TRIGGERnometry - Jaron Lanier - The Truth About Social Media

00:00:00.000 it puts everybody into this behavior modification regime that even if it doesn't do what it's

00:00:05.340 supposed to, it has all these side effects. It makes everybody a bit more vain and crazy and

00:00:10.440 stupid and contentious and paranoid. That does happen. It's really good at that. And because of

00:00:17.400 that, it becomes this giant invitation to the worst actors in the world to try to go in

00:00:22.880 and sabotage other people's societies.

00:00:30.000 Hello and welcome to Trigonometry. I'm Francis Foster. I'm Constantine Kisham. And this is

00:00:56.120 a show for you if you want honest conversations with fascinating people.

00:01:01.420 Our brilliant guest today is somebody we've been trying to get onto the show literally

00:01:05.140 since the day we started. He's an American computer scientist, visual artist, computer philosophy

00:01:10.220 writer, technologist, futurist, and composer of contemporary classical music as well. He's

00:01:15.560 considered one of the founders of the field of virtual reality. Jaron Lanier, welcome to

00:01:19.460 Trigonometry.

00:01:20.640 Hey, how's it going?

00:01:21.800 It's going great. Listen, man, before we get into what we hope is going to be a brilliant

00:01:26.680 and fascinating conversation, tell everybody a little bit about your story. Who are you?

00:01:31.800 How are you where you are? What has been your journey through life that leads you to be sitting

00:01:35.660 here talking to us?

00:01:37.140 Oh, gosh, you know, I find that to be a confusing question.

00:01:40.120 Well, let's see. I'm known for a few different things and to different communities of people.

00:01:50.040 Some people only know me one way or the other. I approximately started virtual reality. I came

00:01:57.900 up with the term and I had the first startup and pioneered a lot of the apps and devices and

00:02:02.280 whatnot a long time ago in the 80s when I was a kid in my wasted youth. I write books that are

00:02:11.160 critical about technology. I'm very concerned that we have a responsibility as computer scientists to

00:02:16.340 think about what we're doing to the world. And when it's not working, we have to be honest about

00:02:21.440 it. And so I've written a bunch of books. The best known one might be the 10 arguments for deleting

00:02:28.440 your social media accounts now, right now. But another one is called You're Not a Gadget.

00:02:33.940 And I'm very skeptical of the idea of artificial intelligence. I'm very skeptical of the business

00:02:38.340 models for social media and other things that involve. I'm very concerned about them. And then

00:02:44.120 yet another thing is, as an actual computer scientist, I do all kinds of things. We don't

00:02:48.920 have to get into that because it'll put you to sleep, but it's actually extremely important and

00:02:52.300 profound. And then I do other stuff too, but I think that's enough to get going.

00:02:58.440 Sure. Well, you've covered three things there that Francis and I are really keen to talk to you

00:03:04.380 about, which is AI, virtual reality, and of course, social media. I think we best start with social

00:03:09.760 media because I feel like of all the different stories we've covered on the show, the different

00:03:14.900 guests we've talked to, the polarization, the tone of our public discourse, the behavior of children in

00:03:23.320 terms of their self-image and their body image. I mean, any issue that is affecting modern society,

00:03:29.780 really, there is a role that social media has played. And by the way, you know, Francis and I

00:03:35.280 are both quote unquote content creators. I love using Twitter, for example, to talk to, you know,

00:03:40.480 hundreds of thousands of people on a daily basis. I think it's an incredible piece of technology,

00:03:45.020 but we're also paying a heavy, heavy price for some of these benefits. Talk to us about your broad

00:03:51.840 take on social media. You talk about the business model and the problems with that. What are some

00:03:56.500 of the impacts that we're seeing in the world, in your opinion, and where they're coming from?

00:04:01.320 Well, okay. The first thing I want to say is that in the biggest picture, something similar to social

00:04:07.920 media, as we know it, would be great. Like I worked really hard to help get the internet to happen,

00:04:14.780 which is a story we can talk about if you want, what my role was. But I still believe in the idea

00:04:19.760 of a more connected world. I still believe in the idea of people being able to have reach in new ways.

00:04:27.580 All of that's great. The problem that came up is the business model, which had a perverse effect

00:04:33.420 of promoting the worst sides of human nature. Okay. So it's a business model problem more than a

00:04:40.600 technology problem. And I really want people to remember that because a lot of times somebody will

00:04:45.040 come up to me and say, well, there's this or that about whatever social media platforms they like

00:04:50.160 that I love. And how can you say that I should quit it and that it's all horrible? And I'll say,

00:04:54.100 you know, I never told you to quit. I gave you arguments. I want you to think. I want you to be

00:04:58.900 responsible for your choices. But the good stuff is authentic. Like it would be absurd to argue that

00:05:06.520 the good stuff is inauthentic because that's ridiculous. I mean, it's plain in front of our faces

00:05:11.780 that they're good things. My personal favorite one remains an oldie but goodie, which is people

00:05:17.900 with unusual diseases can find each other and share notes. And that used to be possible. That to me

00:05:23.520 is a great concrete example of something that the technology has brought into the world that

00:05:29.220 otherwise would not be in the world that has kept people alive. I love that. All right. So I mean,

00:05:34.600 it should be obvious. However, what happened was, for a lot of really stupid reasons,

00:05:42.660 the early internet companies ended up with this bizarre business model. And the business model

00:05:47.860 is you pretend to be socialist in the user's experience, but the way you're capitalist is

00:05:55.240 fake capitalist too. You end up with the worst of both worlds instead of the best of both worlds.

00:05:59.140 So you're fake socialist in the sense that you say, oh, sharing is great to share your information

00:06:04.800 and we'll give you free stuff. But the problem is there has to be a customer somewhere because it's

00:06:10.640 still a world driven by business. And so the customers who come in think that what they're

00:06:15.740 buying is mind control. They think that what they're buying is so much access to user data

00:06:20.520 and so much access to giant algorithms that can predict user behavior that they can just pour money

00:06:26.480 into Alphabet or Meta, Google or Facebook, and out will come massive behavior modification that will

00:06:34.040 throw elections or get people to buy products or whatever it is, or commit to a religion, all kinds

00:06:41.160 of crazy stuff, or just cynically bring about a quicker end to civilization or something, whatever the

00:06:49.760 person's into. The thing is, a lot of that stuff doesn't even work. I've been around the meetings

00:06:57.020 where the big companies sell the biggest clients to pour money into this. And honestly, I don't think

00:07:04.980 you get a lot of mind control. I do think you get two things though. I think you get just a few big

00:07:12.800 companies holding the ability to blackmail everybody by controlling access of everybody

00:07:19.480 to everybody. So in other words, it's just the opposite. What the internet was supposed to be is

00:07:23.400 connecting everyone. And instead we have universal blackmail of forcing people to behave in certain

00:07:29.420 ways to get connected, which is the opposite of what was supposed to happen. So that's stupid.

00:07:34.780 But then the other thing is that even if the mind control to get somebody to buy a particular soap

00:07:41.060 is pretty sketchy, and I think a lot of times the data to support it is sketchy, what it does do is

00:07:46.840 it puts everybody into this behavior modification regime that even if it doesn't do what it's supposed

00:07:52.480 to, it has all these side effects. It makes everybody a bit more vain and crazy and stupid and contentious

00:07:59.380 and paranoid. That does happen. It's really good at that. And because of that, it becomes this giant

00:08:06.540 invitation to the worst actors in the world to try to go in and sabotage other people's societies.

00:08:12.320 So you have the lovely Mr. Putin having these basements in St. Petersburg create fake black

00:08:22.980 activists to try to make the black political movements in the U.S. more radical in order to

00:08:27.300 destabilize American society and make it kind of nutty. That's documented. I'm just using that example

00:08:32.680 because it's well-documented. There's many, many others that are as well. And so you overall have

00:08:40.940 this thing that darkens human prospects. It makes everybody, like I say, just a little like a lesser

00:08:46.340 version of themselves, a little more vain, a little more mature, a little more paranoid, a little more

00:08:50.760 ornery and irritable. And all of that is in service of a business model that doesn't really deliver a lot

00:08:58.540 of real value, but just creates artificial impediments to what would otherwise happen,

00:09:03.000 in my opinion. So I think it's a massive, massive screw up. And it happened.

00:09:11.680 You know, I was around Google when it was very, very small before Facebook existed. I sold them a

00:09:18.220 company. I used to know Sergey and Larry. They were so cute, so full of energy, so optimistic.

00:09:23.400 And the thing is that the surrounding social circumstance was incredibly adamant about two

00:09:33.520 different needs that were totally in contradiction with one another. On the one hand,

00:09:38.580 there was this leftist feeling because it's the Bay Area in California and everybody's supposed to

00:09:43.640 share. The internet is finally a chance to get rid of the evils of capitalism. But on the other hand,

00:09:49.460 we love our business hacker heroes. We love our entrepreneurs. We love our Steve Jobs, you know.

