for inflation? It’s like five grand.
0:37:27 It’s free for a lot of them. It’s free. It’s a website. You just go there. Of course,
0:37:31 it’s got faster adoption. And there’s way more people online as well. So even the absolute numbers
0:37:35 are faster than they were for Facebook 20 years ago, 15 years ago, because there’s way more people
0:37:40 online now. Yes. So that’s again an example of my unfair but relevant comparison. You’re sort of
0:37:44 standing on the shoulders of giants. So of course you can get to way more people quicker. But do you have
0:37:51 to keep asking, well, what? Yes, but why do so many more people look at this and not get it? Or even
0:37:57 worse, the not getting it, I can kind of see because people look at everything and don’t get it. Why is
0:38:05 it that somebody looks at this and gets it and goes back every week, but only every week? Right. Why is it
0:38:11 they can only think of something to do with this once a week? I worry about those people. I mean, I’m just
0:38:18 thinking if these numbers are accurate, the 10%, the 15%, you know, 90% of the people that I spend the
0:38:23 most time with are within that 10%. Well, I’m not. Interesting. Tell me more about that. Well,
0:38:29 here, actually, I’ll preface this conversation with my kids don’t use Google anymore. They use it to
0:38:37 find phone numbers or local businesses or places distance. Everything else, they basically have
0:38:43 defaulted to chat GPT now. Again, I’m going to, all AI conversations seem to be analogies. So I’ll-
0:38:49 100%. And it’s like, it’s like nuclear weapons. It’s like, no, it’s not. The comparison, I think,
0:38:53 is interesting here. It was not perfect, but it’s interesting is to look at early spreadsheets
0:38:58 for software spreadsheets, spreadsheets with paper. Dan Bricklin creates, and I can’t remember the
0:39:03 other guy’s name, who created VisiCalc in the late 70s. And I think to get an Apple II to run it with
0:39:08 a screen and everything costs like 15 grand adjusted for inflation. And you show this to
0:39:13 an accountant and it’s like, you can change the interest rate here and all the other numbers
0:39:20 change. And we see that now and we’re like, yes, 1978, that was a week of work. Almost literally.
0:39:21 That was like amazing.
0:39:27 Yeah. He would do a week of work in half an hour or less. And he has all these stories about
0:39:32 accountants who would, you know, they would be given a one month project and they’d get it done
0:39:35 in a week and then they’d like go and play golf for three weeks because they, partly because they
0:39:38 could probably, they didn’t actually want to tell the client I needed it in a week because the client
0:39:43 would think they hadn’t done it properly. So I look at ChatGPT and I think, right, I don’t write code.
0:39:48 I have zero use of something that will write code for me. I don’t really do brainstorming.
0:39:57 I don’t do summarization of things. I don’t do the things where it’s sort of out of the box,
0:40:03 easy and obvious. And then there’s a sort of mental load of, okay, I’ve got to kind of try
0:40:09 and think of what things am I doing that it could do for me. And that most people don’t think like
0:40:13 that. So there’s a sort of, I said a memory ago, there’s like a matrix. There’s a matrix of like,
0:40:20 who has the kinds of tasks that it’s good at? Obviously. Who is good at, has the kind of tasks
0:40:26 that it’s good at? Not obviously. Who is good at thinking about new tools for the things that
0:40:32 they’re doing? Who isn’t? If you, I’m kind of blown away. You don’t reflexively use AI.
0:40:39 If you were using Salesforce and you had a button that said, drove me an email to reply to this
0:40:45 client, then that gets massive adoption. Well, that’s a feature that we talked about earlier.
0:40:51 But that’s a different thing. Yeah. Is it, is it the chatbot as product where you get this blank
0:40:56 screen and you kind of look at it and you scratch your head and you have to think, well, what is it
0:41:01 that I would do with this? And then you have to form new habits around it. Or is it that it’s
0:41:06 wrapped in product and UI where somebody else has said, it would be really useful for this,
0:41:09 wouldn’t it? And then you look at it and go, oh yeah, that would, I could do that.
0:41:13 Do you think it’s better with qualitative or quantitative analysis?
0:41:22 I think it is presently, and I’m going to get a binary statement. I think today it has zero value for
0:41:28 quantitative analysis. Oh, interesting. Because if, well, let me, let me qualify that. Do the numbers
0:41:33 need to be right or roughly right? Because what all of these things do is they give you something
0:41:39 that’s roughly right. And roughly is a spectrum, but it’s always… You don’t want pi to be 3.1.
0:41:42 Depends how big the circle you’re measuring is.
