The Origins Podcast with Lawrence Krauss - The Like Button, and the Strange Power of Tiny Ideas | Martin Reeves

Episode Date: December 22, 2025

On this week’s episode of The Origins Podcast, I ended up in a place I genuinely never expected to go: the humble “like” button. When the idea first landed in my inbox, my reaction was basically, why on Earth would anyone write a whole book about that? Then I spoke with Martin Reeves, and I discovered that the history of this tiny icon is a surprisingly rich window into innovation, entrepreneurship, human psychology, and the modern attention economy.Martin is a senior figure at BCG’s Henderson Institute, but what made the conversation especially fun for me is that he is not a consultant who wandered into science. He has a background in science, and then wandered into the world of strategy, technology, and ideas, and he approaches the “like” button the way I wish more people approached our digital world: with curiosity, skepticism, and a willingness to follow evidence across disciplines.The central irony, of course, is that the “like” button began as an almost laughably small, practical solution. In the story Martin and his coauthor reconstructs, it is often less about a single inventor than about a messy ecosystem of micro innovations, technical constraints, and cultural accidents. Yet those small choices compound. The result is that something as simple as a handful of code became a universal signal that helped shape social media, transformed advertising, and created feedback loops that are now baked into the infrastructure of daily life.We also dig into why it works so well on us. The mechanisms are not mysterious in the abstract, they are biological and social, but the scale is unprecedented. Approval and recognition are ancient. Industrialized approval is new. And once you start thinking that way, you notice how these same feedback dynamics are spreading into new domains, including the tools we now use to interact with AI.This conversation surprised me, and I suspect it will surprise you too. Indeed, if you are like me, and wondered why the like-button is worth discussing, you will be surprised to learn how much of the modern world is quietly organized around it. You can listen on any podcast platform, watch on YouTube, or view ad free on Substack. And if you are tempted at the end, well, you may even find yourself clicking the very thing we spend the episode dissecting.You can listen on any podcast platform, watch on YouTube, or view ad free on Substack. And if you are tempted at the end, well, you may even find yourself clicking the very thing we spend the episode dissecting.As always, an ad-free video version of this podcast is also available to paid Critical Mass subscribers. Your subscriptions support the non-profit Origins Project Foundation, which produces the podcast. The audio version is available free on the Critical Mass site and on all podcast sites, and the video version will also be available on the Origins Project YouTube. Get full access to Critical Mass at lawrencekrauss.substack.com/subscribe

Hi, and welcome to the Origins Podcast. I'm your host, Lawrence Krauss. In this podcast, we go to a place I never imagined I would go to. The like button on the internet. I spent time talking to Martin Reeves, who's one of the authors of a new book about the history implications of the like button. And when I first got pitched it, I thought, why would you ever want to write a book about

the like button? But in fact, understanding its history can teach us a great deal about entrepreneurship, innovation, science, neurology, society, and the future. And I had a wonderful time talking to Martin Reeves about all of those subjects in this podcast. It was a revelation for me that I found the subject even interesting. and I hope it will be a revelation for you too. You can listen to the podcast on any podcast listing site, or you can watch it on our YouTube channel,

and I hope you'll subscribe to there, or you can watch it add free on our Substack site, Critical Mass, Substack site. And I hope you'll consider doing that because the proceeds from that site go to supporting the nonprofit foundation, the Origins Project Foundation, that supports this podcast and makes it. it possible. No matter how you watch it or listen to it, I hope you will be surprised and fascinated,

and if you even had a chance, you might click like. With no further ado, Martin Reeves. Welcome to the Origins Podcast. I'm your host Lawrence Krause, and I'm very happy to have with me here Martin Reeves, who has co-written a book that captured my interest called Like. Now, I wish I had the book here. I don't know if Martin has a copy. I have just moved, which is why you're seeing a different background than before. But there it is. There is the book.

And you notice it has a thumbs up. It's about the history and associated efforts associated with the like button online. And you might say, why would you write a whole book about the like button. That's the first thing I said when I first heard about it. And I learned later that it's not only fascinating, but it describes a microcosm, a lot of issues that are relevant for our modern world of social networking and technology. And it's fun and interesting. But before we get there, Martin, you may not have seen this podcast before, but it's the origins podcast, and it's called that for a reason. I like to find out how people got to

where they are. Now, you are now involved with an organization called BCG, right? Which stands for the Boston Consulting Group. Specifically, the Henderson Institute, the Research Institute of the Boston Consulting Group. Which I learn explores and develops new insights from business and science, but look at the technology of ideas and harnessing imagination for corporate resources, et cetera, Really neat, I suppose, but what pleased me more was I discovered your background was actually in science and not consulting. I'm always suspicious, I should say about consultants, but I like people who have a real background. And I want to learn more about that.

You got an MA in natural sciences, and what was your undergraduate degree in science as well? Yeah, I studied natural sciences, which is a convenient bracket in Cambridge for studying all sorts of. things. So the thing you should know about me is I'm an unashamed generalist. I sometimes thought I was the last generalist walking the earth. And this permitted me to study, you know, physics and chemistry and mathematics and also Persian and Japanese studies at the same time. So I bring that spirit of generalism to anything I do. Now, so, and you can, so your BA and your MA is a natural sciences from Cambridge. By the way, you know, I'm sympathetic to when I first started, you know, I started at a specific university because they had what's called a general science program,

which allowed me to do history and science at the same time. And I like that. But eventually, of course, one tends to specialize. I was intrigued that you actually have a master's degree in biophysics, but not from Cambridge, but from a Japanese university. So you must have, your Japanese studies must have taken you to Japan. Did you work there? Indeed.

So I got a Mambusho scholarship to go study Eleanor. electrophysiology in Tokyo, and honestly, the electrophysiology probably had an equal pull to the exposure to Japanese culture, and that was yet another arrow in my quiver of generalism, essentially. Well, okay, but then now, you have an MBA. I'm going to assume you got the MBA after all this. Is that right? Yes.

So general science is interesting. Japanese is interesting, but one has to make a living? Is that maybe, why did you get an MBA? Or was it because you were, or because you were specifically interested in basically what we do now, looking out of technology? Well, it's a little personal, but my roommate at university told me that, you know, he admired my brain, but I must know that I would be very unsuccessful in life. And I said, why is that Steve?

Steve is now a professor of geology. And he said, because you're a butterfly, and nobody employs butterflies. And actually, he nailed it. I'm a curious generalist. And so my life journey was basically trying to figure out how you could be paid to be a generalist. And I found that the high variety of problem solving that one sees in consulting might be a solution to that.

And then I migrated towards the research end of that. But essentially, I got an MBA in order to get into consulting where the attraction was generalized problem solving. Yeah, absolutely. I mean, it gave you a credential to do that. We once created a master's degree when I was chairman of physics department in physics entrepreneurship,

you'll be pleased to know as an alternative to an MBA. The business school dean called it an oxymoron. But he didn't realize that basically scientists are entrepreneurs. And that'll come into this. Later on, when we talk about innovation and how it affects things,

What goes on in science is very similar, and I'll probably return to that several times. But it gave you the chance to use your generalist background and apply to this. It sounds like you found the perfect position for your interests. And then, but now I'm interested in how the book, you talk about it a little bit, how the book began. You began to have a conversation. Your co-author is Bob Goodson, right? Yes. And he was actually involved in the like button.

So how much? He was. So basically, Bob is a classical literature scholar. So he spent most of his career pouring over vellum manuscripts in the Oxford Library, looking at the marginalia, the notes that the scribe monks made in the margin of manuscripts that took hours and hours to copy. And he ended up in Silicon Valley. How did he end up in Silicon Valley? he was recruited there when nobody else wanted to work there. So there was this period around 2000 when all of the dot-coms had gone bust, and everybody went back to their day job. So that was the time when a classicist could actually be hired by an entrepreneur to run a company. And I was just getting to know, Bob. He struck me as like a really interesting combination. I'm often guided by curiosity, and I just thought,

what's a classical scholar doing, running a Silicon Valley startup? And so in the course of getting to know Bob over a planned 30 minute coffee, I learned that Bob was a sort of compulsive collector. So he's collected every train ticket and every, you know, makes copious notebooks and dates everything since he was a child. And so to make conversation since he was moving, I said, Bob, you must be discovering interesting things in your many, many boxes of stuff as you move. And he said, indeed, and he pulls out a notebook. and it falls open to a sketch that he'd made, a dated sketch of the like button. And I thought I knew that Facebook had invented the like button.

And so I said to Bob, yeah, that date, you know, are you telling me you invented the like button? And he gave me a very curious response. He said, no, of course not. Well, maybe I'm not sure. And like the idea that you could be unsure whether you've invented something, I thought was quite interesting. And so the 30-minute coffee conversation became a,

12-hour conversation until we were kicked out of the last open restaurant in Mill Valley. And by the end of it, we had the outline for a book. And the thing that drew me to the story of the like button was its sort of Byzantine complexity, just the sheer messiness of the story. In fact, after three years of research, we never really did get to the bottom who invented the like button. It's a very, very complicated story. And the second thing that attracted me was how something so small that, indeed, you know,

no rational person would consider writing a whole book on the like button. How something so small, 12 lines of JavaScript code could create such enormous change, could create an industry, more or less destroy an industry, the advertising industry, create enormous social problems, you know, waves of legislation. And the story continues, of course, because Chad GPT has a like button. And we may ask, you know, what is the like button doing on our LLM? So that was the origin story of the book, if you like. Yeah, no, there, that was a great introduction.

Yeah, why a book by the word, but because it has an impact. And because the story is interesting. As I say, I first thought, who the heck cares about like button? And then reading it, the story was interesting and involved people that I may have known from Silicon Valley or certainly friends of mine know very well. and they're interesting personalities. So the book sort of about its origins, the role of played,

and in many ways, a significant role that I never would have thought of, and has changed a lot about the way the world operates, and has also, as in much of technology, had impacts that were totally unexpected. And I think that's what captured my own interest. Yeah, I mean, so just to comment,

comment on the impact and the unexpectedness. So this little symbol, this little 12 lines of JavaScript code, that essentially registers instant recognition of a piece of user submitted content. At a time, by the way, when the idea of Web 2.0 existed, but nobody was submitting content. They were desperately trying to encourage content submission. With a symbol, with an ancient hand gestural symbol attached, the thumbs up.

Such a tiny thing, but essentially it created social media. Essentially, it blew up the advertising industry, because this is old joke in advertising that 50% of your money is wasted, you'll just never know which 50%. Well, the like that was the first instantaneous

granular, low-cost feedback loop. So now we do know which 50% of advertising is not working. It created enormous social problems and a wave of regular latry innovation and this whole thing is still continuing to continue with with with LLM so so so that's for the like the tiny and the big impact part of it and in terms of

unexpectedness I guess one thing was very striking when we tried to talk to everybody that had a role in the origination of the like button was that absolutely nobody had any idea about the eventual impact so what they were trying to solve at the time was, well, very low-level tactical problems. For example, how can you register that a review has been submitted to a site like Yelp without triggering a page refresh in the era of Dial-Up Internet? Because if you trigger a page refresh, the conversation would be ended for 20 seconds. You'd probably lose both sides of the conversation.

Nobody was thinking, hey, this could be the route to monetize and create a new industry around social media. In fact, we look very hard, and we only found one person that had the vision of what social media would eventually become, and it was actually not a technologist. It was actually an author, Gary Steingart. So Gary Steingart wrote a novel called True Sad Love Story at about the time that all this was happening. And he fairly accurately, in a fairly dystopian sort of way, anticipated everything that was going to happen in social media. And so one of the most fascinating interviews was I asked him, how come the technologists could have to figure out what would happen to their own technology? but you could. And his answer was twofold. He said, well, number one, you don't have to break the future. You just have to find a pocket of the present where the future is already present. So he immersed

himself in these embryonic social media companies that no longer exist. And then the second thing, he said, I don't, I didn't ask the same question. The technologists were asking the question, what can the technology do? He said, I, as a humanist, ask the question, what would the humans do with the technology? And of course, the answer is they'll do everything with the technology. They'll lie, beg, cheat, borrow, fall in love, dissembled. They'll do everything with the like button. And that's the clue to sort of seeing where this thing would go as opposed to the intended function of the like button.