00:09:56.780 And so you're supposed to be this great business person, but you're supposed to hate business.

00:10:01.920 You know, like, how do you get those things to work? And the way to do it is by being fake,

00:10:06.200 where you create fake socialism as an experience and fake capitalism on the back end. You don't get,

00:10:12.720 like, whatever you believe about either of them, you know, this is the worst of both of them.

00:10:19.080 It's like the worst possible solution, but in a way, it's the only one the social pressures of the

00:10:24.880 time would allow to happen.

00:10:26.400 Jaron, can you dig in for us into that particular thing a little bit more?

00:10:30.500 What is the, I understand that me as a Twitter user, I'm not the customer, it's the advertiser

00:10:36.220 that's the customer. But beyond that, how is the business model making us different?

00:10:42.340 And when you talk about fake socialism and fake capitalism, dig into that a little bit more for us.

00:10:48.920 Yeah, sure. So there's a, we have to go a little bit into the science behind behavior modification.

00:10:56.420 So this backwards, it goes into the 19th century with Ivan Pavlov, yet another Russian. I see you

00:11:02.100 have a mock Russian poster behind you.

00:11:03.840 Well, I'm from Russia originally, so that's my heritage.

00:11:06.000 Okay. And I'm from Venezuela, so I've seen socialism done properly.

00:11:16.000 Okay. Excellent. Excellent. So we're all extremely well informed here.

00:11:24.020 Yeah. Pavlov, talk to us about Pavlov and his beautiful dogs.

00:11:29.100 One of the early behaviorists. And so these are people who said, we're going to do a methodical,

00:11:34.840 mathematical, scientific approach to understanding how training works. So, of course, people have

00:11:40.900 been training animals, and for that matter, other people since ancient times. There's nothing new

00:11:45.640 about that. But this was saying, okay, we're in the scientific age, we're going to do it methodically.

00:11:50.260 And so Pavlov started to experiment with like, you're going to put a creature in a cage and control

00:11:58.660 exactly what stimuli are available and exactly what outcomes are available. So you can really get

00:12:03.960 focused information about what works and what doesn't work to understand behaviorism. In the 20th

00:12:10.240 century, probably the most famous heir to Pavlov was B.F. Skinner. He was doing the same thing. In

00:12:18.160 Skinner's case, mostly with pigeons and rats, you put them in a cage. There are a few very

00:12:24.820 interesting things that came out of this methodical approach to training. One thing is that it can

00:12:33.920 get out of control. So if you put an animal in a cage and you say, if you hit this button, you'll get

00:12:41.260 a treat. If you hit this other button, you'll get an electric shock, let's say. They'll just sit there

00:12:47.500 hitting the treat button. What happens is it changes the nature of an organism or a creature from one

00:12:54.680 that has multi-level behaviors that can respond to different circumstances and can learn to one that's

00:13:00.040 very simple, just stuck on a little loop. Okay, that's a very important thing to observe. All right,

00:13:06.120 that's one thing. Another thing that's extremely interesting is that if you want to deeply embed a

00:13:16.540 behavioral pattern in the creature, instead of just having perfect feedback, where when you hit the

00:13:23.860 button, you immediately get the candy every time, you add a bit of randomness, or it might take a while

00:13:29.560 longer. And this bit of randomness, what it does is it absorbs the brain's ability to adapt and

00:13:36.940 focus it entirely on this, because the brain is like, well, what's going on? What's going on? And so it

00:13:41.640 actually strengthens the conditioning. Okay, so this is sometimes called operant conditioning. There

00:13:47.680 are many variations, and I could flip into technical talk to create many shades of distinction, but

00:13:53.640 I think I'm giving you a fair summary of how this stuff works.

00:13:59.600 The people who studied this had different opinions about how it should be applied to human beings.

00:14:06.820 Okay, now, at the dawn of computers, one of the founding computer scientists

00:14:13.640 wrote a book called The Human Use of Human Beings. And this was a book pointing out that

00:14:24.560 with the advent of computers, you could automate this whole thing instead of having white-coated lab

00:14:30.380 researchers doing it. And there's a danger that people could make machines that would just make each

00:14:36.240 other trained in a way that they'd become crazy and very narrow, and it could really be extremely

00:14:41.100 dangerous for civilization. That was a book in the 50s. Sadly forgotten, Human Use of Human Beings.

00:14:51.660 But then we hit the 60s, and something really interesting happens. It's famously known that a

00:14:58.300 little into the late 60s, the beginning of the technology of the internet happened,

00:15:03.060 funded in the United States by military research. I was trying to come up with communications

00:15:09.060 networks that could survive a nuclear attack by being very adaptable, right? But before that,

00:15:15.560 there was a cruder technology for networking that was the first time computers were actually

00:15:19.960 networked at distances for people to use. And this was an experimental network that ran only between

00:15:28.200 universities in the American Midwest. And guess who got the job of designing the user experience in

00:15:34.760 it? None other than RBF Skinner. So the very first network experiment with real people was by a

00:15:42.320 behaviorist. And his attitude, I think, was quite wrong, really wrongheaded. What he thought

00:15:48.040 is that if you could get everybody onto computers that are networked, and you could put everybody under the

00:15:55.620 influence of these algorithms, you could engineer a society that would become stable and productive and

00:16:03.580 happy and perfect. That's the right reaction. It's funny, right? It strikes us as a laugh line. It

00:16:15.260 seems ridiculous. But there are a couple of amazing things about it, which is that BF Skinner was on

00:16:21.380 the Western capitalist side of the Cold War. And yet what he was saying was not unlike what was

00:16:27.680 happening on the other side. This is exactly what people in places like East Germany and

00:16:33.340 North Korea and many other places have tried to do using other means to sort of put people under enough

00:16:39.600 behavioral control and sensory control that you can engineer this perfect, coherent society to your liking.

00:16:45.880 And jumping ahead a little bit, Jaren, it's exactly like what China is doing now to its own citizens, as we know,

00:16:50.660 right? Right. And unfortunately, I think China learned some tricks from our tech companies in

00:16:55.340 this case, something that I find shameful. Yeah. But anyway, sorry for interrupting you. Carry on.

00:17:01.300 No, it's right. It's right. So look, anyway, so Skinner did this experiment. It didn't work particularly

00:17:05.900 well. It was a tiny crude experiment, mostly forgotten, actually. So then the internet happens

00:17:17.400 initially between military sites and universities. But this is the radical 60s. So there's this kind

00:17:22.580 of hippie thing, which, hey, listen, I'm part of it. I'm a hippie. I'm a California hippie, proud of it.

00:17:29.920 But at any rate, there was this hippie sensibility that was anti-capitalism, not exactly pro-Soviet or

00:17:37.080 pro-the other side, but definitely in this other space and somewhat naively utopian on many levels.

00:17:45.380 I say that as somebody who's tried to live in a 60s commune and whatnot.

00:17:50.080 It's hard. People are hard. You can't just pretend people are easy when people are hard. I mean,

00:17:54.160 this is a fundamental issue that the left never gets, right? But anyway, so then when companies

00:18:06.200 like Google show up, the rhetoric around computer culture, which is, and computer culture is

00:18:12.100 incredibly powerful. I mean, it's not, this has never been an entirely top-down thing. This is a

00:18:17.600 community cultural effort. And computer culture is anti-capitalist on the surface, but as I say,

00:18:24.040 also pro-capitalist, as long as you're a hacker hero entrepreneur, you know. So it doesn't make sense.

00:18:29.920 And it doesn't, we were absolutely unwilling to take three seconds to step back and notice our own

00:18:35.540 contradictions, right? It's ridiculous, you know, very human, I think. So Google was born with this

00:18:45.100 ridiculous hippie illusion in front, even though it's actually a company. And I have to say,

00:18:51.540 I'm putting it this way because I'm pretty sure that the original people who made Google

00:19:03.640 would have been just as happy to do it differently, but the community pressure was so intense that they

00:19:08.540 weren't really given a practical choice. It was like this very intense dogma. Like as an example,

00:19:16.300 I'll say some things here that might upset some people who are listening, but like in the

00:19:21.380 early days, 20 years ago, or whatever, but more than that, actually, like the seeming good guys

00:19:29.860 who had nothing but very high self-regard, like the pirate parties in Europe, were feeding Google

00:19:36.340 business while pretending to be anarchists and anti-capitalists. And that was plainly true,

00:19:42.620 obviously true to everybody. They were even funded by Google in many cases. And yet somehow everybody

00:19:48.480 just ignored this. And I think it's an incredible testament to how people will be happily ready to

00:19:55.740 lie to themselves if it's comfortable. And I think that's universal. That's not just a problem

00:20:01.380 with this particular community. It's a universal human quality. And it sneaks up on us all the time.