0:41:49 You know, this is the line about pi that, you know, we can calculate it. And however many digits we have
0:41:54 is like enough to calculate, you know, the diameter of the universe or something, but people still adding
0:41:57 more numbers. So, you know, there’s a little bit of Xeno’s paradox in here. You know, like you get
0:42:01 infinitely close. At a certain point, it doesn’t matter. And this is at a high level, this is some of
0:42:06 the AGI argument that if the thing gets infinitely close to reasoning, without ever actually reasoning,
0:42:10 does it matter? Like at a certain point, the thing, if the thing is always right, if the thing is only
0:42:15 wrong once in a billion years, does it matter that it’s not, that it’s not always right? The problem
0:42:19 today is, it’s not wrong once in a billion years. It’s wrong a dozen times a page.
0:42:21 You don’t want to spit that out and give it to somebody.
0:42:22 And I don’t know.
0:42:23 Yeah, yeah.
0:42:28 So, I had a very early example of this. And I was going to speak at an event at the beginning of 2023.
0:42:32 And the conference people had asked me for a long biography of myself. And I didn’t have,
0:42:37 I don’t have, still don’t have one. And so they’d made one and they’d used it with chat TPT and not
0:42:39 told me. And they just sent it to me to check. And I looked at them and I said, what the fuck is
0:42:40 this bullshit?
0:42:44 That’s 2020. That’s like, but, but generations ago.
0:42:48 That’s not relevant to the point I’m going to make. The point I’m going to make is,
0:42:53 A, it was always the right kind of biography. It was the right kind of degree, the right kind of
0:42:56 university, the right kind of experience, the right kind of jobs. It just wasn’t actually the right things.
0:43:04 But B, I could take that and fix it. So for them, it was useless. For me, it was completely,
0:43:08 very useful. I could spend 30 seconds fixing it instead of spending an hour scratching my head,
0:43:14 which is why I say right or wrong depends, which is a very kind of French philosopher kind of
0:43:19 question. Is the answer, is the prompt, does it, does the prompt have errors? It kind of depends
0:43:27 why you wanted it. Okay. Well that, I don’t have use cases where I want something that’s roughly right.
0:43:33 I don’t have use cases where I want a list of 10 ideas or I want it to brainstorm or I want it to
0:43:38 draw for me an email or want it to write code or I don’t want it to generate some images. You know,
0:43:44 my friend who works at a consultancy and they want pencil sketches of concepts and now they can
0:43:49 just use Vigione to make those. That’s great. Does that sketch, like, does that person at the back
0:43:54 have three legs? Not anymore. No. And if they did, it wouldn’t matter. You could Photoshop that out.
0:44:01 But I don’t do that. I don’t create images. So I don’t have a good mapping of the stuff this is good
0:44:08 for early against the stuff that I do. And the stuff that it, that it maybe would be useful for
0:44:15 is the stuff where it’s actually not yet very good. And the things where you would mitigate that by
0:44:19 saying, well, I would fix it. I don’t do those things.
0:44:24 Okay. So this is a good thing because I wanted to come back to something you said. You said you
0:44:32 think by writing and in a world where you’re taking something generated by AI and editing it,
0:44:35 that’s different than writing. Talk to me about thinking by writing.
0:44:42 I, why she, my, my, my chat GPT use case, which is more mental model than a practical thing is I
0:44:46 write something. And I think it’s actually, I would, what I always would ask in the past is kind of your
0:44:51 point about pattern recognition is I look at something and say, am I adding value here? Am I
0:44:55 saying something useful? Am I saying something different? Am I asking the key question? Am I
0:45:00 pushing further? Am I pushing the, am I asking the next question rather than just answering the obvious
0:45:05 questions? Now I can just say, is this what chat GPT would have said? And if the answer is,
0:45:09 this is what chat GPT would have said, then I didn’t publish it. Not because people can get it
0:45:15 from chat GPT, but because anyone would have said that. That’s a perfect analysis in the sense that
0:45:24 it raises the baseline of what qualifies as insight. The difference is the slope of the insights. And
0:45:31 so you wouldn’t say it if chat GPT is going to say, and push back on this by all means, but the chat GPT
0:45:34 level of insight to use an example, it could be clode, it could be
0:45:43 GROC could be any of them is increasing at a faster pace than most people. And eventually those slopes
0:45:47 intersect and it’s probably intercepted already or intersected already with, you know, maybe up to
0:45:54 intern level. And next year it might be master’s level, or it might even far surpass that. And it
0:45:59 hasn’t some domains in terms of math, um, the year after it might. And so maybe it’s like five years
0:46:06 before it, it passes Benedict. Um, and maybe it’s four years before it passes somebody else. And maybe
0:46:12 it’s like passed me a way long time ago. So I think there’s a, there’s, there’s two or three
0:46:19 directions we could, we could take that. One of them is, it’s an interesting theoretical philosophical
0:46:27 question is originality, which is to say alpha go could do original moves because it could do all
0:46:31 the moves and do it, do moves that no one had done before, not knowing what people had done before,
0:46:34 but it had an external scoring system. It knew that that move was good.