Okay, well, this is a great sort of 40,000 foot overview of this. And I want to now take a deep dive into the actual details. And as you say, Bob was a medievalist who ended up at a startup and was the number one employee at Yelp. Yes. And we'll get to what, I mean, and what Yelp was trying to do was get reviews. And to get reviews, you've got to do, get people to do reviews,

and you've got to encourage people to do reviews. So that was, you know, that was sort of the motivation for getting there. But before we get there, there's a lot of interesting characters. First of all, I think you've already said it. But the most important thing is, what is the like button? And you sort of define it on, on, on, um, on, um, on page eight of, or it's a simple one-click mechanism that allows a user of a site or platform to register an emotional reaction to a piece of content created by another user. And the other

thing is that it's one-click and keeps it on the same page. That's the important thing. But so now, so having said that, and we'll get to why that was interesting, and you've sort of alluded to it in the beginning, I want to, I want to look at personalities because Bob would, never ended up in Silicon Valley if it weren't for someone named Max Lechon. So you want to talk about that story a little bit because I find it fascinating. Yeah, so Max Lechon was a successful venture capitalist, part of the PayPal Mafia. And so he stayed in Silicon Valley when the nuclear winter hit and was trying to back a bunch of ideas that would result in the next way. the next resurgence of internet-based business,

what we now call Web 2.0. You know, he was trying to hire people and nobody wanted to work in Silicon Valley, have ready to gone back to their day jobs. And so he actually made a trip to Oxford, where he met Bob. I think Bob asked him some awkward questions in a presentation. And Bob was interested in,

had a part-time interest in knowledge nets. So Bob was interested in sitting down with chemists and physicists and saying, tell me about all of the knowledge objects in your world and how they relate to each other and where they come from. We'd create vast knowledge maps of different subject areas. And this was sort of, I guess, alluringly close for Max Leveschin to the dream of Web 2.0, the social interchange of content. In fact, you could define the like button as the atom

of content. The simplest piece of content that user can submit is a reaction to somebody else's content. And so Max lured Bob to Silicon Valley, lured him to an incubator where Yelp and a bunch of other companies were being incubated. Bob learned coding became the web designer employee number one of Yelp. And I guess it was sort of Bob's humanities background that led him to the thought that, well, we can't pay people to write reviews. We must actually, you know, use human behavior and motivation and brain science to encourage reviews. And you had this sort of intuition about reciprocity and recognition and as being the social currency that might encourage people to submit reviews. And, you know, why did he reach for the thumbs up symbol? Well, this was

already well-established in American culture. And also, there was a very influential book by a guy called Steve Kroog that most normal human beings would not have heard of, but he's essentially the guru of user interface design. And Steve Krupp published a book called Don't Make Me Think, which essentially was a very revolutionary idea. It was the idea that if you have something innovative and you wanted it to spread, make it look ordinary to suppress the novelty. you know, wrap it in something familiar. Don't cause me to think. And so, hence, Bob reaching out for this ancient gestural symbol to say, hey, nice review.

Well, yeah, and we'll get to the cultural aspects of that first. I don't want to leave Max too early. Bob heard a lecture and he wrote down notes. And I think what's interesting about Levchen from what I can tell, and then I met people know them as well. is his prescience. The fact that people, I mean, all of this, it's easy now to see, and so many young people,

to see a business model, you know, when you go to the web. But early on, it wasn't clear what it was going to be good for or how you could make money off it. And he wrote down, I really liked the fact that he wrote down, this was not long after 2000, I guess, when Max went to England to talk and he was and he was talking and Max was talking about the dawning consumer internet era because he'd done PayPal and and realized he said and he said there's going to be a resurgence of the web and this time around we can understand the business models

broadband will change things fundamentally and we're going to execute properly so he realized the potential of the fact that that you know when you don't have just a dial up as things go fast, it's going to change everything. But he didn't exactly know how, and he did a great idea. He said, let's get some good people and just have them start working and maybe, and we'll try and figure out the business model. So that is the mind of the entrepreneur, because what Max did, and you described it very accurately, was, well, a pragmatist would say it's very implausible, right? Because at the very time when the internet businesses had melted down and nobody wanted to work there and before the dawning of broadband you know we're still in the era of dial-up and before there

was any precedent at all for the idea of a two-way internet because the internet of course used to be a bulletin board it used to be one way used to be sort of advertising information not not sort of exchanging user content and views and so on so at that time you know for max to say I'm going to put my money and my time and my effort into creating something that doesn't exist. I mean, that is an act of imagination. So what caused him to regard that as a reasonable thing to do? I guess it's the power of an entrepreneurial vision. And maybe the ability to see patterns that normal mortals can't see,

like the inevitability of richer exchange of content once we had broadband, like the inevitability of progress towards broadband, like the really missing piece of the internet being the sort of the feedback from the consumer, the sort of the two way communication. So remarkably prescient. Yeah, absolutely. And I think you said that he's a typical entrepreneur, and that, you know, that does bring me back.

and I often thought when reading this book of something we used to tell our master students in physics entrepreneurship and why we thought physics was such a good training for entrepreneurship, unlike, say, engineering, because we said physicists are trained to solve problems when they don't know what the problem is. And of course, that's what entrepreneurs do. They don't often know what the problem they're solving is, and the good entrepreneurs realize they've got something and there's actually a problem that can be solved that they didn't know existed and they go with it and I think that's the mentality. You know, we considered dedicating the book to our families and our loved ones, but it seemed

to us that after we'd written about all of this, we actually wanted to do something different. So we actually dedicated the book to the tinkras and makers who history will never record because, you know, essentially what our sort of investigations uncovered was, was a swath of uncelebrated and unknown heroes that all contributed in different ways towards this process that didn't have a grand plan. You know, we often think about the scientist or the entrepreneur of somebody that has a grand plan, but none of the people we spoke to had a grand plan. They had intuitions, they had passions, but often the eventual solution emerged from the

efforts of the followers, followers, follower of what they did. It wasn't part of any of any grand plan. You know, and again, that's what what I think interests me because people think science is done with the grand plan, but it's, you know, pure science, not even applied science, but it's so similar. Often you have techniques and you, and are you trying to solve a technical issue. And it's only later on you find that solution of that technical issue. It has an application you never imagined. Or you have a mathematical technique and you find out that the universe can use it. And so it's very rare that people sort of have this vision, oh, I'm going to solve all the mysteries of the universe. It's often, it's done just the same way. And so much is

similar. It's a social process. There are lots of people tinkering with techniques. And what's great is because they communicate, you can learn about how other people use that technique and say, oh, I didn't know that the technique was there, but I can use it here. So the history of science and the history of innovation in entrepreneurship are not different at all. And of course, that's something I believe in strongly anyway. There's a big lie here that is, I think, central to the book too, which is, so I firmly believe that that's the case, having looked at the like button and also a bunch of other inventions. And, but that's not the way that the story is often told after the event. I think Howard Varma said, you know, scientists don't always tell

it the way it was. They have some celebrated, you know, post hoc rationalization of what happened. And also, it's not the story that we prefer to hear. If you said to me, you know, who invented the, who discovered the inverse square law of gravity, it's very convenient if we'd say Newton, but, you know, Hulk would have seen it at his idea there in vicious dispute about who came up with the idea first. But that's just too much information for us. Now, the... Yeah, let me interrupt for a second. As someone who's taught physics for a long time, I've always thought that, I mean, we don't have enough time to teach it as it was really learned.

Yes. We understand the logical coherence of an idea after it's done. And I have to say, even in my own papers, there are some papers that have been influential that I really didn't understand until a long time afterwards. But it's easier if you're teaching it to put it into a logical step. But that's not the way the history of science action works. And that's the damage, right? Because so I'm fascinated by the, you know, Ibson, the playwright had this idea of the

life lie, which is the thing that is convenient and valid. to believe that it isn't actually true. And so lies can be very useful. And so the simplification of who invented that, who thought of that, how did they think about that? You know, in a sense, it's a useful lie. It stops us from being good well with the information. It makes things more understandable.

The danger, though, is that if a would-be entrepreneur or scientist thinks that that's the way that you discover these things, then they're going to be probably sorely disappointed. If they're waiting for the grand plan and the complete explanation, you know, in sequential psychological fashion, that it's not going to happen that way. And also, you just hit another point. Who invented this? Often it's not a who. Ideas are circulating in, I would say, the ether,

but a lot of people, a lot of creative people are thinking about similar things, and it's not surprising, and it happens all the time in science, that independently, two different or ten different or whatever, people are coming on a similar solution because it's in there. It's in the Gestalt. It's in the environment at the time. And so it's often, as you point out, the like button, difficult to pinpoint. It's not as if one person said, hey, world, it's the like button.

A lot of people were trying to solve a similar problem. And it turned out the problem changed over time because of what happened with that technology. In the case of Yelp and Bob, it was to get people to write reviews. How do you do it? And you talk about, I guess, or Bob probably talked about this early stage of Yelp when they're trying to get reviews. How do you get people to submit content for free if you don't pay them? And there was a guy, Kevin. I was abused at Kevin S.

Right. Yeah, so I was telling you about Kevin. So in a sense, I guess we asked a very childlike question. You know, children often ask who invented that and when. did they invented? And it seemed quite reasonable that we give a clear answer to that question because somebody must have done something for the first time and they must have done it at a particular time. Well, that's, of course, not what we discovered. We discovered that this process had four characteristics that were almost the opposite of that. It was it was serendipitous. In other words,

what the people were trying to do was not what eventually happened. It was social. It wasn't just one person. It was a collection of people. It was sequential. It was. an evolution of pre-existing things over time, so much so that the boundaries between invention and discovery is sort of blur, because in a sense, the invention is often the discovery of some existing form of something that can be repurposed or combined or improved with something else.

And also situated. So by situated, what I mean is that it seems perfectly logical to think that the conception of something comes before the prototype, and then the prototype gives way to the invention, and then the invention gives way to the, you know, the dumb salesman selling the, you know, selling the invention and the rollout and the exploitation.

Well, actually, there is a strand of truth in the idea that somebody called Kevin S had played a major role in the invention of the like button, a user. So Kevin S was, so picture this time at Yelp where no good. was submitting content. There were no restaurant reviews. In fact, they rigged a bell up to the server to ring on the rare occasions when somebody submitted a restaurant review, while they examined their souls on, you know, how can we get restaurant reviews submitted

before we run out of money. And one day, the bell rang, and then it rang again, and then it rang again. It turned out to be the same guy submitting restaurant reviews, when nobody else was submitting reference reviews. And this was really rather mysterious because his reviews, his reviews, were going into the ether, nobody was seeing them, let alone recognizing them. And also, he discovered like a secret pathway in the software, like a fossil pathway in the software. It shouldn't have been possible to submit one review after another. That was not what the latest design of the software permitted. So they figured out this guy was Kevin S, who was a Bonne River foodie living in San Francisco. So they went to see Kevin and tried to figure out,

like why are you submitting reviews when nobody else is and how can we get more Kevin's and it turned out that he was having fun it was like intrinsically motivating for him to submit reviews and and from this conversation with Kevin they they realized that they needed to shift their attention from the from the business model of framing this as a service to restaurants and instead make it a a service to reviewers and also they needed to make it easy for people that weren't Kevin to find this a rewarding activity and the obvious currency was the currency of recognition. And the obvious problem with that was the page refresh, because any signal you sent back to the server would interrupt the conversation. And so a couple of

people, not just Jeremy Sothelman, the CEO and the founder of Yelp, but also Max Hong at the site hot or not, sort of more or less simultaneously figured out that you could use JavaScript, the language of websites for a purpose that it was never intended, which is to do local computations, like the computation of this sort of showing recognition for a review having been submitted. You know, hence the like button came about, essentially building on a dream, you know, Max Levski's dream, a problem, the problem of the dial-up internet, a need for ingenuity because they couldn't pay for restaurant reviews, an enterprising user, a piece of philosophy that said

this ain't going to spread unless it looks ordinary unless it looks familiar like a like button and a collection of people that were all sharing everything in Silicon Valley at the time at things like the Squid Labs meetups and Tim O'Reilly's sort of unconference so it's making it even harder to figure out who did what so all of that comes together

you get a like button and it eventually ends up doing something that none of them is totally different. I was intrigued that after Kevin, they actually had a launch party. Yes. The question of how do you motivate people? It wasn't even clear, as far as I can tell, it wasn't even clear to people that, you know,

Kevin was internally motivated. He didn't need to be complimented. But in this launch party where people could talk to each other, they complimented each other on reviews. And that was like the aha moment. And, hey, people are getting a lot of feedback. Other people like my review, maybe we should do something. And I think there was a, Amazon had a complement feature, right?