00:20:06.800 Hey, Konstantin, do you want better mental health? I'm from Russia. We don't have mental health.

00:20:14.740 So how do you deal with mental health? You drink vodka, then go out and wrestle bear. If you live,

00:20:20.660 you feel better. If you die, you're not real man. What about the bear's feelings? It's Russian bear.

00:20:26.780 It has no feelings. People don't always realize that physical symptoms like headaches, teeth grinding,

00:20:32.560 and even digestive issues can be indicators of stress. And let's not forget about doom scrolling,

00:20:38.820 not sleeping enough, sleeping too much, undereating and overeating. Sleeping too much,

00:20:45.000 undereating. This is Western disease. Therapy has really helped me in my life to concentrate and focus.

00:20:52.000 It's really important to have someone impartial who you can talk to about the tricky issues that

00:20:57.680 you're struggling to deal with. Therapy has played a really important role in helping me to deal with

00:21:03.320 my ADHD and become better in all areas of my life. Why is he telling them how weak he is? Drink vodka,

00:21:11.220 feel better. BetterHelp is customized online therapy that offers video, phone, and even live chat

00:21:18.380 sessions with your therapist. So you don't have to see anyone on camera if you don't want to.

00:21:24.100 Trigonometry funds get 10% off their first month at betterhelp.com forward slash trigger,

00:21:29.740 especially if they're not real men. That's B-E-T-T-E-R-H-E-L-P dot com slash trigger.

00:21:40.740 We're talking about the worst of capitalism, but these companies in many ways are the worst of

00:21:46.080 capitalism. They're monopolies. They're incredibly powerful. They now have the ability to influence

00:21:52.520 elections like what we saw with Hunter Biden and Twitter suppressing the story. I mean, it's

00:21:58.480 terrifying, isn't it? Well, a couple of things to say there. One thing is I get calls from all kinds

00:22:07.300 of people who are upset about what happened to elections. I have more firsthand knowledge than

00:22:15.580 most people about what happened in the platforms. And there was more suppression of the left than the

00:22:21.280 right. Both things happened. But just to be clear about that, there was more stuff that happened

00:22:27.540 that was anti-Biden than pro-Biden. Like what? Tell us. Because I've never heard this argument. I'm

00:22:32.860 very open to it. Like what? Can you give us some examples? The documented for real and high volume

00:22:41.120 foreign interventions that favored a candidate, favored Trump, mostly driven by Russia.

00:22:47.080 To say that that's all that happened would be ridiculous because all kinds of things were

00:22:52.540 happening. But the weight of it went in that direction. Does that mean anything? No, actually,

00:22:57.100 it doesn't. I know that it's frequently, you kind of want to believe, like if you're more sympathetic

00:23:03.220 to one side than the other side, you kind of want to believe the other side's got more advantage or

00:23:09.180 whatever. Honestly, this whole thing is a big, giant bundle of bullshit anyway. Like I think to worry

00:23:18.220 about whether the other side got more bullshit than your side is really irrelevant. The point is there

00:23:25.040 should be no bullshit. No argument there at all. But John, can I just, can I just challenge you briefly

00:23:31.040 on this point? I'm not challenge. You know, look, I know I've been through this argument a lot. I

00:23:35.560 really want to emphasize, I do believe I have a clear picture of it. On the other hand, I just

00:23:41.340 really don't think it matters or it's that important. If it strikes you as being important, I really want

00:23:45.740 you, I want you to reconsider whether it is. Well, that's why I want to ask you the question,

00:23:50.160 because I am totally open for you to persuade me it's unimportant. The difference that I saw,

00:23:55.460 and look, neither Francis or I are Trump fans. Neither of us is on the right. None of that,

00:23:59.920 right? We're just, as a principle, to me, watching a set of coordinated efforts by different tech

00:24:08.200 platforms to prevent a story about the son of a presidential candidate being published on the

00:24:14.280 eve of the election, that was unprecedented and very, very different to the tech platform messing

00:24:23.080 around with Russian bot farms or whatever. Do you see what I'm saying? Yeah. Oh, look, I, and you know

00:24:28.860 what, um, I, if you're going to, so if, if the question before us was, were there parties who

00:24:39.980 worked together to try to support Biden in ways that make us uncomfortable about the future of a

00:24:45.760 decent, honest society and democracy? The answer is yes, there were. And I, and so I can, I would

00:24:51.540 validate what you're saying. Um, I think, I think that happened in a few different levels. I think

00:24:57.020 part of it was, um, the fact that most people of influence in the tech world are educated and most

00:25:05.120 educated people are more sympathetic to the Democrats in the U S and to liberal positions in general. Um,

00:25:14.000 and they naturally tend to support each other when they start to move in that direction. And even

00:25:19.360 without an explicit conspiracy, there can be an effective one. All right. I, I, I don't want to

00:25:25.380 deny that there's truth about that. Um, I also want to point out that some of what they were resisting,

00:25:31.780 like if the question is, uh, well, they're different questions. You could say, was there less

00:25:37.200 reporting than there might've otherwise been about Hunter Biden? Probably on the other hand, a lot of the

00:25:42.460 promotion of the Hunter Biden story was also equally questionable and for the same reasons.

00:25:46.760 And with even more central or central coordination and more questionable issues and more money behind

00:25:53.400 it to the best of my ability to read it, which is neither here nor there. Um, in my view, in the

00:25:59.720 balance, if you compare all the different kinds of bias that were injected into the social media

00:26:04.000 systems, there was more pro Trump stuff going on than pro Biden stuff. Once again, neither here nor

00:26:12.140 there. And it, it, it makes no sense for any of us to adjudicate that if I'm wrong, it really wouldn't

00:26:17.900 have any effect on any of the positions I'm taking of consequence here.

00:26:21.500 Truly. Okay. Jared, that, that being the case, what do we do with these hugely powerful companies?

00:26:28.620 They, uh, you know, they can pretty much do what they want. Yeah. They can control what they want.

00:26:34.220 They, they disseminate information as they want. They're too powerful, surely.

00:26:39.900 Yeah. So this is very problematic. So, um, going back to the book, I mentioned the human use of

00:26:45.340 human beings, and I didn't mention the author Norbert Wiener, I should mention. And I have to say,

00:26:49.900 this book is really hard to read because this guy's an ultra nerd and there's mathematical

00:26:53.660 equations in the middle of it. And, um, it's a great shame if he'd written it in an accessible style,

00:26:59.420 I think it would have changed the world, but he, he just wasn't that person, you know, at the time,

00:27:03.420 this is going way, way back. This is to the Manhattan project generation of scientists, you

00:27:08.140 know? So anyway, um, the problem is that if you have computer networks running the world, whoever

00:27:15.900 runs the networks runs the world, it's like a much more powerful thing than money. It's much more

00:27:20.060 powerful thing than votes. It's much, it's like the most powerful thing because you're directly

00:27:24.540 in control of the channels of action that exist in the civilization. And this was a danger that

00:27:31.020 was well articulated in advance by a variety of people. Um, I tried to say what I could about it

00:27:36.300 at the time, as did a few others. Um, honestly, most of us in computer science were just so enthralled

00:27:42.540 by the prospect of power that we allowed our, our kind of, um, I don't know, our egos to get the best

00:27:52.300 of us. Oh, we must, this must be for the good. We're the best people. We'll save the world if we

00:27:55.900 get powerful. I mean, it's, and, and of course that's fallacious. No, no, nobody's perfect, you

00:28:00.540 know, and no concentration of power is trouble free, you know? And so, um, that, that's, uh,

00:28:08.060 what happened. And so now we do have this kind of curious situation. Now that said, there's a weird

00:28:16.060 kind of neutered quality to the new kind of power that makes it a little less horrible than it might

00:28:23.100 be. And if you want to look at the comparison, compare what companies like Meta and Alphabet

00:28:27.020 can do to what the Chinese are doing, or if it doesn't matter, the Russians or the Iranians,

00:28:31.900 or these all kinds of other people. Like there's a weird thing about Silicon Valley culture,

00:28:36.380 if I can call that, or tech or tech culture, and that it's so nerdy. It's so, we're not supposed

00:28:42.620 to use the term on the spectrum anymore, but I don't really know what language to use for this.