0:46:36 Mm-hmm. The, because it had feedback.
0:46:41 Yeah, it had a feedback loop because every move has a score. It can evaluate the score every move.
0:46:47 The classic, um, parable of the monkeys and typewriters, or, you know, the Borgias infinite library is there’s no
0:46:48 feedback loop. Yeah.
0:46:52 Yes, the Borgias infinite library contains new masterpieces generated at random, or the monkeys
0:46:55 with typewriters would generate new masterpieces, but there’s no feedback loop. So there’s no way of knowing.
0:46:59 You’ll see this with music now. You can generate new music. It could generate new stuff that you
0:47:06 wouldn’t know. For an LLM, variance is bad. Originality is, is, is a lower score. So what’s
0:47:13 the feedback loop for original but good? Now, it might be that that’s the same sort of false
0:47:19 question as saying, is it really reasoning or is it just right? 99 times, you know, nine, nine,
0:47:23 nine followed by many zeros. Does it actually understand or is it just always right without
0:47:27 understanding? Does it actually know that’s original and different or does it not? And that’s a,
0:47:31 that’s a kind of a puzzling, I don’t think we know the answer to that. And it may be the wrong
0:47:37 question, but it’s kind of a puzzle as to how would these, how far can these things make things
0:47:44 that are both different from the training data and good? And is knowing that this is different
0:47:50 but good, is it really different or is it just matching the pattern on a longer frequency?
0:47:55 Do you see what I mean? And how much could you actually have predicted that given enough data
0:48:00 that it’s not actually outside the pattern? It just kind of looks like it is if you’re zoomed in more.
0:48:04 And if you zoom out more, then that is matching the pattern. How would you know that people will,
0:48:08 it’s like, you know, thinking about music, like, how would you know that people would like punk?
0:48:14 You could get, you can very easily imagine a generative AI system that can make you more
0:48:18 stuff that sounds like, yes, or more stuff that sounds like Pink Floyd. It might not sound like
0:48:22 good Pink Floyd for Floyd, but you could imagine it would make more stuff that sounds like,
0:48:26 like the Grateful Dead. You know what Grateful Dead fans say when they run out of drugs?
0:48:28 This music’s terrible.
0:48:37 You can imagine, I’m being unkind, but like you can imagine the challenge is knowing that now people
0:48:44 are really fed up of 70s prog rock and they would really like something else and that something else
0:48:49 would be punk and that would work. It’s knowing that people in the 40s were really fed up of the war
0:48:54 and would want luxury and that Christian Dior’s new look would work and would express that.
0:48:59 Could an LLM do that thing? How much variance do you need? I don’t know. It’s an interesting,
0:49:05 like, thought experiment to ask that question. There’s a completely different place to take this,
0:49:12 which is to say this is an appeal for boutiques and in-person events and the unique and the curated
0:49:16 in the individual. There’s a shop I always used to always talk about. I’m not sure if it’s actually
0:49:19 still there. There’s a shop in Tokyo that just sells one book in Ginza that just sells,
0:49:24 one book and they change what it is once a month. It may have closed 10 years ago. I’ve been talking
0:49:27 about it for 20 years. But the point is you don’t go into the shop and have to work out what I’m
0:49:32 booked by, but you have to know the shop exists. Or you can be Amazon and you’ve got 500 million
0:49:37 scoots and they know they’ve got everything. Actually, there’s some stuff they don’t have
0:49:41 because they want to be individual. They don’t have LLMH. But for the sake of argument, Amazon
0:49:45 has everything, but you can’t go to Amazon and say, what’s a good book? Or, you know, what’s a good
0:49:47 lamp? They have all the lamps. You can’t just go to it and say, what lamp should I buy?
0:49:54 All of retailing and merchandising and advertising is about where are you on that spectrum and what
0:50:00 else do you do? Do you spend the money on rent or advertising or shipping or what? And how
0:50:07 does that work? And the more there’s a sort of a polarization between, well, if I know I
0:50:12 want the thing, I can get it within 12 hours. But how do I know I want the thing? And as I
0:50:19 alluded to earlier, an LLM is one, paradoxically, an LLM could suggest you the unique individual thing.
0:50:25 Would the LLM also create the unique individual thing? That’s a high word. And a second, that’s a
0:50:31 different question. It’s a question further down the pipe. But the more that the LLM can do what
0:50:37 everybody would probably do or say what everyone would probably say, then the more you push to other
0:50:41 places. That makes a lot of sense. I mean, there’s always going to be a market for insight,
0:50:46 whether it comes from LLMs or people, you have to be providing insight.