And so did TiVo, is that right? Yeah, so, I mean, so, you know, we can tell different types of convenient lies in terms of who invented it. We can say Yelp invented it. We can say Facebook invented it. In fact, both of those companies didn't – both of them really didn't invent it in the sense that Zuckerberg opposed the introduction of the liking. Yeah, we'll get to that. Dr. Montgomery didn't like it at all. Also, Yelp.

I mean, Bob came up with the first sketch of the thumbs up and the thumbs down button with a counter, clearly a like button. But actually, that was not what Yelp did. They ended up going with useful, funny, cool, the multifunctional. You could choose one of useful, funny, or cool, yeah. Right. And that actually did not become the standard, but later became a standard in more recent years, the multi-functional like button. But if you look at everything that's going on at that time, it's really quite messy because you had, you know, who invented it, it depends upon your, you know, exact precise definition of what it is. So, for example, if you say a digital voting function, where that's probably the digital bulletin boards like Telegard from the early.

1980s. You remember, you could get some green text displays on your TV screen and look at the weather on your TV screen. And digital voting was one of the functions that they had. I think, yeah, the TiVo, if we all remember the nightmare of programming our VCR, so TiVo in the late 90s had solved that problem by having two buttons, a red and a green button with a thumbs up and a thumbs down symbol that you pressed, and the software learned what you liked and fed you more programs like the ones you'd like. So that's the semi-physical, semi-digital like button. A huge breakthrough was hot or not. There was this site that, where you rank the appearance of your male or female friends, are they hot or not? And part of the fun of that was,

the serious side to that was the technical problem, how to register that vote without I'm trying to get a page refresh. And I could go on, but there was a community of people that were making micro-innovations that essentially resulted in the like button, that Facebook, Facebook eventually picked up. And Facebook's big innovation was nothing to do with the like button. It was to do with the creation of a business model that exploited the like button. Yeah, that's where they, and it took five years to do that anymore. But we'll get there.

I think, yeah, and the fact that Zuckerberg, the light. button was proposed on Facebook and Zuckerman didn't like it and it had to go through many, many stages to even get anything like it. The Amazon had it because they wanted to, they wanted to be customer-centric so it was a way that they could show customers they cared. But every different group had a somewhat different reason to do it. But the problem, but the technical problem, well, I should say, I was amused that the reason, one of the reasons that, and you go with thumbs up rather than a heart, was, was, um, um, um, um, some, um, what's her first name?

Is it Linda Mueller? Is it, I forget the first name. Yeah, it was CEO of, um, and, uh, CMO, Anna Mueller. Yes. And that was, and she was a CMO at, was it friend feed or, or was it, Facebook, but all I know is she's basically said, I don't want to have my hearts all over my website. No way are we going to have hearts. And then also YouTube did the same thing, but within a few weeks they switched to a thumbs up icons. So somehow everyone started with a

heart, meaning I love it, and ended up a thumbs up. And it was interesting. This person you're referring to Anna, when the company she was working, for almost with the with the heart symbol just for reasons we actually spoke to her for reasons she couldn't recall just had a violent negative reaction to the idea of a heart button being all over her her social media feeds and so I guess this is a bit like science too this is what you make of like the frozen accident so that one person at that particular moment you know nudged the trajectory of something and and you know that sort of essentially locked us into a path So once you've got sufficient critical mass around the like button, it really became a global standard.

It's not a 100% standard. There are still some sites in Asia that use stars or hearts or other symbols. But by and large, most people centered on the like button. Of course, it starts off as a thumbs up and a thumbs down button, but pretty quickly the thumbs down disappears. Yeah, interesting. We'll get that. It was interesting. It turns out disliking isn't particularly useful or people don't like to do it.

It's a little bit redundant in terms of information because it's just the absence of liking. And from business perspective, not liking doesn't really help, I suppose. And it doesn't help you with a business model that's based on attention, which is, you know, Facebook centred upon. And I was intrigued to see, I mean, the rise of it from 20, you know, Facebook was the one who really popularized it in 2009. Let's put it in perspective. If 2005 was when Bob and others were playing, and Facebook really popularized it for an internal

reason that changed the world, but it took four years for them to do it. Even then, there was a lot of resistance, and it wasn't obvious that this was the thing to do. But that was in 2010 or so, and by 2023, there were 300 trillion-like presses, and over 60, trillion like presses per year. Let's go now to how stuff gets conceived, your chapter two, because that, to me, is really resonated. The idea, again, was a technical problem. And clarify for me, you talk about Hong and Young, and they were involved in hot or not.

Yes. Okay. And it was, was it their 12 lines of code, or was it Bob's 12 lines of code? that are the 12 lines of JavaScript. Well, it really depends on exactly your definition of the like button. So there were various people that more or less simultaneously hit upon the idea of using JavaScript to do a local computation that would get around the time limitations of the dial-up internet. And Jamesong was probably one of the earliest ones,

at hot or not, but he didn't, he didn't associate it with a, with a thumbs up symbol and didn't associate it. So it was an instant voting function leveraging JavaScript, which was an idea that the Stoppelman at Yelp also used, but to associate it with an emoticon of the thumbs up or the useful funny call symbols was was yelp's innovation but um again at least 20 innovators in that period of time in that 10 year period and um all contributing and all vaguely aware um without without being aware of the precise details that other people were were doing similar things the ideas were indeed were indeed in in in the air and we spoke to the uh we spoke to these people about sort of

Well, where did you get that idea? And some of them were, like, vaguely aware that Paul Burkart at Gmail, for example, was doing something similar. But most of them couldn't remember where they'd heard it. But they could remember going to lots of meetups and basically boasting about their latest work, which essentially became public property. That was the culture of Silicon Valley at the time. And it was this, as you say, it was in the air at the time in many ways in the, in the, in the, in the, in the, in the, in the, in the, in the sea of.

ideas that are going on. But interestingly enough, there were so many things here that were done that were done for not the original purpose. So the purpose of all these people, they were all creative people trying to solve a problem, which was how to do something without changing page, without going off the page. Well, there was a commonality in what they were trying to do. It was essentially technical, and it was to do with constructing a feedback loop in the absence of high-speed information exchange. But each tactical purpose was subtly different. So, for example, what some people were trying to do with the like button

was to declutter content feeds. Because if you do succeed in getting users to submit content, one of the problems you have is that most comments are uninteresting, right? People all say good, great. And if you have to scroll down a page of great and goods to get to a substantive comment, that would make for very uninteresting newsfeed. So that was one motive.

Some people were going after that. Others were going after the drama of an instantaneous vote. Other people were trying to rank content using cumulative votes. Other people were trying to encourage people to buy things by showing that other people like those things and people rated them as reviewers. All subtly different, but none of them amounted to the grand plan. of a new business model for social media.

So it's a curious sort of mix of convergence and divergence around tactical or sort of tinkering elements that emerged to have a bigger purpose that nobody had conceived of at the time. Which is characteristic, not an exception. It's in science and every other areas. It's these things you don't know what they're going to, you don't know, you don't really have an idea in advance

of how significant it's going to be. but you have the technical expertise to solve it. And what intrigued me, I didn't even realize that they use JavaScript, but they actually repurpose JavaScript for a task that JavaScript inventors never envisioned, you know, to basically allow you to work within a page without leaving it. And so they already, even JavaScript was not planned to do what JavaScript did. And they repurposed it. And then they did something, and even they didn't know,

what it would be used for. I love you talked to Hong to think back that time and you found him again, I'm willing to take much credit if he invented the like money. He said, yeah, so look, I mean, I don't know. Okay, we definitely invented

independently, but I could never say with certainty that we were the best of my knowledge we were, but who knows, you know. And that kind of Well, that was Lawrence, that's another sort of peculiar characteristic of the interviews. which is, so not only no grand plan,

but it was actually quite hard to get people to tune into the day when they made their contribution. And the reason was that that day, from the perspective of the moment, was just like any other day. Every day the programmers were facing, and the entrepreneurs were facing tactical issues. And this was just like, you know, tactical issue number 514. It later emerged that it was a particularly,

you know, pivotal detail, but at the time, that day, a number of people sort of said to us, you know, there's, you know, that day didn't, was not different from any other day. They didn't have any special sense of any significance to that particular day. So I think this is, I think the idea is, is modular recombination. We think of innovation or discovery as being some, you know, great instantaneous insight that's born complete in the minds of creative geniuses. But the truth is, it's the recombination, the tinkering, which leads to the recombination of elements and ideas, some of which turn out to be extraordinarily useful, but there's no grand plan. As Newton said, you know, anything it did achieve by standing on the shoulders of giants,

I later learned actually that that wasn't quite as magnificent as it seems because Robert Hook had accused him of plagiarism, stealing his idea for the inverse squirrel of gravity. And Newton in a private letter to him had said, and by the way, Hook was a small man with a stoop back. Yes. He said basically this phrase essentially is thought now to have meant that if I did steal it from anybody, it wasn't. from a short, crockid person like you. So even perhaps you had the idea of romanticizing his own contributions and denying the sort of, you know, the community of ideas and this sort of tinkering and serendipitous process of science.

Yeah, and in fact, Tom said if we had invented as someone else would have, because a lot of people do the same thing. And that is so often, I guess, there are rare times, and those are the ones people pick up in science with, you know, they might think of Einstein and general relativity or even special relativity. But when you look at, well, special relativity, you look at it most, or you realize someone would have done it. There are a few times, and Newton, I happen to say, is unique.

But even Einstein was, I mean, David Hilbert was very cool. The mathematician was very close. I've got a question for you as a physicist, Lawrence. So I've challenged myself to try and find an invention that, upon close scrutiny and invention or discovery was ex-Neil or de novo. It was born complete in the mind of a genius and there was no precedent. And I thought I found it in relativity.

I thought it found it. Yeah, relatively, definitely not. And actually, I can't find it. I can't find it. I don't think of, I mean, Einstein was amazing, but you can understand when you read Einstein's papers, you can you can say, oh, if I had that line of thought, I could have, I mean, I could have seen it. Certainly a special relativity. General authority required an incredible intellectual effort because of our new mathematics. But if you have the idea, the line of thought, you can see how you go from A to B.

And it's interesting, I've talked to my friends of mine, including some Nobel laureates. And there's only two. I can think, I find it in Newton. You can't see where he, when you look at Newton sometimes, you have no idea where that came from. And the other thing, interesting enough, another person that several of us talked about is a mathematical physicist named Ed Witten. Again, you know, you just don't see, it's not, what's the line of reasoning? It's like, wow, where did this come from?

But it's extremely rare. And usually there are many different people thinking about it from different ways. And if someone hadn't done it, someone else would have pretty soon. And that's the great thing about science, too, because, again, it's a social activity. And James Wright has written about this a lot. It's not done. The conventional public view that it's genius is working late at night in their rooms in doing something.

It's just not how science is done. And generally, if someone has been working in their room for 20 years, you know, they're a crackpot. Right. They'll talk to others. And you should be wary of the people who make those kind of discoveries. So one of the things I've tried to do since writing the book is to think really hard about whether the like button is an exception or the norm. So my anecdotal survey of the particular collection of inventions and discoveries I've looked into deeply all turn out to be the pattern of the like book, even if they, the like button, even if they look otherwise on the surface from the received stories. But there's actually, of course, some new big data science around the science of science,

which puts a number on it. So you're probably familiar with the paper that looked at 1.2 million biomedical papers and classified the knowledge categories of the grants that funded the proposals and the findings of the papers that resulted from those grants. And they found this alarming number, which is that 70% of papers contain knowledge categories, reference knowledge categories that were not present in the grants. And you can reasonably assume the scientists had every interest in anticipating all interesting findings from their grants.