00:28:46.380 There's this kind of dryness to it that in a way, it's not exactly traditional power seeking,

00:28:51.980 and it doesn't have traditional power goals. It's more just this kind of

00:28:57.740 nerd supremacy without a particular direction or sensibility. It's, it's, it's truly different,

00:29:05.100 I think, than previous concentrations of power. I mean, like you go to Silicon Valley and every other

00:29:11.420 center of power in history has made itself beautiful and impressive, and Silicon Valley has

00:29:15.660 not. It's just kind of another shitty suburb, you know, and that's strange, but it reflects this kind

00:29:21.100 of incredible nerdiness. And so if you talk to Zuckerberg, it's just this kind of very formulaic

00:29:33.580 first order approximation of what an idea would be. Like the world should be more connected, so we'll make

00:29:37.740 it more connected, and that's to the good. But there's nothing more. There's no, there's nothing

00:29:41.900 there. I mean, of course, the individuals like getting rich, but they don't necessarily have

00:29:46.460 anything to do with that, you know. Same thing at Google. It's, it's kind of an odd thing. It's like

00:29:52.380 we will organize the world's information, which I think, by the way, is a misnomer. Let's, but that we

00:29:57.340 don't need to go into the philosophy of that. But the thing is, it's a weirdly neutered thing. And so

00:30:02.540 the main problem with it is different from the problem of, oh, I don't know, a Putin-like figure

00:30:11.340 who wants to conquer a neighbor, or a Xi-like figure in China who wants to create this sort of

00:30:17.900 Chinese-centric model of the world that gives no room for even the identity of Tibetans or Uyghurs

00:30:24.460 or Taiwanese or anything. It's not like either of those things. It's more, it's a weird thing. It's

00:30:32.700 like the nerd kid in high school who doesn't even really have an agenda, but just is sick of being

00:30:41.180 the one dumped on, you know. It's a strange, inert kind of power to a degree. But because it has the

00:30:47.580 side effect of destroying the personnel, harming the personalities of everybody else, it still is,

00:30:54.220 the effects are still too negative. And of course, the other thing to say is, whenever you create a

00:30:58.940 center of power, it'll eventually be seized by the worst people. You start off with Bolsheviks,

00:31:03.820 then you get Stalinists. That's how it works. And when you get rid of the Stalinists, you don't

00:31:07.740 suddenly get Democrats. You get Putin's, you know. Like, centers of power are seductively horrible

00:31:19.020 for history. So there's a lot of reasons not to do it. But it's also important to understand that

00:31:25.020 when we talk about power of tech companies, it's a different kind of power than we're used to

00:31:28.940 historically. Jaren, it seems to me that, like all revolutions, it started off with utopians,

00:31:35.900 people wanting to build a better future, people having an idealized view of the world and how

00:31:41.260 they were going to change things to the better, and they were going to make everyone more connected,

00:31:44.700 and society was going to get richer and just better as a result, without realizing,

00:31:51.100 A, the consequences of their actions, but B, as well, you've got to monetize this thing. And if you're

00:31:57.740 going to monetize it, you want people to stay on the platforms. And how are you going to do that?

00:32:02.860 You're going to do it through algorithms and playing with people's emotions.

00:32:06.300 Yeah, just over, over the last weekend, I keynoted, or I co-keenoted with Neil Stevenson,

00:32:14.780 who wrote Snow Crash and many other science fiction novels, the big crypto conference,

00:32:20.140 which is called Consensus, put on by Coindesk in Austin, which is kind of like,

00:32:26.620 we call it a dumpster fire in our current lingo. But anyway, whatever. The thing is, I thought,

00:32:32.460 wow, I'm going to come in here. This is a hostile audience. You're not going to want to hear me say

00:32:36.780 how scammy this whole thing is, and how whatever. But I put it in a positive light. What I said is,

00:32:42.060 look, instead of trying to make money from finding the next person to come along, who buys some token

00:32:47.580 for more than you paid for it, why don't we get these things to form value chains? Where if you have

00:32:52.700 a token for some piece of art, and some person uses that art and animation, and then somebody else adds

00:32:57.420 music to the animation, and somebody else adds a story, and somebody else does this and that,

00:33:02.140 and then it eventually makes money because people are willing to pay for it, then if royalties go

00:33:05.660 back to everybody, then instead of selling the token, you could earn royalties and dividends on it

00:33:10.940 and fund a society. Because one of the fundamental problems with tech culture and money in general is

00:33:17.580 that it has forgotten what money is for. If you go back to Adam Smith, or anybody who's been concerned

00:33:24.620 with how money works, the idea is that when you have money, that money goes to work for society,

00:33:30.540 that's how you earn interest on it, and the money goes up in value. The worst thing is to stuff your

00:33:34.860 money in a mattress and pull it out of circulation, right? And yet, the tech industry has sort of been

00:33:41.660 like putting money in a mattress, and that's even more so with the web three and crypto stuff.

00:33:46.700 And so what's funny is, when I said this, I thought they were going to be angry, but I had this whole

00:33:51.180 haul of people who started applauding, and then they stood up. And it's like, I think everybody knows

00:33:56.780 in their gut that the way we've been doing things is silly, and we need to be more focused on real

00:34:02.220 productivity, real investment, real accomplishment, real achievement, real building, real innovation,

00:34:07.820 and just moving bits around and screwing around with other people's behaviors doesn't actually do

00:34:12.140 anything for the future. And everybody really is in it for the future at the end of the day,

00:34:16.700 that's why people get into tech. So I, it might just evaporate, maybe I'm kidding myself,

00:34:22.620 but I have this feeling like there's this, and this is a very young audience. So these are new people,

00:34:27.900 and not the same old people I've been talking to for decades. And I just have this kind of optimistic

00:34:33.420 feeling that they're open to it, and that there's room for tech culture to improve. I really believe that.

00:34:39.980 Yeah, well, I know what you mean, John, because I think, as we, I said to you right at the beginning,

00:34:44.220 when we started, I think, while many of us are grateful for the tremendous opportunities that

00:34:49.900 being connected with 8 billion other people have given us, at the same time, we cannot ignore some

00:34:56.780 of the terrible downsides, whether they are for society, or for our children, or for our own well

00:35:02.140 being, or any of those things. So, and I think a lot of people are starting to wake up to that. And

00:35:07.180 the other thing that resonated with me so strongly is when you were talking about,

00:35:10.940 you know, these are nerds who've got their own sort of not quite as evil agenda as it might be.

00:35:17.100 But I was thinking the whole time you were talking, yeah, but that's because they own

00:35:20.380 the thing that they created. What happens when they all get forced out, or they die, or whatever?

00:35:24.780 Who's going to come in and replace them?

00:35:27.420 Yeah, yeah. I think about that all the time. Like, who's going to, who exactly is going to inherit

00:35:33.580 Facebook or Meta? Right. You know, it's not set up to allow anything. I mean, now, Cheryl just left,

00:35:40.220 and instead of a new Cheryl, it's just even more concentrated. And that is not the way to create

00:35:45.260 an institution for the benefit of humankind. You have to see beyond yourself, and it just hasn't

00:35:50.460 happened yet. So let me ask you about that, Jaren, because you talked about how this could all have

00:35:55.340 been done differently. And you spoke about it at the time, and the business model is wrong.

00:35:59.980 How do we fix social media? Or how do we make it better? How do we make it slightly less worse,

00:36:07.100 at least? Yeah, slightly less worse. I'll live with that. I think it's a really huge

00:36:15.180 problem to say, oh, I'm going to fix it. I'm going to make everything better, and everything's

00:36:19.020 going to be perfect. Because whenever somebody says that, you know, this is going to be really

00:36:23.820 crappy. If it's like, I can make this slightly less horrible, you know, that's the person you want

00:36:28.060 to listen to. So I think there's a few things that can be done differently. One of them, if anybody

00:36:34.300 wants to, the current issue of the Atlantic magazine in the US has a piece of mine on how to fix Twitter

00:36:41.100 that I wrote for Mr. Elon, and perhaps to no effect whatsoever, we shall see. But what it proposes

00:36:48.380 is an idea that I think is the least invasive and least impinging on free speech way of just improving

00:36:58.700 the conversation. And that is to demand that people form small associations, like form little zines,

00:37:06.300 or bands, or clubs, or brands, small enough that they can all know each other, and only publish through

00:37:11.900 those and rise or fall in terms of reputation and monetary income together in that band, like you

00:37:19.580 two are doing with your production here, right? Now, here's the reason why I'm saying that.

00:37:26.700 If you look at all the people who've studied societies that you'd want to live in, they've all

00:37:32.140 come to the conclusion that that mechanism, which is often called a societal institution, is at the core

00:37:37.500 core of maintaining quality, all right? Some of the people who talked about that are the Tocqueville,

00:37:44.060 Hannah Arendt, but you know the one that really got me going on it some time ago was a friend of

00:37:48.780 mine named Mohamed Yunus, who won a Nobel for starting microlending. And there's a lot of ways that

00:37:55.340 microlending has perhaps not achieved all that some hoped it would achieve, but let's leave that aside.