0:50:51 You know, our world as quote unquote content creators is a very wide spectrum of people who
0:50:56 do very different stuff. And there’s, you know, there’s people doing AI slop and there’s people
0:51:01 doing, you know, the, what is it called? So, you know, the passive income thing. But there’s people
0:51:04 who do very different kinds of content coming from different places for different reasons. You know,
0:51:08 Scott Galloway does very different kind of stuff to me. Mary Meeker does very different kind of stuff
0:51:13 than me. You do very different kind of stuff to me. It’s just part of that is about who you are and
0:51:18 your story and the authenticity of it. And some of it is about no one cares who you are, but you’re
0:51:22 saying interesting stuff. And some of it’s a recommendation algorithm or something else.
0:51:27 There’s a book by Zola about the creation of department stores called Bonheur de Dame,
0:51:32 which means the happiness of women. And it’s basically about a 19th century Jeff Bezos calling
0:51:37 a department store into existence out of thin air through force of will. And like he invents
0:51:44 fixed prices so that you can have discounts and loss leaders and mail order and advertising.
0:51:49 And, you know, he puts the slow moving expensive stuff at the top of the store and he puts food
0:51:54 and makeup on the bottom of the store. There’s nothing new under the sun. And meanwhile, the shopkeepers
0:51:57 on the other side of the street are saying, like, have you seen what that maniac is doing now?
0:52:02 He’s selling hats and gloves in the same shop. He’s got no morals. He’ll be selling fish next.
0:52:05 And of course, he’s got this counter. Like there’s like the whole plot point is about
0:52:09 loss leaders. So there’s like you can again, you’re going to step back and think, well,
0:52:13 people have freaked out about industrialized mass produced product before. People are freaked
0:52:17 out about there being too much content. There’s a line that Erasmus was the last person to have
0:52:23 read every book. There’s too much AI content slop on the internet now. Like, yeah, how many books
0:52:26 do you think were being published in 1980? Do you think everyone was reading all the books
0:52:27 then?
0:52:32 Yeah, same thing. Just different scales, I guess. What advice would you give students today?
0:52:36 Well, when I was a student, we’re all supposed to be learning Japanese. I think that was just
0:52:41 the tail end of that. You know, I was sort of lucky to have a sort of very expensive and old
0:52:47 fashioned and handcrafted education that was all about learning how to learn and learning how
0:52:56 to think. I think there are skills that people used to sneer at that probably shouldn’t have
0:53:00 been sneered at and certainly shouldn’t now. I mean, I’m old enough to remember when people
0:53:05 would just sort of smugly say, well, I’m not computer literate. So that was like, it was like
0:53:10 being a car mechanic or something. I don’t know how to do that. That’s not my problem. That’s
0:53:14 somebody else’s problem. And I don’t think anyone now, partly this is because of mobile. I don’t
0:53:20 think anyone thinks like that anymore. Should you learn to code? No, I think you should find
0:53:23 out if you want to learn how to code. I think this is like saying, should you learn an instrument
0:53:29 or should you, you know, go take theatre classes? That may or may not be what you should be doing.
0:53:33 Of course, what does learn to code need in 10 years time? That’s a different question. But
0:53:37 should you, I don’t think you should presume you will or won’t be a software engineer.
0:53:42 I think you should presume that you will need to be curious and that you’ll have many careers
0:53:48 jobs and different kinds of jobs. I think you should be focusing on learning how to think.
0:53:51 But I don’t know, I think you should be presuming that everything will change.
0:53:56 Everybody says something like learning how to think. I feel like you would have a really good,
0:54:00 what does that mean? Like break that down for me, because it probably means different things to
0:54:06 people. So every now and then I’m slightly perplexed to get an email asking for career advice because I
0:54:11 think if you looked at my LinkedIn and I sort of company shut down, company shut down, company shut
0:54:17 down, like lasted a year there, that didn’t work. Coming from the UK and seeing the US system, I never
0:54:24 really liked the US idea that like, if you want a good job, you should be doing maths and business and
0:54:31 engineering. Now, that may be how people hire students here. But I never liked the idea that
0:54:36 you’re like studying philosophy or studying history or studying literature is useless because you’re
0:54:41 learning about history. That’s not what I learned. Yes, you know, I wrote lots of analogies about
0:54:46 history, none of which are actually things I studied at university. What I learned studying history at
0:54:53 Cambridge was how to ask what the next question is, how to break this apart, how to read 100 books or 50
0:54:58 books in a week and find the bits that you need, how to synthesize lots of information, how to ask,
0:55:03 well, what does that actually mean as opposed to what it looks like? It means, do you believe this?
0:55:08 Is this credible or should you just jettison that idea? How do you put this together and think about
0:55:14 how you explain something? And that’s what my friends who studied English did, or my friends who studied
0:55:19 philosophy did, or my friends who studied engineering did. That was what you were being taught how to do.