So one number that we might bandy around here is the science is at least 70%, at least the majority of the time, serendipitous, in a way that belies the hero stories that we tell about it. Oh, absolutely. I'll put that more directly. I've said it before probably on this podcast. But as a scientist who for the entire time my professional career was getting grants, I view grants proposals as creative lying because you make a proposal for what you're going to be doing three years,

grant could be three years or five years, say, what are you going to be doing three years or five years down the road? But in my case, I felt if I knew what I was going to be doing three years down the road, then that was a failure. You know, because events would happen. There'd be interesting discoveries that would change. So you try and write a vague thing that's going to, I'm going to go in this direction, but it can be applied, you know, to occur with reality because the universe surprises us every day. And that's true in fundamental science, but clearly true in innovation, especially true now in this world where new things are happening every single day and that are changing the world. the interesting thing, and I want to talk about that, because I get reading your book, I kind of give Zuckerberg more credit, but for understanding the significance, potential significance

like Biden, even though he was told against it. But one of the things you talked about that was interesting when you talk about how stuff gets conceived is when something works, how quickly it gets adopted. And the example you gave was Humphrey Davy, someone who I hold in high esteem, mostly because he hired Michael Faraday, who was an uneducated bookbinder's apprentice. But Humphrey Davy was dealing with a technical problem was how to avert blasts and coal mines and came up with, you know, and lamps would boom.

And but the minute, what's amazing is the minute he discovered, as you point out, the minute he came up with a workable device within weeks or, I mean, it was used everywhere, right, within a very short time. And similarly, what was good about Hongs like Barton, you point out, is that easy feedback, but the other thing that made it so useful, and this is where we get to Facebook, I think, is that it automatically provides tremendous amounts of data.

And the realization that data would be useful was, you know, I think it, well, clearly Lefjian realized, two because he talked but but but but Zuckerberg I think you know it took five years but when but but realized that it could be used as a business model and in particular

by doing a plugin I never this shows how ignorant I get him you're saying I hadn't realized that when you give plugins when Facebook gives plugins it links back to you so you get the data and that I hadn't realized oh that's that that's that

That's the obvious. No one's being altruistic here and that you could actually use the data and the data could become the currency of the modern world. A lot of things to pick up on there, Lawrence. So quick aside on Sanfrey Davies. So I actually gave a lecture of the Royal Institution and I gave a demonstration of a methane explosion

using Humphrey Davies' lamp. And the story as told and celebrated is that, you know, a reverent Robert Gray of Bishop's Wehram in County Durham wrote to him and said, you know, we have to sort of prevent these explosions that are killing miners and derailing the Industrial Revolution, the mining of coal that powered the Industrial Revolution. Within weeks, so Humphrey Dabry had an idea and a working model, which was to wrap a copper gores around the flame in the miners' lamp. And within three months, he was rolling in.

it out and the rest is history. Well, I decided to take a closer look at that story and actually in granular detail, not so clear. So Stevenson had tested a prototype in mines before Sir Humphrey Davy had even constructed his first prototype. There were bitter recriminations between them. Part of the reason we know this is that Sir Humphrey Davy had called a Stevenson, an illiterate thief and not a particularly clever one. And so to save his honour, Stevens had to produce documentary evidence to say, look, this is the timing of events. I was testing a prototype, but you tested yours. There was somebody called Clancy that had a lamp in certain parts of the UK, the Clancy lamp or the Georgie Lamp or the

Stevenson lamp, were the names and the products, not the not. the Davy lamp. I actually, when I was demonstrating the Davy lamp at the Royal Institution, it didn't work. It actually triggered a methane explosion. It turns out that Herpy Davy's first lamp didn't actually work very well in the presence of drafts. There was a slight draft in the lecture theatre. So the idea that this thing was born perfect. There was an incredible convergence and copying between the, they all ended up basically with a sort of a chimney, a gauze arrangement, and difference in some subtle details, but tremendous convergence between these three variants that we know about.

So even where we have clear hero stories, not so clear if you go back and look at the dates in forensic detail. But the point you make about data, yes, indeed, the brilliant, the incredible impact of the light button, essentially in retrospect, looks very simple. It's that it creates a feedback loop. which permits you to, and an atomic unit of information, whether somebody likes something or not, which turns out to be a treasure trove of information that can be to be used to target sales, advertising, news feeds.

Because by simply saying we like this or abstaining from liking something, the cumulative total of those signals essentially tell. us more about us than we may know ourselves. You know, you get information on, you know, when they do it, also where they are, I mean, it gives you, by giving a like, you're giving all sorts of information that can be used, that you had no idea could be used. Which is the AI revolution that already happened. So right now we get a lot of speculation about, you know, what's the, how's the air revolution going to play out?

We already had an AI revolution. It was called the backend algorithm revolution of social media. So the thing that we've been talking about until now was the JavaScript computation, the front-end algorithm, which simply registered that a review had been submitted and liked by somebody. Once Zuckerberg and Facebook realized that they could actually sell segmentation information to advertisers, then you saw this proliferation of massive machine learning algorithms as the back-end algorithms. of social media that mind social graphs, you know, who are you liking, who is a connected by a like, you know, content graphs, what is being liked, segmentations of those, like, you know, which self-similar groups of people are liking the same things and the same, and the same way and socially connected, which was the massively valuable social media

revolution, sorry, the AI revolution that is, that is, that all, that, that all, already happened. But I think the prescience of Levchen and Zuckerberg was to realize that new asset, the mine of social graph information before it was even technically possible. Yeah, and I want to explore that a little bit more. Lebskin in particular, I found I would like to meet the guy because he seems like a visionary. But the fact that there's convergence, is because it's distributed, because it's a social network, literally science and innovation are social networks.

And I like that, I want to read the quote by Jeremy Stoppilman, who was he the CEO of, he was the founder of Yelp. He said, I'm not a historian and I haven't done all the research into who did, what, when. But typically, when there's something new created,

there's like a cocktail of different contributors, people working on similar things, all of them adding a little seasoning and salt and pepper from different directions. Press a little further, he allowed that priority ups real-time feedback buttons, he hadn't encountered anything similar on the Internet. And he also recalled having dinner with a key team member at Facebook

just a month or two after launching their new feature and freely talking about the success was having. And so in Silicon Valley, everyone shares some of their secret sauce. That's the way it used to be anyway. And I think that's the way happily it is in science now, at least fundamental science, people happily, once they publish, change, share their secret sauce. But that's one of the reason there's convergence, I think.

Well, it is. And it's clean, it's absolutely clear if you look at the granular information, but it's not the story that we tell. So one of my favorite books is this book by Ross and Nisbet, who was sort of social psychologist called The Person and the Situation. And essentially what the book is about is a fairly academic statistical treatment of the idea that as humans, we love to talk about. the agents in a situation, you know, what Fred did, you know, what Newton did, and not the

context. But that in psychology, if you know everything about a person, you've done all of the personality tests, you know the person's history, you can never predict they point out more than 30% of the, of the, of the behaviors of the person. And the bigger contribution is often from context, which is just messier, right? If you want to analyze the context, you essentially remove the convenience of the scientific hypothesis, that it's just the agents and their actions. We have this sort of replicable, comparable phenomena, and you end up dealing with complex, unique situations. So we love to tell stories about agents, but in fact, yeah, you have this sort of soup of ideas that multiple people are tapping into. So if we paint a picture of that,

what was going on in Silicon Valley at the time that we have this phenomena of multiple originators and contributors. You know, why is that? It's that we already had a symbol, a gestural symbol that was part of language and culture. No accident that the Gladiator One movie was from that period of time too, perpetuated the myth, but in culture a myth is a fact, if you know, people believe it, perpetuated the myth of the Romans did this and this to decide the fate of the fallen gladiator. We had a common problem, which is feedback in the era of the dial-up.

We had a bunch of startups that were very fragile, so they urgently needed to experiment their way towards a conclusion. They were sharing everything. That was the culture of the time. And they were borrowing, you know, many pre-existing inventions. They were borrowing the module, if you like, of, of JavaScript and the idea that it could be used to do social computation and combining it with the like symbol. So it really was plainly a firm end of ideas with, you know, many people

circling the solution. And when the solution was found, it propagated like wildfire, partly because it worked, but also partly because it had, it had been influenced by this, this Steve Krug idea of, you know, memes which travel. which is, from a scientific point of view, it's like another dimension of evolution. You know, our social power, our superpower as a species is, is not really the opposable thumb,

which is shared with some primates. It is actually the capacity for social learning. We don't actually have to jump off a cliff to find that it's a bad idea. We can listen to somebody's story about jumping off a cliff or observe somebody jumping off a cliff. And the light button essentially is the epitome

of social learning, because it involves choices about who we learn from. It involves homophily, which is this delicious ambiguity of the word like. I like the content, but I am like you. And it also involves reciprocity and recognition, which is thank you for letting me learn from you. Hence, nobody ever needed a manual to figure out, unlike the VCR, nobody ever needed a manual

or an explanation as to how to use the like button because it was already in human bio. culture. Absolutely. I'll talk about the neuroscience in a bit. You just brought up evolution and you just got, it was a wonderful segue. There's only one bone that I pick here that I disagree with you. He said, you talk about development of this. He said, people create novel technologies not by discovering something new or inventing, but by putting together different Lego box, so to speak. Okay. And then you say, and thus there is relentless progress, a kind of technological evolution, although it's not of the Darwinian kind that operates by natural selection.

Instead, new things come along as completely new combinations using new principles that keep adding to your Legalset. Well, I mean, I'm sorry, that's exactly how regular biological evolution works too. There are systems, there are biological systems that are using some, that are using some new code, and they get adapted and applied. And, you know, the example of molecular motors and all of these. things that look like their intelligent design or the eye the eye didn't develop there are all these components that were used for different things and that evolution discovered independently and then

combine together a new combination so i see it as exactly darwinian evolution and it is natural selection because the combination that works is the one that proliferates so i i i don't know i should agree with that i maybe i didn't express myself clearly what i was what i was referring to with that sentence you quoted is the sort of the statisticians use the urn experiment you put your hand in you pull out a ball and you look for the probabilities and so on

the thing about the the thing about Lego and invention is that the set is constantly added to because not only the bricks go back but the modules go back

so the so the you know when we when we invent a GPS. Not only the resistors and capacitors that go into the GPS, but the GPS itself goes back into the urn, and we may pull it out then with another module and make an iPhone. So it's sort of modular hierarchical recombination.

That's how the eye evolved, I think. Exactly. And yeah, so I find it, I mean, it's not too surprising that it is evolution because that's the way things work. And it's a very effective tool for getting, for solving problems when you don't know what the problem is. Because that's what evolution does, right?

I mean, it's, you know, it turns out that this thing happens to allow you to reproduce more effectively, but it wasn't the reason that thing was developed. There's no reason behind it at all in that in case of evolution. Right. And I think the additional human dimension to the evolution here, Let's say that it's technology and ideas evolve in a form of natural selection. I think the uniquely human element is this extra dimension of evolution,

which is not just the survival value of the behaviors and innovations, but the ideas themselves are subject to evolution. With the same iterative effectiveness as evolution, but also the same quirkiness that evolution doesn't explore every possibility, every branch and does get locked into frozen accidents and so on. It's not a complete search

algorithm. No, no. But, you know, it's like intramarers that fail. I mean, evolution fails in most of the time. People don't think evolution is directed. It's no in no way is it. And all these misconceptions about evolution. You know, 99.99% of species that ever existed

have been extinct. I mean, it's true for businesses. probably true as well. I mean, there's a remarkable similarity. So I find that fascinating, that similarity with science, a similar evolution. Before we go on, I want to, before we go on to the thumb and neuroscience, which I want to get to, I can't resist going back to Levitgen, who I find, as I say, fascinating. He talked about, he gave a lecture to or talking about substrates, and like the Lego blocks, but this is what was fascinating to me. If PayPal, the substrate was transactions.