00:38:00.460 There's one part of it that's worked really well, which is he had a bank in Bangladesh

00:38:06.540 trying to serve a super poor community where nobody had a credit rating. If they gave somebody

00:38:11.180 a loan, chances are the person would not pay it back because like, why would they? It was just free

00:38:15.260 money. They weren't even used to the whole thing. And they said, okay, we're not going to do loans

00:38:19.740 to individuals anymore. You guys have to form a group. You have to vouch for each other. If one of

00:38:23.580 you doesn't repay, you all are going to pay the price for it. And this is a way that we can distribute

00:38:29.100 the process of credit, creating credit, or if you like creating quality. And it worked. They

00:38:35.180 suddenly got better repayment rates than traditional banks are used to. Incredible. So when people have

00:38:40.540 a bit of a shared stake, they start to watch out for each other. And if somebody runs into legitimate

00:38:44.860 trouble, other people help them. If somebody wants to screw around, they get rejected. They get ejected

00:38:49.660 from the group. And suddenly you have quality going up. Magic, right? And so I would like to

00:38:56.060 see that in social media. And it would be strictly by free association. If you don't like your group,

00:39:00.220 you can always quit. You can get into another one. A group can decide who to accept, who to reject.

00:39:04.380 But the point is that the groups are small enough that the people really know each other. And then

00:39:09.580 what you get out of that is a few things. One, if somebody starts turning into a jerk, and

00:39:15.100 everybody turns into a jerk online from time to time, because we turn into a jerk from time to

00:39:20.300 time offline. I mean, let's be honest about ourselves. The other people say, hey, you know

00:39:25.820 what? Cool it. This is like a bit much. But another great thing is the group can post often enough to

00:39:32.540 keep the brand going and to keep subscribers, whereas individuals have to post too often to do so and to

00:39:37.980 stay sane. So that's a good thing. Another really good thing is that since the group is divvying up

00:39:43.740 benefits, and I want them to get subscriptions, donations, micropayments, I'm really into money

00:39:49.820 for this. I think there should be more money online. I really do. Because we live in a market world.

00:39:54.380 And to say, well, the online world won't be a market world just makes people become gradually

00:39:58.620 more and more obsolete as technology gets better. It's ridiculous. If we're going to be socialists,

00:40:05.420 we have to do it everywhere. We can't just do it online. So I'm really pro-capitalism online,

00:40:10.380 because I think it's the most viable way to make things workable. But anyway, if there's a bot,

00:40:16.700 if there's a fake person in your group, and that bot is getting some of the money, you have suddenly

00:40:20.300 a motivation to get rid of the bot. And now when Elon Musk is saying, oh, there's too many bots on

00:40:26.300 Twitter, he's absolutely right. A friend of mine at Facebook tells me that 99% of the new account

00:40:32.540 applications are from bots by their estimate. And the reason why is that people who want to mess up a

00:40:38.380 society put in bots in order to sway opinion, right? So there's every economic incentive to

00:40:43.980 make bots. Well, now there'll be an economic incentive to get rid of bots. And otherwise,

00:40:49.260 it's the platform's job to do it from on top, which is a losing game. But if it's distributed,

00:40:54.620 just like creating credit and micro-lending, all of a sudden it becomes doable and we can get rid of

00:40:58.860 the bots, which are terrible. They're poisonous. There are many other benefits. You can read the

00:41:04.780 piece in the Atlantic if you want. So another thing I'll point out is that I've tried to do a

00:41:11.660 ranking of the relative degree of terribleness in different online platforms.

00:41:17.100 How often are people jerks? And nothing's perfect. But the thing is, it's human beings,

00:41:23.660 so nothing's ever going to be perfect, right? But there are differences, okay? So

00:41:28.940 some of the places are really a cesspool. Facebook is junk. Horrible. YouTube is really

00:41:40.780 pretty bad. You can manage your experience on it to make it better, but it's really pretty awful.

00:41:47.980 There's an experiment. Now, some people challenge me and say, no, this isn't true. But I've done this

00:41:53.100 many times. And in my experience, having done it dozens of times, which is not a huge sample,

00:41:57.420 so not up to scientific standards. But if I get a room full of kids and I say, okay,

00:42:03.100 just either start with your account or start with a fresh identity where Google doesn't know who you

00:42:08.620 are and just let YouTube recommend follow-up videos, how many tops does it take until you end up in some

00:42:15.580 really weird, creepy, paranoid, ugly thing? And I usually found it to be in the teens, like in the late

00:42:21.260 teens of times, sometimes sooner, sometimes later. But that's horrible. That's horrible. So YouTube's

00:42:27.340 garbage. YouTube's a cesspool, despite many, many wonderful, valuable things on it.

00:42:31.580 What's something that's better? I might be biased because I work with Microsoft,

00:42:37.260 but GitHub, this place for sharing projects, is pretty good. There's an occasional jerk on GitHub,

00:42:45.660 but mostly people have a stake and they have a shared stake with people in their teams.

00:42:50.060 So it's a little bit like the thing I'm talking about for social media, where if you're working

00:42:54.460 with a bunch of people on a coding project, the last thing you're going to go and do is mess up the

00:42:59.900 value for everybody by mouthing off at somebody for no reason, like all of a sudden you feel a bit

00:43:04.300 of a sense of responsibility to your mates, your compatriot, right? That's a good thing.

00:43:11.820 So, and I believe in monetization, because I believe when people have a stake, when they have

00:43:18.140 something to lose, they'll think through their actions more. If the only thing you have to gain

00:43:24.140 his attention and you have nothing to lose, then you'll be a jerk because you have the incentive,

00:43:28.860 right? It's really simple. So I want to monetize the stuff. And then people say, oh, but I don't

00:43:33.580 want to have to pay. But the thing is, if you're earning, paying won't feel so bad. Like if somebody's

00:43:39.340 gainfully employed, it doesn't bother them to pay for stuff that keeps other people gainfully

00:43:44.140 employed so they can afford to buy whatever they do. Like if people are part of the cycle of the economy

00:43:50.220 and are part of a social contract that makes sense, they're more ready to pay for stuff.

00:43:55.340 And we already know online that people pay for stuff. Sometimes it's possible. It doesn't have

00:44:01.420 to be all the time. It doesn't have to be like this religious thing that you never have an exception to,

00:44:06.140 but we just need a lot more of it. And Jaron, you're talking about, you know,

00:44:12.940 social media platforms and you mentioned Facebook and I agree with you and YouTube as well.

00:44:17.500 To me, the worst one is Instagram. It's the worst one. And there's somebody who used to,

00:44:25.820 I used to teach for many years. Go for it. Go for it. Tell me.

00:44:28.940 It's the ideal platform for me to do my bikini modeling. So

00:44:31.980 I don't know what you're telling me, but yeah, no, no, I hear you. I hear you. I get it.

00:44:38.780 But just, but the effect that it has on particularly on children and young girls in particular is awful.

00:44:46.540 Yeah.

00:44:47.580 It's awful. And, you know, and then I was reading, they didn't do it, thank the Lord,

00:44:53.100 but then they were talking about Instagram for kids.

00:44:56.140 I know. Oh, and, and like, you can just imagine the meeting where that came up and

00:45:01.340 somebody saying, how can we expand? How can we expand? Well, there's this population of kids,

00:45:05.820 you know, you know, it's just like, it's incredible that that conversation could even have happened.

00:45:12.140 Yeah. Amen.

00:45:14.540 Jaron, let me, were you going to?

00:45:16.140 No, no, because no, no, I was going to follow up because to me, the thing that is awful about

00:45:20.700 Instagram, it's the obsession with the physical self and the constantly looking at other people's

00:45:28.780 lives and other people's bodies and what that breeds in young people, particularly in women.

00:45:34.060 And, you know, I walk around London now and there's so many people having plastic surgery

00:45:40.460 and I haven't, there's no studies in this, but I go, this must be something to do with

00:45:45.260 what people are seeing constantly online, constantly on Instagram.

00:45:49.180 Right. Well, look, um, a couple of things to say there. Um, one thing I've learned through

00:45:58.220 hard experience is that if a causation connection seems right, it doesn't necessarily mean it is

00:46:05.260 right. So you say that the plastic surgery thing tracks Instagram, it sounds right to me, but

00:46:12.700 I'm trying to keep us to somewhat high standards of scientific study, even though that's hard with

00:46:17.420 this because it happens so fast. And the only people with the data are the creepy people doing

00:46:22.460 it. So it's, it's difficult, but I wouldn't, I would just put that in the category of kind of

00:46:29.100 makes sense, but not confirmed. It's my advice to you. Like, like we should do our best to try to,

00:46:36.460 because we can make all the arguments we need to and chart a future path without convicting every

00:46:42.140 little possible wrong along the way. Cause it won't even happen. Nothing will happen anyway.

00:46:47.100 So, uh, that's one thing I'll say. The next thing I'll say is, um, uh, the pumping up vanity has

00:46:57.020 been happening, happening for as long as, uh, women could spend money on anything or men,

00:47:02.140 for that matter, just people like there's a sense in which it's ancient and, and yet there's a sense in

00:47:08.380 which it's more immediate, more programmed and more, um, kind of nasty now than I think it was before.