0:55:26 you weren’t being taught to be an English language, to be a historian or to, and I’d hesitate to think
0:55:33 that, you know, you can only build a company if you’ve had it or only work for Goldman’s or McKinsey or a big
0:55:40 law firm. If you had a particular kind of education, a particular kind of degree, I think you should be
0:55:48 looking for what’s going to challenge you and push you and give you the ability to learn and think in
0:55:54 different ways. But again, this is me, you know, what are the skills that you have? How does your brain
0:55:58 work? How do you think about things? And it took me, God, 20 years to work out what I was good at.
0:56:04 Although I’m not sure that you can know that as a student. So you have to try and find what you’re
0:56:10 good at, as well as, you know, learning to think. Maybe learning to think is what I do. Maybe that’s
0:56:13 not what you should be doing. You should be learning what is it that you should be learning to do? What
0:56:16 are the things that you’re good at? Try all the different things. I don’t know. It sounds like a,
0:56:21 like in a university commencement speech now. I don’t fucking know. But you don’t know what you’re
0:56:26 going to be good at. So you kind of want to try and like create options for yourself.
0:56:36 What did you learn about investing working at a 16Z? So there’s a bunch of like maxims or sayings.
0:56:40 That’s probably, I wouldn’t want to dignify them as like theses or anything else, but there’s a whole
0:56:45 bunch of maxims and sayings in venture, which, you know, we could have a podcast talking about,
0:56:49 but there are better people to give you a podcast talking about the mechanics of venture. But like
0:56:54 you have, you’re understanding what startups are and how they work and how the machine works and
0:56:58 startups are an industry and Silicon Valley is like a machine for creating startups.
0:57:03 And still too many people kind of look and say, well, that was a dumb idea. And it’s like, well,
0:57:08 that’s the wrong question. The question is, if you look at a startup and you think, could it work?
0:57:12 And if it did work, what would it be? And could those people make it work? And then you understand
0:57:16 more, you like the mechanics of, well, how does social media work? And how do people build companies? And
0:57:21 what is it like to create a startup, which is a whole other conversation. I think something else
0:57:27 that I learned was, um, calibration. This is sort of, again, another metaphor I always think of,
0:57:32 which is that if you go to like a really great art gallery, like you go to the MoMA or, you know,
0:57:39 the Met or the Louvre or something, everything there is a masterpiece. If you go to a smaller,
0:57:44 weirder art gallery, like London’s a gallery in London called the Wallace Collection, or I was in
0:57:50 Rome a couple of weeks ago and I went to one of the sort of old aristocratic palaces. And this palace
0:57:57 is like, it’s like 10 or 15 rooms of pictures. And they’ve got like a quite good Tintoretto and a
0:58:05 maybe Titian and a Raphael. You see it glowing across the room. And they’re like, oh, that’s why he’s
0:58:11 Raphael. And it’s the same when you see lots and lots of startups, like, oh, that’s why he’s
0:58:17 Oh, no, that’s why this is bollocks. Like, you get 10 minutes in and like, God, I’ve got another
0:58:21 45 minutes. I’ve got to pretend to be interested and polite. So the founder has a good experience.
0:58:31 Um, it’s seeing that contrast and texture and seeing what good looks like, seeing what worked,
0:58:36 what didn’t work, what people tend to say, how this tend to work, pattern recognition as much as
0:58:41 anything else. You also get, there’s a whole, all sorts of other kind of cultural contexts you get
0:58:45 around, you know, Silicon Valley is, can be very high school. You know, I was used to say it’s an
0:58:49 industry town. And I always used to say, it was like being a college town where there’s one subject.
0:58:54 So everybody you meet is doing the same thing. And so in some ways that’s very powerful.
0:58:58 You, you know, everybody around you is doing a, you want to do a PhD, everyone around you is doing a
0:59:01 PhD. Here’s the world expert on the subject. Of course you are. It’s like being a middle class kid.
0:59:03 Of course you’re going to university. What do you mean you’re not going to, everyone’s going to
0:59:06 university. Of course you’re going to do great work. Of course you’re going to start a company
0:59:11 and you’re surrounded by the people who’ve done it. Um, you want to get a CTO who’s done it five
0:59:16 times. You want to get the head of grace. He’s done it five times. They’re all there. The other side of
0:59:20 that is you’ll never meet anybody who isn’t working on exactly the same stuff and isn’t interested in
0:59:24 what you’re working on. So you have no external context. You have no external perspective.
0:59:28 The nearest theater is in LA than, you know, Chicago. I think like, is there a theater in LA?
0:59:31 You want to go to an art gallery, you’ve got to go to LA.