What's the substrate in which the business is going to be based? It's transactions. And I don't know if he was the first one to say it, but he realized that the substrate of the modern world, the Web 2.0, 2.0 would be data. And that is the substrate. It's data. Data is the commodity that has changed everything.

And I don't know if he was the first one to really think. in those terms. But I found it fascinating. I think you're going from transactions to data as a way, because for much of the web, it was, development of the web was trying to find a business model. And only when you realize, and I, again, I think to credit him, at least from what I can tell now, Zuckerberg too, realized that data was a commodity. And that plugins was a way to make data more of a commodity because he got more information. Other people, other people would be doing things and you be getting the information for free and the data which would make you more valuable. And so not only did Lezsche can have that really important realization, but I also didn't

realize that there was another kind of a ha moment or another kind of tool that he was involved in, and it's called the Leveskken something else, I forget their name, but it was really optical character recognition. You want to talk about that for a second? Yeah, so one of the, this is not very related to the light band, but very related to this idea of technological evolution. So another sort of technical puzzle that needed solving at roughly the same time as the light button was occurring was the was fraud basically, internet fraud.

I mean, posing as somebody that you're not and so on. And they, so they needed, and especially sort of computers posing as humans, you know, bots trying to pervert payment mechanisms. So they needed some form of proof of humanity, if you like. And the insight, and Lecham was involved in this, but he wasn't the only one. You know, it's one of those stories where there was actually like two technologists,

two inventors competing. And so one of the insights was the observation that, you know, what can only humans do? And it was like the observation that optical character recognition was really bad, like the software didn't work that well, and that really bad handwriting

was something that only humans could decode. So you had this idea of like, you know, messy overlapping letters and a human figuring out what it said as being the proof of humanity. And it's quite a fun story because as Max was developing his own version of this capture technology, he was actually in communication with a hacker that was trying to break his system. And so they had these sort of like caustic interchanges about, you know, who was winning in the arms race of using our human cognitive quirks to distinguish between sort of botten humans. And essentially, Max, one, having pulled an all-nighter, Max won and got a sort of a grudging recognition from the hacker that he had won. Yeah, that's one of the rare occasions where the offensive technology wins over offensive. But once again, the Captcha bit, which was designed for fraudulent transactions there, Captcha became, you know, it's everywhere, right?

Whenever I log in anything, I now have to do some. There's a bit of a Humphrey-Davie story in terms of like who, you know, because there are two individuals contending for recognition. But it's, yeah, it's the result of a successful evolution, lots of unsuccessful things tried that we don't even remember and then this widespread universal standard that has really stood the test of time. I mean, most technology comes and goes or doesn't even come, but capture and the like button stayed. Yeah, and the other thing that's important differentiation between sort of natural evolution and technology is and hardware versus software is that normally it takes

decades for this, even if you have a new idea for the technology to take over. It's often and resist it for a while and take over. In this case, the resistance may happen, but it can happen, but it's so much faster with bits rather than it's, as I say, with data rather than materials. It can be done so much faster. And that's the issue, that's the good and bad of what's happening to us right now, is that the new technologies for which you don't know what they will be useful for

can take over before you know what they're useful for. and then we'll get there to the end and the end of the book as we talk about the good and bad of the like button. But just briefly, because I found history as someone who's interested in history, the thumbs up versus thumbs down. The fact that, as you say, Gladiator was around at the time,

but there'd been James Dean and there's all that. And the thumbs up was first, you say, written about in 1917, but the person who promulgated this myth was a preacher named, Talmadge. If you could just briefly tell that story, because if we are talking about thumbs up, we might as well at least talk about how it's not right, true. So a very important aspect to the like button is the graphic and the cultural element,

the gestural symbol that we're familiar with. And it was deeply embedded in American culture, and too many iterations to explain in detail. But you know, you have the fighter pilots of World War II that couldn't shout, my plane is okay, over the noise of the engines. So they gave an an okay sign you had the you know the thons with the sort of communicating uh something is cool and you um and but most americans believe that it will it will have come from it came from the for the roman colosseum although interestingly the italians called the thumbs up the american gesture and it all actually originates from a painting a painting from the 1870s by a paint a french painter called jerome that was very popular with a rising american middle class he

He painted sort of classical scenes. And there was a famous painting where he paints the vestal virgins in the Colosseum giving the thumbs down to the fallen gladiator, a very dramatic painting. And this really sort of sees the imagination of Americans and was amplified by the father-in-law of the painter who had perfected a cheap painting reproduction technology that was able to produce these classical paintings in high fidelity and sell them at low prices to the rising middle class.

And also by an incredibly prolific preacher called the Reverend Talmadge, who his numbers are impressive even in modern day terms. So he had a, he had a, he had a, his spirit. His sermons were franchised to, I can't remember the number. It was either 15 or 30 million, I believe. People every week read his sermons because it was franchised to thousands of local newspapers. And he was particularly fond.

We think he was familiar with this painting. And he was constantly using the thumbs up and the thumbs down as a sort of moral metaphor. Will we give the thumbs up to this social phenomenon or the thumbs down? It was like sort of redemption and damnation, the thumbs up and the thumbs down. Very fiery, very eloquent, very widely distributed. And so we think that we know that this is where the thumbs up came from. Well, we have good evidence from coins and Roman literature that that was not what the Romans did, in fact. The Romans probably, they probably had a gesture like this,

which is the sheath thumb, which is the sheath sword, if you like, the symbol of the sheathed which is salvation, and probably the downward pointing thumb or the unsheathed thumb, which is, you know, finish him off. But, you know, in culture, a repeated lie becomes a fact. And in fact, in Italy today, it more or less has the American meaning, but historically, this symbol in Italy and in parts of the Middle East meant something rather obscene, as the Italians call it the American gesture. So this is the cultural strand

and depending the technology. Now, if you're a very technocratic, scientific person, you may regard this as an irrelevancy, but it's absolutely central to the evolution of the like button. It's essential to that feeling that nobody needs to explain to us how to use this symbol. We know exactly what it means, and we are delighted as human beings

to use it. Yeah, I think that's a wonderful story, especially in a book about the thumbs up signal that most of us use it, you know, wrong, or at least where do we think it came from us wrong. Now, you spend some time, you know, why do we like, like? So, I mean, I think all of us intuitively understand because we get a rush from being liked,

but there are obviously a lot of science behind that. And you begin that chapter talking about someone in an MRI who's, you could see the brain activity when they're liking things. And why don't you just talk about, I mean, clearly it's, there's dopamine, there's, there's, there's, but why don't you talk a little bit about what you learn about the science, both the neural activity, also from my friend, Nicholas Christakis, who I've had done a podcast, but about that about hypersociality, about social learning, about the, the evolutionary basis. it's really how we as a species have been successful. And finally, about homophily, which is a term that you talked about, but the fact that we like things like us. Yes. And so why do you talk about that?

Because all of those contribute to why it's so addictive, to why it's so successful. Well, one of my favorite books is Theo Wilson's, conciliants and you know the idea that you you can't understand something unless you drill across disciplines essentially you drill down through the sociology you know into the into the psychology into the biology into the chemistry into the physics and it's the idea that we can have explanations across all of these all of these disciplines and it seemed to me that you couldn't I mean I could have stopped by by dealing with the history of the technology of the like button, but really, we couldn't understand the full significance

unless we looked at the culture, which we just talked about, and also the biology. So the biology, two significant things we talk about. One of them is, you know, why did it really stick and take off explosively? Like, what biologically is going on that permitted that? And also created tremendous sort of psychological and social problems. And it turns out that if you put people in a functional MRI scanner, that we can measure the dopamine release in the nucleus accumbens in the brain, which is the same signal that you get from sex

or gambling or other sort of pleasant activities. is, so it's, it's, it is the, the internal behavioral reinforcement mechanisms of the brain that is triggered by the like button. Interestingly, being liked and liking is exactly the same signal. So actually, we, we, we, we like to like and we like to be like. It's exactly the same signal. And it's, um, and it's addictive, of course, you know, the, the, the, the, the, the, the, the, the, the, the, the, the, the brain is up, our behavioral repertoire is modified and either reinforced or weakened by these dopamine events. So it is an evolutionary, very powerful and useful function. But it can be hijacked.

We do have problems of addiction and so on. And we have those problems for the like button, for social media too. And so we spent some time pondering, you know, why did all the problems arise? because it's the same, it's the same natural mechanism that we, we would employ in the playground. You know, the kids will like each other and receive a frisson from that, or they're receiving an unpleasant frisson from being ostracized by their friends. I think it's the volume. So how many people could I, I meet called Lawrence in a day,

and how many times could I be liked or, or dislike by them? You know, maybe about five, let's say. That would be tough to meet, to be liked by five people called Lawrence. in one day in the real world, online, of course, we can put a couple of extra zeros on that. And, you know, that at a formative stage in the brain development, in the social development of young, especially for the female teenagers, is the source of the problem. They just cannot handle that volume of evaluation by others, and it can cause depression, addiction, all sorts of problems.

that's, so that's in a way the biological potential and potential for abuse that the, the light button is tapping into, but it's also tapping into some, some higher level evolved behaviours. So, Nicholas Christakis, who you referred to at Yale, he talks about the social, what he calls, a social learning suite, which is the, the evolved behaviour which characterize our species, which permit us from, to learn, not just from our own direct experience, but to learn from each other. And some key behaviors that are built on top of the dopamine mechanism, the plumbing, if you will, are homophily. You know, we like to learn from people like us, because it's logical that the lessons derived from experience by people

like us are probably going to be applicable to us. And so, So hence, this sort of ambiguity of, I like the content and I am like you. I like the content, I like you, and I am like you. Homophily, the like one taps into homophily. It also taps into mild hierarchy. Why mild hierarchy? Because the hierarchies of the animal kingdom are usually based on violence or the threat of violence. In human society, it's the milder form of hierarchy, which is who is worth learning from.

and who's worth learning from people with lots of likes on their light count the people that other people appear to be learning from so we have this hierarchy of the comparison of the light count of different we can call them sort of influences social media influences we want to learn from people that other people are learning from

and then we have recognition what is the currency for this for being permitted to learn from somebody like us that other people appear to be learning from it's that we recognize their teaching or their influence by by providing this social currency of the like button. Thank you for letting me learn from you. So really not surprising that in a sense, these engineers in Silicon Valley that were coping

with a technical problem of creating an information feedback loop on the internet, knowingly or unknowingly, were tapping into culture and also tapping into a rich biological source. I think not very consciously, because if you talk to them about, you know, why did you do to this, where the idea come from, it's probably some survival bias, which is we don't hear about the ones that we're not thinking about these things. And it's probably just, you know, subliminal recognition. I mean, I think these things were in the culture. These things were these ideas were known.

and they were tapped into. By the way, a delicious detail on the cultural aspect that I forgot to mention is that the like gesture received an extra boost in American culture from the Gladiator One movie directed by Ridley Scott who intended to turn down the movie. He was by then an established director and he knew he was going to be offered the movie.

He didn't want to make what he called another side. Sandals and Toga movie, but he was shown the very same painting by Jerome, and he was impressed by the drama of the painting to actually take on the Gladiator One project that perpetuated the myth of the thumbs up and the thumbs down symbol. And this was happening, I think it was 2005, precisely when it was all happening. So it's a delicious sort of cultural conjunction That's okay, it clearly, I mean, it's, it clearly tapped into something deep. And you hit, you hit this idea of, you know, I've run workshops on xenophobia, which is in groups and out groups. And so why you don't trust out groups.