00:47:15.420 Um, so, um, it's, it's easy in this stuff to get caught up with this argument. Well,

00:47:20.700 it's always been that way, which is kind of true to a degree, but then the question of the amount

00:47:25.980 and kind is also important. Then there is a difference down. Sometimes it can be a little

00:47:29.740 bit hard to be articulate about what the difference is, but it's, it's definitely there. And I think it

00:47:33.900 has to do with operant conditioning with this use of behavior modification as technique.

00:47:38.060 Well, right. I was going to say human beings are the way they've always been, but the tools

00:47:42.860 are much more powerful. And I think that's, that's where we are, which is one of the things I wanted

00:47:47.260 to talk to you about because you talked somewhat almost romantically and wistfully about the early

00:47:52.780 days of, of the internet. And I remember that time, uh, freedom, connection, every, anything is

00:47:58.540 possible. What are you talking about? Say again? You weren't born yet. What are you talking about? Well,

00:48:03.340 maybe not in those days, but I certainly remember like first encountering the mass

00:48:08.220 phase of the internet, right. When the ordinary user was using it. Right. Uh, and I was, I used

00:48:13.660 to play a lot of computer games in the time and I'd have friends all over the world that I would speak

00:48:18.140 with. We'd build connections. We'd play games together and we'd, we'd make the most outrageous

00:48:24.700 jokes to each other. And it was free and it was kind of edgy and it was like the wild, wild

00:48:30.380 west. And that was fun. And that was cool. And now we are in a very different place where

00:48:35.740 rather than talking about how much freedom we all have, it's much more about how do we prevent

00:48:40.740 quote unquote harm. And we've talked about some of the harms that come with social media and the

00:48:45.180 internet, but also there's a lot of this, like, you can't say this, you can't say that amplifying

00:48:50.540 this message causes this and that. Do you think the internet will ever be free again, even in the way

00:48:55.980 that it was? This is why I'm proposing this group structure I just talked about, because

00:49:01.180 on the one hand, if we just say, well, we have to lay off because all this, all of this telling

00:49:06.940 people what to say is leading in no good direction. Well, then the whole world turns to pot because

00:49:12.540 we let manipulative, creepy people have the most power. But then on the other hand, if we let things

00:49:18.140 just, if we get to control to the point where we control those people, we also control too much.

00:49:23.660 And if you want to see what that looks like, look at China. We actually think that looks like,

00:49:27.500 and it's not good. And I think, I think China has become a little crazy, for instance, in its

00:49:33.020 management of COVID because of the inability for people to just communicate in a straightforward

00:49:38.620 way. I think, I think they've really hurt themselves. And so neither of those paths are good.

00:49:45.180 This group thing I'm talking about, I don't know if it'll work because it hasn't been tested enough.

00:49:48.940 So maybe it also is hopeless, but at least right now, it's, it's an idea that's still standing,

00:49:53.740 that gives us an alternative to shutting people down or just letting society fall away to the

00:49:59.660 worst people. It gives us a path through that dilemma, at least it appears to, you know, so far,

00:50:04.700 or maybe some variation of it will, but I mean, it's at least a direction because I really don't like

00:50:10.060 either of the alternatives right now. Right now, if we don't go into the direction of the groups I was

00:50:14.380 talking about, we either have more and more censorious behavior online, or we have more and

00:50:20.700 more power for the most creepy people, or we have to just shut down the net. All three of those are

00:50:26.540 horrible. So this is another option. It's another direction. Maybe, maybe. Do you have a website or

00:50:32.540 do you plan to have a website? Because if you do, then Easy DNS is a company for you. Easy DNS is the

00:50:40.060 perfect domain name registrar provider and web host for you. They have a track record of standing up

00:50:45.660 for their clients, whether it be cancel culture, de-platform attacks, or overzealous government

00:50:51.500 agencies. He knows about that. So will you in a second. Easy DNS have rock solid network infrastructure

00:50:58.780 and fantastic customer support. They're in your corner, no matter what the world throws at you,

00:51:04.620 unless it's your ex-girlfriend, in which case you're on your own. You know about that.

00:51:09.660 Move your domains and websites over to Easy DNS right now. All you've got to do is go to

00:51:15.020 easydns.com forward slash triggered. That's easydns.com forward slash triggered. Use our promo code,

00:51:21.900 which is also triggered and get 50% off the initial purchase. Sign up for their newsletter,

00:51:27.340 Access of Easy, which tells you everything you need to know about technology, privacy and censorship.

00:51:34.780 It's interesting you mentioned that, Jaren, because we use a platform called Locals,

00:51:39.420 which is a bit like Patreon, but there's a community element to it. So it's not just

00:51:45.260 money for access to extra content. It's a community of people who love our show.

00:51:49.500 And, you know, there's about 15,000 members in total. There's about a thousand super active users who

00:51:55.740 contribute or maybe 1,200, something like that. And we, at this level, at least, we don't have to do

00:52:01.660 any moderation. People aren't, generally speaking, having a bitch fights with each other every three

00:52:07.980 minutes. People are generally respectful. There's a sense of, like, everybody's pulling in the same

00:52:12.620 direction. And one or two times when people have got out of line, me or Francis have gone,

00:52:17.660 guys, remember, this is like a cool space in which we all hang out and have a good time.

00:52:22.940 And everyone's going, oh, yeah, that's cool. Right. And people come and go. But I think your idea

00:52:27.660 sort of works. But I suppose the question for me is, in an internet that connects everyone to

00:52:34.380 everyone, how do you do that? Right. Well, all the platforms were

00:52:40.460 cute when they were small. Yeah. Yeah. Earliest days of TikTok weren't toxic. TikTok is currently

00:52:46.300 toxic. Right. The very earliest days of Instagram were actually cute, as hard as it is to remember

00:52:51.820 now. So smallness works. The question is, how can you make a network of small things that maintain

00:52:59.180 the good qualities of smallness while still scaling to the world? So that's the idea of

00:53:04.140 presenting to you as an attempt to do. And it's going to be a project. It's a civilizational level

00:53:12.460 project. I can't tell you that I know exactly how to do it. I just believe that through process of

00:53:17.900 elimination, it's the last idea standing I know about. And I don't see anything that kills it

00:53:24.460 offhand, unless you just want to be cynical and say, ah, it's all going to be shit, which is,

00:53:28.060 you know, it might be. It might be, but maybe this is a path forward. I think there has to be a small

00:53:38.780 enough immediate community that it's possible for people to really talk to each other and for

00:53:44.540 not to become lost in this giant ocean of bots and unknowability. That's absolutely essential.

00:53:49.260 social. But there has to be, there's another side of it I haven't talked about. Right now,

00:53:59.980 people, the way people perceive the world is social. This is something we often don't realize

00:54:05.740 about each other. There's an experiment I sometimes do with my students where if you go out on a crowded

00:54:10.380 sidewalk and you just start pointing at something, everybody will start looking there, even if there's

00:54:13.980 nothing there. And the reason why is that we evolved to rely on each other for cues about danger

00:54:20.380 and where to focus attention. This is something that's deep in us. Okay. So if you have this giant

00:54:25.900 unknowable crowd, whatever the, that thing, that giant crowd indicates, you will start to feel.

00:54:32.460 So if there's enough bots in it to try to indicate, indicate that your society is shit or that the other

00:54:37.100 society is shit or whatever it is, you'll start to feel that. Now, if you can turn it into a

00:54:42.700 collection of groups as I've been talking about, maybe you can cut back the total number of players

00:54:47.980 by a hundred times or something like that, maybe even a thousand times. Now, your interface to that

00:54:55.260 world, instead of looking like this feed from a zillion people, will look like a newsstand with

00:55:01.180 covers that you've started to like. And you can add or remove from your newsstand and start to dive

00:55:05.340 into it, each of which representing a group. And in that circumstance, the way you receive information

00:55:11.900 is given context and package. So there were horrible Nazis and whatever, whatever kind of

00:55:20.060 creepy person, jihadists, whatever these people have existed before. But the thing is, they always

00:55:25.660 sort of labeled themselves like we're in this magazine or we're in this, whatever, when they're

00:55:30.060 just part of the feed, it just makes the whole world feel horrible. Part of it leads to this rising

00:55:35.580 paranoia for everyone. If everything is once again, compartmentalized where there's at least some

00:55:40.780 context, context hasn't been destroyed. Then if you see some horrible message is, oh yeah,

00:55:46.620 those are the Nazis, whatever, fuck those people. So it gives you that ability to not have it be part

00:55:51.740 of your social queueing. It's an extremely important issue I hadn't talked about before. So we have to

00:55:56.540 think about realistically how people work cognitively and how to just build a technology that works with us

00:56:02.380 as we are, instead of as how we imagine ourselves.