0:59:35 Who’s the best positioned right now? So from the outside looking in, you know,
0:59:42 SOC is, is going all in on AI. Elon seems to be going all in on AI. Uh, which of sort of the,
0:59:48 the leading companies do you think like, Hey, why, why are they all of a sudden shifting and, and really,
0:59:53 you know, they were dabbling in it before, but now they’re, they’re really committing, you know,
0:59:58 tens and hundreds of billions of dollars. And then who’s the best positioned in this sort of
1:00:03 space. If you had to pick one and invest your entire net worth in it, who would it be?
1:00:05 Well, that’s, that’s, that’s several different questions.
1:00:06 Let’s pick them apart.
1:00:11 Like this was a technology that had been kind of floating around before chat GPT 3.5 and
1:00:15 everyone kind of saw that it didn’t work very well. And then chat GPT was, no, actually it
1:00:20 works well enough. And then there’s this explosion of interest since then. And so I think last year,
1:00:28 the Google, Google, Microsoft, AWS, and Meta spent about $220 billion of CapEx last year.
1:00:32 And we’ll probably spend something over $300 this year. And so that’s basically more than doubled,
1:00:36 almost tripled, I think, from a couple of years ago. So there’s enormous surge in CapEx in investment
1:00:41 in this. And we’ve got these stories about like, well, Meta bought half 49% of scale.ai for $15
1:00:48 billion, apparently looked at both of the other recent OpenAI spin out, SafeRoute, Superintelligence,
1:00:52 and what’s the other one? Thinking Machines, which are both basically pre-product, pre-revenue
1:00:58 labs with somebody from OpenAI at multiple tens of billions valuations. And he apparently,
1:01:03 like Sam Altman, complained that Mark is offering people $100 million to join. So Mark’s in beast
1:01:08 mode. Microsoft has this kind of weird situation in that his own models aren’t actually very
1:01:13 good, but it’s got this very kind of weird relationship with OpenAI. OpenAI, Sam Altman
1:01:17 is, I was going to say polarizing figure, but actually opinions about him tend to be fairly
1:01:22 unanimous and tend to be fairly negative. Like everybody who’s ever worked with him quit.
1:01:29 OpenAI itself still kind of sets the agenda, but much less so than it was two years ago.
1:01:34 And I wouldn’t want to do like a detailed, like calling the scores on whose models are good and
1:01:38 whose labs are good. But, you know, objectively, Google is clearly like firing on all cylinders
1:01:44 now and is making great models. Lama seems to have been an embarrassment.
1:01:49 And so Metro is kind of scrambling to catch up. Apple is a slightly different position in that
1:01:54 they are always sort of taking the position that they don’t want to be first. They want to do it
1:01:59 right. And that they don’t need to be doing whatever the latest consumer internet thing is like. They
1:02:04 don’t have a YouTube. I think Craig Federighi said in an interview after WWDC, like, we don’t have a
1:02:07 YouTube. He didn’t call it YouTube. We don’t have a YouTube. We don’t have a car showing service.
1:02:12 We don’t do grocery delivery. We also don’t have a chatbot. Okay. That wasn’t quite the question.
1:02:18 The question for Apple is how much does integrating an LLM into the operating system change the
1:02:24 experience of what it is, potentially shifting the competitive balance with Pixel, which at the
1:02:27 moment has basically only gets bought by people who work for Google and people who work for the
1:02:32 tech press and like literally no one else buys Pixels. Well, because Google doesn’t want to compete with
1:02:36 Samsung. We could go down a whole smartphone industry rabbit hole. There’s this sort of
1:02:41 question for Apple around, does this actually change the experience of what a smartphone is,
1:02:48 what the ecosystem is? Does it end up kind of getting Microsofted in the sense that you’re going
1:02:53 to still, for the time being, you’re still going to buy a smartphone. It’s not at all apparent
1:02:56 there’s going to be another device. And if there is, it’s a long way away and it might be an Apple
1:03:00 device as well. But you’re still going to buy a smartphone. You’re still going to buy the nice one
1:03:05 with a good battery and the fast chip to run the models and the good screen and the best camera,
1:03:08 which will still be an iPhone because Apple still has the best chip team and a whole bunch of other
1:03:13 hardware advantages. But everything you do on it will be from someone else. And it won’t be someone else in
1:03:17 the sense that it’s an app from the app store. It will be someone else in the sense that it’s a model
1:03:22 running in the cloud, which is what happened to Microsoft in the 2000s, which was everyone had to
1:03:26 get on the internet. To get on the internet, you needed a computer. You weren’t going to buy a Linux
1:03:30 computer. You probably wouldn’t buy a Mac either. So everyone bought a Windows PC, but they were using
1:03:38 it to do web stuff, not Microsoft stuff. So Microsoft kind of lost that. And so that would
1:03:42 be the concern for Apple would be you’ll still buy your new iPhone and you’ll buy the iPhone air this
1:03:47 autumn because it will be center and lighter and it will be a lovely phone and you’ll use it to do
1:03:56 chat GPT. But the counter argument would be to say, yeah, you’ll use it to do chat GPT and DoorDash and
1:04:03 Uber and Instagram and that cool new game and the other new cool game that’s enabled by LLMs and to do
1:04:08 TikTok and to do and, and, and, and, and, and it will be kind of the same except they’ll…
1:04:12 So there’s a sort of, you see what I mean, there’s this sort of slight fuzziness around what the bear
1:04:19 case for Apple actually is. Does it sort of end up like Microsoft did? How bad is that exactly? What
1:04:20 does that mean? Yeah.