But the converse of that is really, it's sort of obvious in the fact that we, we see people like us as the best, I mean, as, as, as, as, as, uh, as examples of ourselves as prototypes that we so why not if you want to learn something you want to learn for someone like you because you figure if they're like me they're going to it'll it'll work for me exactly and and the other thing that you talk about this hierarchy of who you want to learn from well and you know we have established experts and we you know they have credentials and things but it is interesting that you know the fact that you know people who have a lot of likes are somehow viewed as more reliable if they like something. And as I say, I think it contributes something I'm not particularly happy with,

which we'll get to, which is influencers. But it is interesting that one of the things that you talk about that discovered in these MRI things is that patients are more apt to like photographs that have already received a lot of likes from others. So, you know, again, I think it's what we might call peer pressure too. You know, if other people like it, I should like it. And that's an interesting issue. But fundamentally, all of this works and feel so good for us, the dopamine and everything

else, not just because nature wants us to feel good, but because being liked increases your survival chances, ultimately, I think, if you're in the group, being liked increases your survival chances, and therefore people like to be liked. But at the same time, and this will get us to influence us in a moment, this one quote that you have, Lemke, it talks about, he says, essentially, what social media has done is taken the human connection and drugified it by distilling it down to its most essential reinforcing properties, namely the like button, basically. And so it's taken all of that and put in one incredibly powerful form that has, like all good things, with great power comes great responsibility, but we're not sure that we've taken any other than. And in a way, that's the explicit dark side of Steve Krug. So Steve Krug's big idea, this interface designer, interface guru, you know, he said basically, don't make me think. and the word he uses friction, you know, eliminate friction.

If you want something, an idea to spread, if you want a website, you'd probably eliminate friction. Well, friction is quite useful to stop, you know, to stop bubbles happening, to stop excessive things happening, to stop things running away. And so in a way, a consequence of successfully removing friction from the act of liking results, in this ability to be liked or disliked many, many times a day, more than we can evolutionally cope with, more than we will have ever experienced in nature, which is essentially the hijacking of this natural mechanism. Now, is there a cost to that for young adults, young teenagers?

Absolutely. They're very sensitive to the playground politics of who likes whom. Yeah, no, in fact, we'll get there into Jonathan Hayden and others. The other thing before we leave this section, which is interesting, you know, in terms of the dopamine, the interesting thing is it's released in anticipation of a reward. So, and that's very important here that people are, you know, it's not, you don't get toward, it's anticipating the reward. That's right.

It's not that you, yeah, the dopamine mechanism is a sort of difference engine. It doesn't just operate on, you know, something nice happened or not. It's, it operates in expectation of good things happening, the thrill of anticipation, but then also the reconciliation with, of the expectation with what actually happened. So an undershoot or an overshoot can either produce a dopamine deficit or a dopamine overshoot. That's, that's the way it works. And the biology actually is,

is very important in terms of dealing with the social side effects because, for example, there's an accommodation phenomenon, which I think is very well known for gamblers too, which is you have to turn up the volume control if you want to experience the same thrill on repetition. There's a sort of dulling of the sense and therefore a tendency to excess, which is part of the problem. Now, in real life, we have the, the frictions of social physics prevent us from overdosing, from overusing this dopamine mechanism. But in the digital world, all of those frictions are removed. Okay. Now, speak of frictions and removal, how you do it. There's a chapter, and we'll go out fairly quickly because it's technical, what happens when you click? But I guess

the important thing is that when you click, you're doing much more than you think you're doing. The code does much more that every, you know, that traditionally, you know, there are unique user ideas, cookies that give you information about you to verify you're the same person that was doing it before. But all of that means that that leads to a understanding of you by clicking. You are immediately giving a tremendous amount of information. I don't know whether you want to talk about the specifics of the coding that are involved in that. or not, but it's a really important feature of what's happening. So you want to expand on that? Yes, well, it's a, so the world of the back-end algorithms, the algorithms that analyze

your social graph and your content graph and update them every time you click or you fail to click or the people that you like or are connected to click. That's a fascinating world. I mean, the algorithms are all subtly different. Essentially, they're all doing the same thing, which is figuring out who you are and what you like and what people like you like. And they're using that to do a couple of things. They're using that to predict which news feeds you may want to see. And that's important because we're in an attention economy.

The business model is to retain your attention and collect further information and use that for monetization through advertising. So, you know, it's important that the service of providing you with newsfeeds is as targeted as possible. And then we have, you know, the advertising sort of value proposition. So it could have been in a different, in a parallel universe that users got charged for their email and their news feeds. But that was not the business model that Facebook hit upon. And so essentially, we're providing with these free services. We inadvertently or perhaps knowingly provide this information on ourselves in an atomic form. It's just like, not like, like, not like.

But cumulatively, it says so much about us and our connections and our wants. And that is useful for targeting and measuring the effectiveness of advertising. So that's where all the money is coming from. That's where they, because that's another part of this that we haven't spoken about yet, you know, which is, how does this thing sustain itself? as a business because we can imagine everything I've set up into this point being true and resulting in a you know some really interesting curiosity sites that are fun to play with but it wouldn't be a big business the reason as a big business is essentially it solves a really fundamental and massive problem in advertising which is we don't know which part of the advertising

is is is is effective now and as we you know as we think about um the thing to The thing about these algorithms is that the sort of primitive sort of use of the word algorithm is a set of instructions, a fixed set of instructions. Where we can reasonably impute, you know, a design intention to the programmer. The sort of algorithms we're talking about here are machine learning. They are massive data arrays that are in constant motion, massively, massive dimensionality. the operations were designed, but whether you get this news feed or that news feed or you get connected with this person or that person, that is an indirect outcome of these algorithms.

So it's very interesting listening to the regulatory debate where we say, well, you know, Facebook shouldn't do that or Facebook should tell us this. They wouldn't know because these algorithms are vast self-sustaining, self-evolving entities that have indirect consequences that are not knowable to the, to the, to the, to the, to the, to the, to the, to the, to the, to the, they have, they have, they have a life of their own. So this is if, if you like, another dimension to evolution, which is algorithmic evolution. You know, why the algorithms precisely the way that they are with the weights that they are? This is the, the problem of the explainability of AI. We, we, we, we, we, we, we need to, we need to think about more subtle intervention strategies. It's just not as simple as saying, you know, don't do that, redesign the algorithm. Well, and you know, but it did take, but what there were people who realized the value. And as I say, I guess I hadn't appreciated how much Zuckerberg played a role in that. You talk about when he, when he, in a developer conference of 2008, which is I guess a year

before they even introduced their own like button now that I think about it. He touted the marvelous potential of Facebook's social graph. Most people had never even heard the term before. And he said, the social graph is changing the way the world works. We had a time in history when more information is available and people are more connected than they've ever been before. And the social graph is the center of that. The idea of accumulating explicitly that social graph and utilizing it for business,

which would essentially inevitably become advertising, as we'll talk about. but that explicitly recognizing the value of that social graph and the data that comes from, not just from the like button, but the like button gives a tremendous amount. And then there was a feature called edge rank, which was something implemented in some way to sort of eventually lead to monetizing that. Can you talk about edge rank first, if you don't mind? Actually, that's beyond my... Oh, okay.

It's like page rank in Google, but it's a way of prioritizing content according to infinity, weight, and decay time. Basically, it's an algorithm that takes that data and ranks it in a way that the algorithms can then use to somehow convince you to do things. So we probe quite hard for the details of the, the algorithms, the different social media companies, and basically it's all, it's all mostly secret. There are a couple of professors out there that are trying to sort of, you know, infer what

the algorithms are doing, but a couple of things are clear. One of them is that the algorithms are constantly evolving in their sophistication. The second is that they're different between the companies, very different, you know, different metrics, different things being maximized. And thirdly, that they, you know, the dimensions of these algorithms and these data sets are, um, are mind blowing that they, they sort of have a life of their own. It's, it's, it's, so therefore, um, in a sense, there's a, there's a, there's a, biology of these algorithms. There's a, there's a natural history. Discovering what the algorithm actually does, it's emerging consequences. It's something that you have to do experiments on.

It's not like looking at the code and saying, oh, it's designed to do this and this, and therefore it does those things. Most of the consequences are emergent. Now, you know, it raises the question, is the like button still important? Because now we have all sorts of metrics and, you know, we have like predicted lifetime user value, propensity to buy, you know, duration of attention deviation and

evolution in attention span value of attention we have many many metrics as far as we can tell and we spoke to the algorithmic specialists in specialized agencies that monitor the effectiveness

of these algorithms and metrics the light button is not not the only thing which is being measured but it but it is still relevant probably because of its elegance and its simplicity. If you were to choose a single metric, this one looked like it will last the test of time because it is the most basic thing a person can do to say I like something or not like something. And while it's ambiguous, it's just one signal. There are many ways of

interpreting this signal. It's cumulative value is enormous. in terms of just how much it potentially tells about this, and how much it predicts that behavior. So the like button is no longer the, I mean, there are literally like 100 metrics that are used by advertising analysts now, but the like button is still very, very important. It's still essentially the atom of user content.

You end the chapter about what it does with a sort of a very literary, paragraph, which I like, so I'm going to read it. We imagine the process as a light rain beginning to fall on the surface of a still pond. Each droplet causes ripples and the concentric circles spreading across the water increasingly overlap. It's the same for all the times you've clicked the like button. The tiny waves of influence begin at the interface of the app, flow to the code-based processing them, and collide and overlap in vast lakes of data storage.

Through a mysterious and complex process that no one could retrace, all those taps of the like button combined in ways large and small to change the course of the future. And they have certainly changed the course of the present because what likes have done ultimately and the data associated with them has has has has is the business of the internet and the business of likes. And it's basically changed the economic activity of the world and advertising. And as as someone said, perhaps in terms of sure impact, the like button was one of the most successful pieces of code ever shipped. And that came from a business magazine. And you begin that, you begin that chapter

talking about someone who was an influencer and now basically trains, you know, helps people become influencers. And I'm going to be up front. We'll talk about later. I hate the idea of influencers. I find it because there are people who've done nothing but somehow influence without having done anything to be worthy of influencing for it but anyway in my opinion they just they're famous because they're famous and um but nevertheless the fact that the life button created data that no one thought about at the beginning made data become the most most valuable commodity in the world and um and um and um and changed everything and in terms in terms of how i advertising is done. So I've said it, but I'm sure you'll say it more cogently now. Why don't I turn it over to you?

Well, if you go back to this big idea of Web 2.0, the idea is that the internet becomes two-way, not just for reactions to content, but also the democratization of content creation itself. And so whether it's creative endeavors or advertising, essentially we don't need to be a rate. station or an advertising agency to create content anyone anyone can create it with cheap consumer electronics in their in their bedroom and that's the the world of the influencer which is probably the most exotic part of the book for me because i you know i haven't spent a lot of time in my life hanging out with uh you know 16 year old millionaire influencers that make short form videos but the way that that world works is that um a much larger number of people can can can

now create content and the digital format of these platforms and this content can permit permit a lot of complexity so in in the in the in the bricks and mortar world you sort of need like big agencies and big companies to do these things because it's there are scale efficiencies but in the in the digital world we can we can actually we can actually create content economically with thousands of influencers and micro-influences is one of the terms from the industry. And that is all sorts of consequences. I mean, essentially, everybody can be a publishing house. Everybody can have a megaphone. Actually, big businesses can leverage this body of people. There are all sorts of ways in which you can incent or contract with the key

influencers and the big potential benefit there is authenticity. I mean, you know, an advertising agency pushing a, being paid to push a product is, I mean, it's somewhat effective, but obviously of, you know, not

limited authenticity from the point of view of the consumers. You know, people making videos in their, in their bedrooms according to their own sort of personal passions and interests. I mean, there's often a high degree of credit. I mean, there's a, there's often a high degree of

credibility afforded to such sources. And it was fascinating for me to walk through this zoo, this new ecosystem of players and see an entirely complete new ecosystem, business landscape. So you have the influencers themselves. You have agencies, new digital agencies, which have grown up to specialize in some of advertising. You have metrics companies, analytical companies, that measure the effectiveness and the reach of thousands of influencers. You have universities for influencers. You have sort of training courses they can go on to be coached in the art of influence. You have black market like factors.