00:56:05.100 Jaron, your idea is brilliant. The one area of pushback that I would say is,

00:56:11.020 aren't we therefore just creating more echo chambers by doing this?

00:56:14.300 Yeah. I mean, look, what'll happen, this is why I want it to be monetized.

00:56:21.500 If you have a bunch of people just saying, oh yeah, you're great. And those other people are

00:56:25.180 shit. And then you have a whole bunch of those, how many people are going to subscribe to that?

00:56:28.620 Well, a certain amount, it might even be a fairly large amount, but it won't be most people.

00:56:33.020 Like what'll happen at a certain point when you get to the point where you have

00:56:39.260 bands and brands and stuff, there's a bit of a filter that happens where if it's just self-indulgent,

00:56:46.060 echo chamber-y stuff, it'll start to fall. There's an amazing thing that happens. Like for instance,

00:56:52.860 if you look at the history of music, you might say, well, the music business is filled with producers

00:56:59.740 doing calculated artificial stuff and crappy artists who got in through nepotism and blah,

00:57:04.700 blah, blah, blah, blah. And all of that is true. And yet, if you look at the things that were hits

00:57:09.500 and the things that persist and the things that lasted over the years, at least in my opinion,

00:57:13.980 there was kind of a sieve in which quality stuff rose, you know, overall, with some exceptions,

00:57:19.020 and we can disagree about individual cases, but overall it happens. And I think it would start

00:57:24.220 to happen online too. That makes a lot of sense. One of the reasons I think the people that like

00:57:29.580 our show, like our show is they know we're not right or left. We're just trying to find the

00:57:33.820 answers in a complicated world. And we've got our own biases, which we're upfront about, but we're not

00:57:39.660 pushing an agenda. Because I do think that gets very boring. And we've seen other people go through

00:57:43.420 a process where they just end up in one thing and they're not curious anymore. And I think that's

00:57:48.140 very off-putting. I know what you mean. Look, we've done a lot on social media. We could talk for hours

00:57:54.540 more, of course. Talk to us about AI. You mentioned you were extremely sceptical. And about VR as well.

00:58:02.380 Just give us your, you know, a Cliff Notes version of your thoughts.

00:58:05.500 Oh yeah, sure. Well, look, on AI, just to be clear, the actual algorithms I'm a total enthusiast of,

00:58:12.220 and I've worked on and contributed to. To my knowledge, my little group, a little startup was

00:58:18.700 the first to do deep fakes, as we know them, and also the first to do snap filter kinds of things.

00:58:23.740 Because I think these types of algorithms have their uses and can, you know, and they're

00:58:29.180 scientifically interesting. The thing that bothers me though, and I think that tends to create a cover

00:58:34.380 for the worst uses of these things that make them, that can make them feel anti-human and indeed be

00:58:39.660 anti-human, is this ideology that we're building a life form in the box. Now, within, and even the term

00:58:46.780 artificial intelligence suggests an intelligence, you know, it suggests that there's some entity,

00:58:51.900 some creature, some, you know, something alive in there. And I really don't like that way of

00:59:00.140 thinking. I want to think of them as tools, not as creatures. Now, as it happens, the single person

00:59:05.900 who did the most to promote the creature way of thinking about computers was named Marvin Minsky.

00:59:11.180 And he was one of, probably my most important mentor when I was young. And then he started to

00:59:15.580 argue with him about this. We're going back to like late 70s, early 80s, something like that.

00:59:21.260 And I'd say, you know, this is bullshit, Marvin. Like these things aren't alive. Like, look at this

00:59:25.260 thing. And I know, God, I hate to say this. Marvin wouldn't want me to say this, but he would say, look,

00:59:32.140 this lab is funded because the military believes that if we don't build these creatures that our

00:59:36.380 enemies will. So just play along and then you get a salary. You know, it's like a marketing thing.

00:59:42.460 And so I said, okay, okay. You know, I did it. And I think it still is sort of a marketing thing,

00:59:47.980 but it's become sort of a religious thing. It's like, we, the nerdy guys get to create life. Screw

00:59:52.940 those women, like, or don't, don't screw them or whatever. Like we're, we're gonna, we're gonna,

00:59:59.340 we're gonna make these life. And, and the thing is, everything gets clearer. Totally. So there's no

01:00:04.860 absolute truth on this because nobody can really know what else is conscious or what else we like.

01:00:09.820 We don't really have, is evolution itself as a process conscious? Is it intelligent?

01:00:17.980 You can debate that. I think you can see it either way. The terms have a lot of potential

01:00:22.300 wiggle room and definition and whatever. I mean, it's interesting, but you can't ever come to a

01:00:27.580 conclusion because there's no way to define your terms. And we don't have, if I, if I can get slightly

01:00:32.700 sophisticated, my language use, we have no empirical channel to gather evidence to help the argument

01:00:37.660 along one way or the other anyway. So it's not resolvable, but we can ask pragmatically,

01:00:43.740 which way of thinking makes us more competent? I think that's a reasonable question. And thinking

01:00:50.540 of them as tools instead of creatures makes us more competent because then we can evaluate them

01:00:54.780 as tools. If we think they're a creature, then we're giving them too much deference. We never make

01:00:59.260 them better in the ways we need to. So like, if you, that's why I really don't like it when people

01:01:04.540 say, Oh, I'm going to have this program that'll make music art for you. And it'll be better than

01:01:07.740 a person. And the reason why is that then you'll change to make it seem better.

01:01:11.980 The way I used to, the way I used to put it is you'll make your, do you know what the Turing test

01:01:17.900 is? It was this idea from Alan Turing who started computer science. And it's ironic because if you

01:01:25.340 know anything about the story, he wrote this just before he committed suicide because, and it's a,

01:01:31.740 it's a long, amazing story, but at any rate, if you take it at face value, what it says is

01:01:38.700 if you can't tell whether a program is a person or not, you might as well call it a person. Otherwise

01:01:43.740 you're prejudiced. Now, of course he was, he was being essentially tortured for being gay

01:01:49.660 and treated as a non-person. So there's a whole level to this thing that we could talk about, but

01:01:54.380 let's leave that aside for a second. Um, the, the thing is, if it's, if it's true that somebody

01:02:01.660 cannot tell if the machine is acting like a person or not, maybe they're behind it, you know,

01:02:06.860 they're just texting with you. Um, then it's one possibility is the machine elevated itself and

01:02:13.180 became a person. The other possibility is that the human lowered themselves and became an idiot to

01:02:17.580 believe the machine's a person that the test doesn't allow us to distinguish between the two.

01:02:22.060 This happened just last week where an engineer was fired from Google to sort of believe their

01:02:27.100 program had become alive, but it's also possible that that guy became an idiot.

01:02:32.220 There's no empirical or scientific distinction between these two things. We can't tell which

01:02:36.700 is which. Anyway, um, people demonstrably are willing to make themselves idiots to socialize.

01:02:42.780 We do it all the time. We all become morons to try to impress a date. We all become morons to try

01:02:48.060 to impress a potential employer. We all become morons all the time with each other

01:02:53.660 in the hopes that it'll do some good. And every once in a while, maybe it does. It usually doesn't,

01:02:57.100 but we do exactly the same with machines that we think are people. So let's not think they're

01:03:00.860 people. Let's call them tools. If whatever it is, isn't working, let's say that tool needs to be

01:03:05.660 better designed. Let's not say, Oh, you know, uh, I guess that's just the way it is. It's a creature.

01:03:11.500 No, no, no, no, no, no. And I think this, so this issue of treating them as, as, as tools instead of

01:03:18.380 people is actually one of the core problems that got us into the trouble we're in because Google

01:03:23.340 always conceived of itself as an AI company. And they always thought, well, if it's not doing what

01:03:28.380 we want, the thing is it's an emerging AI and we have to give it the space. It's like a child.

01:03:32.460 You let it grow and go through its tantrums or whatever. Um, and that attitude is absolutely

01:03:37.260 wrongheaded. They're tools. If they're not doing what we want, we change them.

01:03:41.500 And Jaren, on that, speaking of pragmatic questions to ask in this sort of conversation,

01:03:46.940 uh, you know, these tools from the inside out, we don't, is it not the case that the algorithms

01:03:52.380 that say determine our ability to search for things on Google, they're so complex now that

01:03:57.820 actually it's not within the realms of possibility for a single human being to assess that algorithm

01:04:03.580 in its entirety. Right. So this is a, uh, this is a large topic to talk about. And I am,

01:04:12.060 I'm not here representing them at all, but I'm, I'm, uh, the so-called prime scientist at Microsoft

01:04:18.140 and our office is funding and helping guide the open AI algorithms like GPT and DALI and Codex that some,

01:04:26.140 some of your, uh, uh, viewers might be familiar with. And so this question of, can we understand what

01:04:33.180 these things even do? Um, in my opinion, part of the issue is that the culture, tech culture wants

01:04:44.140 them to be creatures instead of tools and kind of set them up to be hard to understand

01:04:48.860 more than was really needed. Now that's going to be a controversial statement and I'll get a lot of

01:04:53.340 pushback from people in the field. And yet I think it's true. Like for instance, one of the problems

01:04:58.860 we have, do you know what DALI is?