1:04:27 If we all go to wearing something like this, that’s lit up by AI, then that’s a bigger shift, but it’s
1:04:34 very unclear how close that really is. Um, just for the optics. There’s another axis here, which is
1:04:39 what happens to Google search? Where does that money go? How do you map the search activity that
1:04:46 goes to an LLM that, and how do you map that against where the revenue is? And also from the other side,
1:04:52 how do you map that against where the publishers are? How do you think about whether you just shift your
1:04:57 habit and you’re actually using chat GPT as Google? It’s basically doing what Google does, but you’ve
1:05:02 shifted the brand and you’re going to that search box instead of the other search box. I don’t think you
1:05:09 can count them out in absorbing that in. So who would you, out of the public companies,
1:05:14 you had to put your money in one of the top like seven? Well, then there’s a valuation question. And
1:05:19 I, well, a long time ago, I was a public markets analyst and I was bad at being a public markets
1:05:22 analyst, equities analyst for a bunch of reasons. One of which was I was never interested in
1:05:26 share prices. Well, we’ve given the standard disclaimer. You’re forced to. What would you?
1:05:36 It’s hard to see iPhone sales slipping from what we see now. Even half the cool, sexy Google stuff is in the
1:05:43 Google app on the iPhone. I think meta and Google, there is this sort of big question around where the ad
1:05:50 revenue goes and how much the ad revenue gets pulled away to different places. I think Instagram is probably
1:05:56 in a very good place in terms of changing what advertising looks like and how that works. I mean,
1:06:03 I had a slide in my presentation, which was like, um, what meta and Amazon want to do is to make LLM’s
1:06:08 commodity and for the sold at cost. Yeah. Now this is why meta made it open source because they want to
1:06:11 make it commodity and for the sold at cost. And they differentiate on top with meta stuff with
1:06:15 Facebook, social, Instagram stuff. And they want the model itself to be just infrastructure. Amazon would
1:06:19 also like it to be commodity infrastructure that sold its cost because that’s what anybody wants is. They sell
1:06:23 commodity infrastructure costs and they do it better than anybody else. And they make a lot of money from
1:06:27 doing that. Go to Amazon’s financials and basically all the money comes from AWS and the ads. People
1:06:32 who complain about AWS don’t come and realize the ads make it. Amazon did 50 billion, 60 billion dollars
1:06:39 of ad revenue last year. So Amazon seems to be fine, but there’s a bunch of stuff to navigate around
1:06:44 how does this change how people buy stuff on Amazon. Who does that leave? Microsoft, it was this line
1:06:49 from Bismarck that the great man is somebody who hears God’s footsteps through history and grabs onto his
1:06:54 code as he walks past. And like Satya is like tried to, first of all, he tried to grab onto VR and AR
1:07:01 with HoloLens saying that we don’t talk about that anymore. Now it’s AI, their own models are not
1:07:06 really ranking. I mean, they hired Mustafa, but like they’re still struggling. They’ve got this weird
1:07:12 contentious relationship with Sam Altman and OpenAI and it’s basically not their models. On the other hand,
1:07:18 like they’re going to sell an awful lot of Azure to run all this stuff, which again is this tension.
1:07:24 Is it, is it that everybody’s used, just uses ChatGP2 to do the thing? Or is it that someone is going to
1:07:31 come to you with a great accounting product to run Farnham Street and it runs on Azure and it uses
1:07:37 some LLM? Who cares which one it is? It’s just, it’s just better. You know, you can edge to your bank and it
1:07:42 does the cool stuff and you can edge to that and it does the cool stuff. You know, my use case for an LLM is do my
1:07:49 fucking invoicing for me. It’s not even that, it’s work out why exactly it is that that client’s ERP
1:07:55 doesn’t like my bank account and not have me spend the next three months bouncing emails back and forth
1:08:02 with somebody in India about one second on getting this done. That would be a great use case that LLMs
1:08:07 can’t do that yet. If they could, that would be great, but we’re not there yet. So Microsoft and Google
1:08:12 are in this sort of position of being the incumbent. Both, you know, how can I put this? I’ll give you
1:08:15 a more systematic, again, I’m sort of thinking my way through to the answer to your question.