You have, you know, factories where, you know, people are using arrays of like computers on mobile telephones to evade the fake-like prevention software by having humans actually, you know, liking things in an irregular pattern that's hard for an algorithm to detect. You've got markets for like buttons. You can actually buy, you know, a thousand likes. There's a price. There's a market for that. You've got this entire ecosystem that's unleashed by the like button and the economy

that created it. Simply by creating an information-rich, cumulatively valuable, low-cost, universally recognized feedback loop. All of this comes from the genius of this dozen lines of JavaScript code that changed the world, not only the visible ways I've been talking about, but in terms of the complete reconfiguration of the advertising and marketing ecosystem. And by the way, in case anybody thinks we're talking only about social media, You know, UPS as a like button, open AI as a like button. This is now, I'd say, not so much a symbol in the language of social media as a universal symbol for any digital business. And it has those features, as you point out, that make it so valuable.

First of all, it's exceptionally clean and valid and analyzable. and so that it's great for, you know, people, it allows you to assess personalities in a way where you can test your hypotheses and when people started to do it, they started to look at Facebook likes and it was a great way to do it. It's also super abundant. There's just so much of it, four million posts every minute in 2015. And then the amount that's invested,

and the amount of sales, hundreds of billions of dollars, you know, arising from this, have changed the way advertising is done, has changed the universe of advertising in good and bad ways. And it's click-throughs are the number of people who click through, you know, on a sponsored post. It takes into a website to see more things. knowing that someone is like the similar post the past is valuable. All of this becomes a way to better manipulate people and more effectively manipulate people. And we have to ask at some level whether it's good.

But again, I want to turn to Zuckerberg. There are two quotes from him that, you know, again, Facebook probably the most, initially the most successful utilizer of realizing the power of like. And in 2007, he was already beginning to realize the usefulness for social media. He said, when Facebook was offered, he claimed a completely new way of advertising online. For the last 100 years, media has been pushing out to people. But now marketers are going to be part of the conversation.

And they're going to do this by using the social graph in the same way our users do. So he realized before the like button. But then as a Harvard business school, Leslie John said, one reason Facebook advertising could be effective is that a brand social media pages reaches a highly desirable audience, likes illuminate a path for targeting ads. And then more importantly, perhaps the fact that once again, Facebook brighten that path even more by encouraging other websites to adopt its like button feature. This is what we talked earlier as a plug-in because it shares their information, a brilliant change.

And all of that immediately means that they have access to all sorts of data, which makes them valuable to advertisers. And that becomes sort of that implements in reality something that was talked about before called a segment of one marketing. So maybe you want to explain that. Yeah, so I was actually working in my company. I can't remember the exact date, but before all of this happened, when a guy called David Edelman proposed the idea of segment of one marketing. It was the idea that traditionally we think about marketing in segments, you know, single mums and soccer mums.

and empty nesters and this sort of thing. You know, why? Because our information was not very granular and our analytics were not very good. So we generally couldn't handle more than a handful of segments. So we decide to, you know, serve females or serve males or serve upper middle class males or whatever. But very crude segmentation that didn't change very often.

The idea of segment of one marketing was that you could target every message and product to an individual at a point in time, that that would be possible. At the time, this was absurd because we lacked any of the analytical or data assets to do this. But now it's possible. And in fact, we interviewed David for the book, and he's now a professor at Harvard. And he just wrote a book on customization about the idea that we can not only target pre-existing products and messages to individuals, but actually we can, if the products are mainly digital, we can actually customize the product to every individual and every occasion. They can have exactly what they want when they wanted, and we can, with high

accuracy, know what that is. So that is a huge, that happens slowly. So we sort of, it may not be apparent just how big that changes, but that's a huge shift in marketing. Let me contrast the before and after just to illustrate how big a change that is. So in the old days, you had something called a marketing department where people would buy batches of data, you know, usually reports that are, you know, less than an inch thick, and they would analyze it by hand or with a, with a PC. And they would create a marketing strategy that might last for a year that would essentially say, target these 5 million users, if we contrast that to today, we have more information than we can consume, we have massive analytical capacity, we can target individuals, we can change, we can

construct a million segments, we can change those segments every second, and we can even customize the product and the information to the individual user, and we may not even have a thing called a marketing department because all of this might be just a function of the recommendation algorithms on the website. That's a huge change in the world of advertising. I point out it's a huge change because instead of hiring marketers, you're hiring scientists because what you're getting the scientists do is build algorithms to understand how to use the data. And that's the important things. So advertisers are hiring, you know, scientists because they're, because they'll teach you how to manipulate people better. And the, and, and here's the numbers. I mean,

after the like button was introduced in Facebook, it value, it went, in 2008, its valuation was 15 billion. By 2011, it was, it was, it was, it was, it was earning one billion profits on 3.7 billion in revenue, and by 2012, it was, its market cap would be $100 billion. That, I mean, and as, and interestingly enough, if you don't think it's from the like button, you point out, you know, when Facebook did its public offering in 2012, the Wall Street Journal covered the announcement, along with quotes from Ford and Coca-Cola marketing management, crowing about the results of their Facebook ads and the report was actually the second

in a multi-part series issued and the title of that series was the power of the like and so it has literally created drives our economy in a way and I would say negatively allows us each of us to be manipulated so much more effectively

than we were before and created this whole group of people who are now rich called influencers as part of this. this overall manipulation that just seems, you know, to me as an old person, it just takes a way that, you know, used to be learned their value. I'm not sure I would agree with that judgment, Lawrence. Not that I'm asserting the inverse, but I think there had been a number of books before

we wrote our book, which had passed a judgment on social media. And so we thought that that had been done. But we wanted to create more of a natural history that said, what is everything that happened good and bad? And how did it happen? And with the knowledge available at the time, what could we have done differing? I wanted to take a more neutral course.

And so I think there are certainly some harms that we deal with in the later chapters of the book, resulting from... We're about culture, by the way. But by and large, I think this is just an evolution of business, a different way in business. And there are benefits. I mean, I'm an amateur musician amongst other things. I'll hold off on that because I want to talk about your experience in music.

So let's get there. I didn't mean that the whole thing is bad. I just meant it's changed marketing and change advertising to make it unbelievably more effective by exactly doing what you said before, that some. Someone said, I forget it what rich person of a department store said, 50% of the money I spent advertising is a waste, but I don't know which 50%. And now this tells you exactly what isn't wasted. And it's effective.

And I don't mean it's bad. Influencers, I find tedious. But the fact that people, I mean, I benefit every day from being targeted in a way. I mean, sure, I buy things I need to do, but now I find things that resonate with what I was thinking about before. and we'll talk about your experience. So I do want to move. I think it's an appropriate thing to move to this.

The last two chapters of the book are unintended consequences. And then the future, I want to go over them fairly quickly. But look, unintended consequences are the norm. I mean, it's not just of apps. It's of science in general. Well, absolutely. Yeah, I mean, we talk about, so economists talk about externalities, which is like an unanticipated negative consequences.

Francis. The like button, an aspect of the like button story is that clearly we couldn't anticipate even the benefits. So what chance do we have of anticipating the disbenefits, the problems? Virtually none. So there's a sort of a, you know, as with our invention stories, there's a sort of a cartoon simplification of regulation which simply says, oh, bad things happen. The regulators must have been asleep. You know, wake up and more preemptively act against the obvious harms. Well, it wasn't obvious. And not only that, the regulators often create more harm because they're doing the wrong thing.

We'll see it. We talk about the future a little bit. I mean, the UK harms act, which is horrific. Yeah. It ends up, you know, it's to regulate something, but the effect is to get, is to kill free speech and to attack the victims more than the promulvators. No, I believe the regulators are sincere in their intent, but it's a very difficult job because essentially, you know, this is the problem from a regulatory perspective.

There's a new technology that you were not trained on. By definition being in a position of power, you will probably be the last person, the least expert person in the economy on that particular technology. It's moving extremely fast to this digital scaling. So whatever problems emerge really quickly, we couldn't anticipate the benefits, let alone the harms. And so unless you want to ban everything, you know, just ban the possibility of negative harms from technologies.

I mean, you've got to wait and see how it goes. So in a sense, we have the bumbling of the regulators may be a feature. This may be the equivalent of the entrepreneurs tinkering. The only difference in this particular case, I think there are two differences in the case of digital technology. One of them is it moves really fast. So you have to be probably more nimble than you've ever been before as a regulatory establishment. And the second one is you're dealing with increasingly cognitive technologies that are more and more likely to have.

This is not like something you shouldn't drink because it's poisonous. This is like inside your head. This is a particular type of harm that is high potential for these cognitive technologies. So we had a very interesting dinner in Washington where we assembled the top. regulators and regulatory theorists and we said basically if this is the problem

how can we up our regulatory gain and some really interesting ideas came up so one of them was the idea of a public observatory so given that the commodity here is data that data is proprietary right who owns that data the big social media

companies in order to see things coming we probably need some public access to that repository of data because the harms and the benefits are in the patterns of the big data. There's something around education, which is, you know, maybe at the speed of evolution of physical technologies, we could wait for the regulators to catch up in their education. But these, you know, for instance, take AI right now,

we probably need some urgent accelerated training of the regulators on these technologies so that they're, you know, they're competent enough to pass judgment. The other one is just the idea of the, you know, just the sort of implicit assumption of the legal system that you pass regulations that should not change a lot. You know, you sort of, you pass a regulation and then it's on the books forever. Well, when we're dealing with high unknowns and fast progression and significant consequences, we probably need more of an adaptive evolutionary principle. We probably need something which is more like adaptive regulation, which is Sunday. un-setted provisional regulation that we learn from and then can then can evolve further. Also, we, you know, another really interesting thought is, you know, one of the strengths of the

federal system is that essentially different places, different geographies, can try different things. And that could be a huge benefit if you think of it as a vast parallel program of experimentation. But if it's ignorance and competition, then it's, you know, it's not automatically going to produce that accelerated learning. So to make a, you know, it's a matter of fact that China, Europe and the US will go different directions on legislation, different states will go in different ways. But we could use that as a feature rather than the bug to accelerate our learning. So this to me is less, you know, less character. charismatic than the story of the like button and Jerome and Talmadge, but in a way a more fascinating problem. How do you regulate, how do you innovate in the regulations that regulate the technologies, which is like a double dose of unpredictability? I mean, how do you play that game? Because we have to play that game. I mean, the consequences of doing nothing are, it could be equivalent to the consequences of, you know, doing the wrong thing. We have to play that game. There are, there are emergent hearts. and there will be emergent harms with AI too.

How exactly do we play that game? Like Louis Pasteur said, fortune favors the prepared mind. And you can't always be prepared. And I want to get to this future issue and AI. But I want to just list. So here's some of the unintended consequences. First of all, it killed newspapers or it killed newspaper advertising because it killed the standard old advertising norms.

And it, it affects privacy, obviously. And some laws have already been, you know, there have been many suits about privacy. You're giving up your privacy when you do a like button, among other things. Allowed neuromarketing, which means you can probe, you can use neurological techniques to, in many cases, take advantage of. And that's something that Jonathan Hates talked about. and meta, you know, has admitted to using algorithms that get young people, take advantage of young people's neural lack of maturity in some sense, but also adults' lack of maturity, to, you know, and social pain to get them,

to keep them on. So it's allowed neuroscience be implemented in a way that does take advantage of people. And the other thing is that these are the bad things. it's produced social polarization in a way because by knowing what people want and giving them what they want and exposing them what they want, it gives people echo chambers and it allows people to exist in a world that is homophilic, that is just like you, and not realizing that there are other worlds that are just as real. So those are some of the unintended consequences, and we have to live with them, but you do point out that there are lots of positives and the connections.