01:05:00.700 No. Oh, you should look it up. It's D-A-L-L dash E. It's the latest, um, image generation stuff. So

01:05:07.580 you just give it some text and you say, Hey, I would like, I would like to have an image of a

01:05:14.060 robotic pumpkin, uh, swimming the English channel in the style of Turner. And you know what? It'll

01:05:20.540 synthesize that. It'll come out. And the way it does it is through this process of sort of randomly

01:05:26.940 starting to create images and compare, comparing them with images that have matching keywords

01:05:32.220 nearby and going back and forth until it gets something that it reads as a match. And typically

01:05:37.900 what comes out actually looks like the thing you asked for. It's just astonishing actually. Now,

01:05:43.020 what's wrong with that? Well, there's a few interesting things about it. Um, one thing is

01:05:47.020 if you just give it nothing but the one word Asian, nothing will come out the porn because all it does

01:05:51.260 is it reflects the primary keywords that were online. And I don't think most, the billions of

01:05:56.780 people from Asia really want that result. I don't think that's great.

01:06:04.780 If you, so there's something wrong with the way we're doing it. And so then what we end up doing

01:06:08.940 is it's a little bit like the problem of trying to censor online speech to fix online speech. It's

01:06:13.660 like this game that you can never win. So now what you do is you create this whole organization that's

01:06:19.260 trying to go through the output of these giant models and try to have them not give terrible results.

01:06:24.700 But you know, it's a never ending, it's a never ending struggle. And it's not clear that you ever

01:06:29.740 even get to the point where it's okay. You know, it's just rough. And at what point is this kind

01:06:34.860 of safe for the world? You know? Um, so it's, that's a tough one. Can I tell you what troubles me

01:06:41.500 about your answer very quickly, right? I asked you whether these algorithms are too complicated for a

01:06:47.980 single human being to understand, and you're a very smart guy. And you gave me an answer that's

01:06:53.660 too complicated for me to understand. Oh, I'm sorry.

01:06:56.700 No, no. But what I'm saying is, what troubles me about that is, if it was the case that a single

01:07:02.620 being could understand them, I imagine you would have said yes, but you didn't.

01:07:05.980 I, let me give, give me a chance to finish. Okay. What I was saying is right now they can't

01:07:10.700 be understood, but I think they could be, we could do things differently so that they could

01:07:14.300 be understood and actually better. That's, that's the big picture I'm trying to get across. Okay.

01:07:18.620 So let's suppose instead of saying, we'll just randomly take whatever's people upload into

01:07:24.220 the internet and then use that as the basis for these types of algorithms. Um, what if instead

01:07:32.700 we said, you know what, we're going to put out a bounty. We need better input data that reflects

01:07:38.540 what Asia is. And then what some of these clubs I was talking about, we sometimes call them mids,

01:07:43.660 but whatever, there's like these data trusts, these clubs of people like the ones who would publish,

01:07:47.420 um, they would start saying, okay, we're, we've just assembled a hundred thousand images of Asia.

01:07:53.660 And then the, the people running the algorithm would say, wow, we just tested that. That improves

01:07:59.100 our performance. We're not just getting porn anymore. You know, we'll pay you this much.

01:08:02.860 And then they say, well, you know, we're going to collectively bargain with you. We think we

01:08:06.220 deserve that much. And then they, they, they go back and forth and what they get is a fair market

01:08:10.860 price. So what you're doing is you're using the market to fix it. And now people understand what's

01:08:16.860 going on. And the reason, you know, they understood it is they got paid, right? If they didn't

01:08:22.540 understand it, they wouldn't get paid. And so basically you, you end up in the same situation

01:08:27.100 that you have in capitalism, where understanding is reflected in payment, because if you're not,

01:08:33.820 you only get paid, if you're generating enough value for the person to be able to pay it or being

01:08:37.660 willing to pay it. So the successful negotiation is your validation of understanding, whereas

01:08:42.620 otherwise, what the hell would it be? So what you do is instead of saying, oh no,

01:08:46.220 how can we fix this horrible thing we've made? You say, um, how do we motivate people to make

01:08:50.780 a wonderful thing? All right. That's the answer to say, oh, I'm going to have these eggheads,

01:08:56.220 people like me who are going to understand it is ridiculous because you shouldn't even trust us.

01:09:00.540 Like, suppose we could understand it. Do you want to trust me to, no, of course not. That's not how

01:09:05.420 civilization should work. Do I want to trust a marketplace made of groups of people who are earning

01:09:10.700 their way to improving the performance of these things? Yes. I like that. And then what I like

01:09:15.340 also is that then every time there's some new algorithm or some new robot, instead of the people

01:09:20.700 saying, oh no, this thing's going to put me out of work. I thought I was going to make a living

01:09:23.660 as an artist, but now this thing's going to make the art. What will I do? I'll just be living in the

01:09:27.660 street. Instead they say, oh, this is a great opportunity. I'm going to join one of these groups and

01:09:32.060 we're going to improve that thing because right now it's garbage and we're going to make it better.

01:09:35.420 We're going to get paid for it. So it takes, it inverts the whole thing and turns what we call AI

01:09:41.260 into this endless, infinite set of new opportunities instead of this thing that

01:09:45.020 makes people feel obsolete. And that's what I call understanding.

01:09:50.700 Well, that is a beautiful message to end on. How realistic that is, we are going to find out.

01:09:55.100 Jaron Lanier, it's been an absolute pleasure. We're going to ask you a couple of questions from our

01:09:59.660 supporters for our supporters. But before we do, in that local group that I told you about,

01:10:04.780 right? The small group, people working together. But before we do, we've got one final question for you,

01:10:09.980 which is always the same.

01:10:11.660 What's the one thing we're not talking about that we really should be?

01:10:16.860 Oh, that's interesting.

01:10:17.980 Does AI fulfill the same emotional needs as your ridiculous royal family for people?

01:10:28.140 You're going to have to expand on that a little bit.

01:10:33.100 Well, as an American, I was looking like, what is this royal family? Why is it there? Why is,

01:10:37.500 what is all this fascination?

01:10:38.780 It's our class system, Jaron.

01:10:41.660 Well, no, but the thing is, it's this, see, I think AI sort of is too. It's like, it's this,

01:10:46.860 it's a set of principles that come from very powerful people that are supposed to run society.

01:10:52.620 You know, I'm starting to feel like this whole gambit is falling flat. Never mind.

01:10:56.780 My answer is, forget it, forget it.

01:10:59.660 All right.

01:11:00.860 Jaron, it's been an absolute pleasure. If people want to find your work online, where's the best

01:11:12.140 place to do that? Where is the best place to engage in this wonderful mind of yours?

01:11:15.900 The last book I wrote was, wait, which one? I'm not sure if I remember the order. The last two were

01:11:22.780 called Dawn of the New Everything, which is a memoir of the early days of virtual reality in the 80s.

01:11:29.660 And the social media book I wrote was called 10 Arguments for deleting your social media accounts

01:11:35.820 right now. In the UK and UK adjacent countries around the world, it's published by Boadly Head,

01:11:43.020 which is a fun, a fun publisher you have. And, uh, um, I think on the, the VR memoir,

01:11:51.340 the head is wearing a headset. Um, it's an actual, it's a head anyway. Um, and then there's two more

01:11:57.980 books coming out, but I'll, I'll leave you, I'll leave you in suspense about those. Well, uh, if, if,

01:12:06.380 if you did give us your time once again, when the books come out to chat about what you're talking

01:12:11.260 about and then we'd be very, very grateful, uh, don't go anywhere. Cause we're going to ask you

01:12:14.940 a couple of questions for our locals, but in the meantime, thank you so much for joining us.

01:12:18.700 And thank you for watching and listening. We'll see you very soon with another brilliant episode

01:12:22.540 like this one or or show all of which go out at 7 PM UK time. And for those of you who like your

01:12:27.340 trigonometry on the go, it's also available as a podcast. Take care and see you soon guys.

01:12:32.300 Do you see a connection between the emergent use of social media and virtual reality? And on the

01:12:40.220 other hand, the growth of the feeling that we sort of all get to choose our own identity.

TRIGGERnometry - July 25, 2022

Jaron Lanier - The Truth About Social Media

Episode Stats

Length

Words per Minute

Word Count

Sentence Count

Misogynist Sentences

Hate Speech Sentences

Summary

Transcript