1:08:21 For Google and Microsoft, they have an incumbent business that is potentially disrupted pretty
1:08:25 profoundly by this, but they also have a cloud business that sells all the new stuff for this.
1:08:30 Amazon has an incumbent business that doesn’t get disrupted by this, at least much less,
1:08:33 obviously, and a cloud business that will be very happy selling all of this stuff.
1:08:39 Meta doesn’t have a cloud business selling this stuff and has a bunch of new ways to make money
1:08:42 from all of this new stuff, except they’ve got to have some better models.
1:08:50 Apple, is this a competitive threat to the iOS ecosystem? A lot of stuff would have to happen
1:08:56 first. And they’d have to drop a lot more balls before that was to happen. And meanwhile,
1:09:00 they’re still going to sell you the nicest glowing rectangle to do all of this stuff.
1:09:05 And then we’re talking about glasses and VR, which is a whole other two-hour conversation about
1:09:10 when does that happen. There’s still people poking around in crypto, like Web3, Web3, remember? Maybe
1:09:14 that. Someone said that people still working on crypto are like those Japanese soldiers on islands
1:09:18 in the Pacific. You don’t know the war’s over. But there’s still people working on crypto. So
1:09:22 that’s like another disruptive thing coming down the pipe. Who are the other incumbents?
1:09:32 Netflix is a TV company. It was a car company. Elon Musk. There’s a whole Tesla conversation
1:09:37 fascinates me because Tesla bears balls think it’s a software company and Tesla bears think it’s a car
1:09:42 company. And at the moment, it’s a car company. I mean, yes, they launched autonomous driving.
1:09:48 But what did they launch? They launched half a dozen existing model cars with test drivers.
1:09:54 Tesla’s doing a geofence drive that everyone else was doing 10 years ago. Is that going to scale?
1:09:59 Is they finally going to get the flywheel of having all the camera data, meaning they will work with just
1:10:05 cameras? We’ve been… I don’t know. There’s a conversation you could have asked 10 years ago.
1:10:09 In fact, I wrote stuff 10 years ago. Are there winner takes all effect in autonomous cars?
1:10:13 Will Tesla get it working with cameras before everybody else gets it working with LiDAR? Well,
1:10:18 Waymo’s got it working with 50 grand of LiDAR or whatever that stack costs. It’s tens of thousands
1:10:22 of dollars of extra stuff on the car. So they’ve got it working with all of that stuff. Tesla does
1:10:29 not have it working with cameras. Will Tesla get it working with cameras before Waymo can get rid of
1:10:37 the LiDAR? We could have had that conversation. I literally was on podcasts seven, eight years ago
1:10:43 having those conversations. We don’t know the answer. Maybe. We don’t know. The interesting Tesla
1:10:47 point is people always looked at it and said, it’s the iPhone of cars. No, it’s not. What’s happening
1:10:53 is that cars are becoming Android with no iPhone. And Tesla is just selling and is in that metaphor,
1:10:58 Tesla is just another Android phone maker. And they’re competing with the whole Chinese industrial
1:11:02 policy to make more. And there’s just going to be a flood outside the US. They’re protected by tariffs.
1:11:07 Everywhere else, it’s very clear what’s happening. It’s just a flood of EVs that are just as good as
1:11:13 Tesla’s. We always end on the same question, which is what is success for you? We live in the luckiest
1:11:18 time. You know, we are not worried about rockets landing on our heads. We’re not worried about our
1:11:22 children as dying from diseases. We’re not worried that the bank might be closed tomorrow and all of
1:11:28 your money’s gone. We’re doing something interesting that we enjoy and that pays the rent that we want to
1:11:34 be able to pay. I get paid to fly around the world and give slides for money. So I think I’m doing okay. I could
1:11:44 always be doing more. But I’m always looking for the next question. Like, I’m always trying to be curious.
1:11:47 This was a great conversation. Thanks for taking the time today.
1:11:48 Great. Thank you.
1:11:53 Thanks for listening and learning with us. Be sure to sign up for my free weekly newsletter at
1:11:59 fs.blog slash newsletter. The Farnham Street website is also where you can get more info on our membership
1:12:06 program, which includes access to episode transcripts, my repository, ad-free episodes, and more. Follow
1:12:12 myself and Farnham Street on X, Instagram, and LinkedIn to stay in the loop. If you like what we’re doing
1:12:16 here, leaving a rating and review would mean the world. And if you really like us, sharing with a friend
1:12:19 is the best way to grow this community. Until next time.