I think the fact that people can be more connected to good things is a good thing. So I want you to give your example of music, to give it sort of an anecdotal example of that. Yeah, I mean, I don't want to be an apologist for social media. We do have harms, and we do have a difficult regulatory game to know what those harms are, know what the mechanisms are and have effective interventions. and the science of that is really complicated. If you waited for proof,

you know, the problem would be, you'd be on to the next problem, essentially. But I think we did see a tendency in relation to new technologies to take the benefits for granted and to focus on the negatives. It's important that we maintain a balanced perspective. I mean, how is the like button

and its associated business models change the world for the better, well, I think we're more richly connected to each other. I can more easily stay in touch in more ways with, you know, more people than I could previously. I think I do receive a better service. If I think about the pamphlets I used to receive in my mailbox and the curated information I now receive, you know, I actually, I guess there are different philosophers in relation to social media. Some people tried to sort of cut themselves off from it. I tried to make sure that the, I received the best possible service in terms of being targeted by the things that I want

to be targeted at me. For instance, I'm an amateur musician. I'm quite interested in the latest in digital organs, for example. And so I, digital pipe organs, I would want that information come to me. And I'm, and I'm much better taken care of than I was before the age of these. sort of these targeting algorithms. In fact, I could do something really remarkable. I, so I play wind instruments and I was quite keen on buying a didgeridoo and learning the didgeridoo as one of my instruments. And it turns out that this is a huge object of scamming. You know, they essentially tourists sold fake didgeridoo's and so the, how do you determine whether and digitally do it is authentic or not.

It's a sort of complicated problem. It's a dirty market. Well, thanks to the, thanks to this sort of edifice of social media, I was able to do research, do like research in a very sort of obscure area, do deep research. I was able to find Aboriginal cooperatives that made these things in Australia. I was able to talk to them. I was able to educate myself on what's the difference between a sort of a tourist a drudu and a real do-drudu.

And I was able to, I was able to buy one and have it delivered to me within two weeks. This sort of ability to navigate offerings and information is we take for granted. But, I mean, I've been a musician for most of my life. And this is a, this is a revolution. The other side of this is, which my daughter points out to me. So one of the things my daughter does is she's a, I come up with the technical word for it, but she essentially, she makes animations collaboratively with other people. And so she is absolutely a beneficiary of the creator economy.

So she's able to literally put together quite sophisticated animated films collaborating with with others and designing protocols for collaboration and, you know, reusing sort of art artifacts to express herself. And so I was, I talked to her extensively to try to sort of tap into the, you know, what her generation thinks about all of this. And essentially a major part of her creative existence

is possible now, but wouldn't it have been possible. 10 years ago. So we shouldn't take the benefits for granted. But the other sort of important aspect here is as we think about the harms, we must be careful that we're not solving for yesterday's problem. So we're still struggling in the regulation with, you know, how to prevent you know, young female teenagers from psychological distress by intractrum and social media. But we're already on the precipice of the next problem. There is a like button on the on chat GPT. You know, what is that doing there?

Is that to tell us what we want to hear when we ask a question? Is that to tell us the answer to our questions in a style that we want to hear it? Is it to learn our preferences so that we can subliminally advertise as part of the answers to our questions, you know, things that they think we should be buying? well I'm sure there's a I'm sure there's very creative utilities waiting to be tapped into there and also new new regulatory nightmares and so it's a it's a very it's a very pragmatic discipline sort of regulatory science because you it's a bit like management you you have no choice but to do something at a timely fashion even without complete proof and then you need and then you need to move on to the next problem and you'll probably get it wrong some of the time but that may be the best that

you can do. And you need to often go on to the problem that you created and solve that one. And absolutely, the regulation itself, of course, as you point out, you know, creates distortsions and problems. And it's not just in the internet. I mean, it's happened all the time when people have intended consequences, you know,

cane toads in Australia, you know, I used to leave an example of, and then you have to deal with that. The other positive, by the way, that you didn't mention that your adult one with you and your daughter was right, but, but there is some people are getting positive affirmation that they couldn't get otherwise, that people who are otherwise isolated are getting positive at the opportunity. It is true, although, you know, it is, it is interesting to ponder how such a symbol,

such a simple symbol of human affirmation can end up having, you know, any negative consequences at all. Essentially, how do you go from liking to hating? Yeah. I guess they're sort of intimately connected because the failure to like is the exclusion of somebody. Exactly. You're exercising your understandable homophily, trying to affiliate with people like you, you know, essentially implicitly excludes somebody. And in your in-groups when you're discussing things and exchanging norms, I mean, it's, you know,

implicitly you're excluding the outer group and you're permitting a process of divergence. And so it's quite, I guess this is like a universal in human history. It's, you know, positive intentions don't necessarily only have positive consequences. Oh, absolutely. The ripples on the pond. In fact, okay, well, let's end with the future. And I must admit, when someone calls himself a futurist, I stop listening. Because science fiction generally,

and futures always miss the most important things, including, by the way, the Internet, generally, if you look at science fiction, get flying cars in the 1950, no one was talking about the Internet. And so you do try, of course, fortune favors for mine, and it's an interesting thing to think about where likes may go,

and I don't want to spend that much time talking about, because it's the future, and it'll probably all be wrong, probably. But you make some interesting points, which I at least want to give a chance to elaborate on briefly, and then I want to go to summarize. But one, so which direction were the like button go? And you point out something interesting to me, that regulations, what the light, first all you point out, the like button is probably going to morph into something else.

And it's probably already happened, but we don't even know. It may just be something you don't have to press. You may, it may be, you know, raising your hand on your Zoom call or, or something else, or emojis. But you point out something interesting that I hadn't thought about that. regulators, especially in the UK, may help keep the like button going, because if you repost something, you're likely to get arrested in the UK if it's not politically correct, but you don't get arrested for liking it. So it may encourage people from moving away to likes. And I thought that was rather interesting. Do you want to comment on, well, that and emojis and other things?

And then I want to briefly. It's all speculation, but, you know, there are these. far-sighted individuals that we talked about, the Gary Steingard, the author on with his accurate vision of the internet and Max Lefchin and his vision that a two-way internet where data becomes the key commodity is possible. And plus the consequences are in terms of benefits and disbenefits are potentially enormous. So we should speculate without being over-invested in our speculations. So it seems to me relatively certain that the like button is not dead. I mean, it continues to be a valuable signal. So we'll continue to see it. Because it's now a lingua franca, right? Even if it weren't, even if we sort of replay history in it, it happens not to have been

the optimal symbol. It's the one, it's the one we have. So it's like we're locked into QWERTY, whether we like it or not. So I think that's true. I think that I think that I'm pretty, sure. I would bet a decent sum of money that we're going to see the light button play out on AI as a social technology in order to recover recoup these enormous investments in infrastructure that in data centers that the pioneers have made. I think I think it needs to be a very penetrated social technology basically and there are already signs that it is and these different products have have like buttons and so what is the feedback loop for AI you know you obviously got questions and answers themselves and we've got this liking layer and will there be other layers

but there'll certainly be some feedback loop for AI and it probably it could be like fundamentally qualitatively different because so my my left chin like speculation would be that the function of the like button for AI could be could be quite different. It could be linked to reinforcement learning. So instead of sort of segmentation of preferences for the purposes of monetizing advertising, you know, essentially the AI world is struggling to harness new

data sets to keep scaling the models. And you know, what data do we have? We have the sentiments associated with all of the questions and answers of AI. And so that that may be a sort of a new data set and a reinforcement learning mechanism that is unexploited. And then we have some more speculative, well, then we have something else, I think, which is reasonably certain, which is, I think, the regulatory game in messy, effective and ineffective ways will continue to play out.

And I think that will lead to, you know, secondary consequences, positive and negative. And it would be some time before we see that play out. So there are some places like Texas and Australia where we've, you know, prevented young people from having access to social media and most of the places that's not the case. I mean, we'll get a natural experiment there. How about that play out? We'll have to wait and see. But there'll be certainly, I think, some regulatory convergence, probably sort of, you know, too late to meet the need.

But I think there'll be regulatory dynamics that will shape this business. then I think we have some more speculative technological possibilities. So already we have certain technologies. They're not scaled or used for this purpose yet. But the neuralink technology is a functional technology. We can use Elon Musk's neuralink to have people control their wheelchairs and so on. This technology exists and is functional. So can we imagine a world where we don't need to tell somebody we like something

because we know, because there's a direct neural implant or less traumatically involuntary facial muscle movements. So we now have, there already exists software that is able to look at high-definition video feeds and tell you that you're projecting 13% contempt at this current moment. We don't need to ask your opinion about whether you're contemptful or not. We can actually tell because you cannot, you can't avoid giving off these signals in your involuntary facial muscle movements.

Will that become a part of the future? Possibly. It looks like the multiverse and the virtual worlds is moving a little more slowly than some people anticipated, but we could come back to that. But I think it's, you know, so the final chapter is very speculative. These are some of the directions and possibilities that we see. Now, interestingly, the final thing I'll say is, you know, I imagine Steingard would laugh at these because they're all sort of technocratic perspectives

or economic perspectives, whereas, you know, his genius was to ask, what will the humans do with this? And therefore, you know, I think we have to bring in the sociology and the culture and the humanities to interpret what humanity may, may, may, make of these things. So I'm pretty sure of one thing, which is I'll be doing a second edition of like with a new chapter in a couple years' time, and there'll be something we didn't anticipate, and that's what I'll be writing about. Well, it's the way I don't anticipate that that's why I think, that's why I say science, that's why I like science more than science fiction, because science

fiction creates a new future based on the present, but science, you know, creates a new future there. But in any case, one of the things you didn't mention, which I want to leave and with, I guess, before I give you the last word in a different way, is the other, the other eerie possibility that AI will allow you, allow us to anticipate lights even before they're given. And that is fascinating. But of course, as you point out, brings up Minority Report and all of the science fiction of the Phil K. Dick, the idea that AI will allow, You won't even know you're about to like it. It'll anticipate you're going to like it. And then X will happen? And will we be able to anticipate that people will like hate mail or hate this? And then you worry about all of that. And so, you know, we can talk forever about it.

But it is a chilling and at the same time interesting possibility. And like the universe, I was just talking to a colleague in a different podcast. If it can happen, it often will happen. and we should be, and fortune favors the prepared mind. So I'm going to, this has been fascinating. I'm going to give you the last word because I'm going to read the last two paragraphs of your book, or the regular party book. The like button turned out to have a great deal to teach us about human nature

and the evolutionary psychology behind the hypersciality that distinguishes our species. It held lessons too about the nature of technology innovation and how different the messy history of invention often is from a neat, purposeful and linear way its story is told. Studying the impact of the like button gave us a deep appreciation of the power of feedback data and the many ways it can be used, some of them very profitable. And we come to the end of our tour through the like button, knowing that most of its story still lies ahead. We know that people crave compliments and that not enough of them are dispensed in the world. For some reason, we just don't give out likes and abundance without prompts. One explanation comes from research suggesting that people

underestimate the positive impact of bestowing a compliment and are therefore more likely to suppress a compliment that they fear might not be perfectly expressed or received in the spirit intended. Whatever the reason, if a like button reduces the friction, that means a more compliment-filled world. And that's a nice optimistic view of the future, which may or may not be right. But what those last two paragraphs demonstrate to me is when I first got, I thought, why a book the like button and i hope our conversations indicated all of the remarkable things that you can learn about science about innovation about humans and about society and it's been fascinating and i really i for me it was a it was a unexpected pleasure and i want to thank you for for being part

of this well thank you for a really interesting and probing conversation lawrence and uh i guess that you end up uh with uh with that positive comment coming from you that's um that's that's a great compliment indeed for me. So thank you very much for this opportunity. It's great. Thanks. Hi, it's Lawrence again. As the Origins podcast continues to reach millions of people around the world, I just wanted to say thank you. It's because of your support, whether you listen or watch,

that we're able to help enrich the perspective of listeners by providing access to the people and ideas that are changing our understanding of our understanding of our ourselves and our world and driving the future of our society in the 21st century. If you enjoyed today's conversation, please consider leaving a review on Apple Podcasts or Spotify. You can also leave us private feedback on our website if you'd like to see any parts of the podcast improved. Finally, if you'd like to access ad-free and bonus content, become a paid subscriber at Originsproject.org. This podcast is produced by the Origins Project Foundation,

as a non-profit effort committed to enhancing public literacy and engagement with the world by connecting science and culture. You can learn more about our events, our travel excursions, and ways to get involved at Originsproject.org. Thank you.