Sharp Tech with Ben Thompson - Mailbag: Down the 230 Rabbit Hole, A Third Door Moderation Approach, Twitter Follow-Up, Andrew Responds to the Nicest Angry Emails Ever

Episode Date: November 7, 2022

A question about Section 230 spawns expansive thoughts from Ben, a proposal for an alternate solution to moderation, new questions on Twitter as the Mastodon takeover approaches, and responses to listeners who balked at Andrew's broad social media dismissal last week.

Hello and welcome back to another episode of Sharp Tech. I'm your host, Andrew Sharp, and on the other line, Ben Thompson. Ben, how you doing? I'm doing well. I mean, I continued with my usual policy of mostly ignoring Twitter over the weekend. And I think that by default means I'm in a better state than pretty much everyone listening to this podcast or otherwise. Oh, man. It's been a fun ride on there.

And we are going to get to a number of follow-ups we had to our Twitter. We've got a lot of great emails to get to today, so I'm not going to do too much prelude. Email at sharp tech.fm is the email address if you want to send us questions or comments for future episodes. But let's dive in. Before we get to Twitter, Dylan asked, do you have any thoughts on how the Supreme Court case on Section 230 might impact algorithmically delivered content? And the Supreme Court case for anyone who's unaware is Gonzalez v. Google. It's a case in which the family of a 23-year-old American in Paris is suing Google after their son was murdered by individuals affiliated with ISIS. And I'll read an excerpt from an article on Vox.

Their theory is that ISIS posted hundreds of radicalizing videos inciting violence and recruiting potential supporters to YouTube, which is owned by Google. Significantly, the Gonzalez family's lawyers also argue that YouTube's algorithms promoted this content to, quote, users whose characteristics indicated that they would be interested in ISIS videos. So what do you think, Ben? I mean, how far down the rabbit hole do you want to go on this one? Section 230, it's actually, there's an aspect, I think since there's going to be a little more well-known, I've written about it several times. in very broad strokes, there were a couple of cases where internet companies were held responsible for content posted on their site, even though it was posted by users.

So it was user generated content. And Section 230 was sort of a direct response to that, basically saying, you know, sites cannot be held responsible for content posted on them. Number one, and number two is the critical part, even if they moderate. because there was actually two cases. The first case was where they were held guilty. And that was, I think, thrown out by another court. But then number two was actually this prodigy case that had to do with Wolf of Wall Street, the same firm.

Jordan Belford. Yes. Well, I don't know it was him specifically, but it was that was, what's the name, Oakmont Hill or whatever it is or other. And so someone posted something about, you know, that being having problems, and they sued, and they won. and the court in that case said that prior precedent that was thrown out didn't hold because prodigy did try to moderate and therefore by the act of moderating they assumed responsibility for all content on there. And so Section 230 was written pretty explicitly directly at that use case, basically saying

like a good Samaritan provision where if you try to moderate, that doesn't make you liable for everything. Now this is pretty well known. what's less well known is the broader context of Section 230. And a couple of years ago, I actually went back and read the entire congressional record around the entire thing. And all of the talk and discussion was about protecting kids from porn on the internet. Like that was the entire context of it. And it was like, look, if we don't allow companies to moderate, there's going to be, you know, porn all over and our kids won't be safe on the internet.

That was that, that was completely the context. there was no real. And so one thing that's really interesting is over time, Section 230 is in many respects a court-built statute where just sort of precedent over precedent has really expanded it to be this sort of broad, wide-reaching thing. And I think it's like I don't say that critically. I think every, all those decisions have been very logical and do follow. It's one of those things where if you're a textualist and you look at the text,

I think all these court decisions do fit with the text of the internet. that what's very or text of the statute, I should say. What is interesting is that the intent of the people passing it was not necessarily aligned to where Section 230 has ended up. And so I think that's something that's a little underappreciated by people who take these super absolutist positions about Section 230 and call everyone opposed to it, morons. Actually, I think you can make a good case that some of the people who supported and pushed Section 230 at the beginning are actually ideologically consistent with people who oppose Section 230 today because, it ended up being used used in ways in a much more expansive sort of application than was sort of ever intended or expected. Now, again, I'm not saying that's wrong per se. It is interesting context around this, though, that I think is sort of underappreciated. Yeah, and I think if you go

back to when the law was written, the internet was just a radically different place in dozens of different respects. And so the idea that that law should probably be revisited is not necessarily crazy. I mean, they didn't have any idea what they were legislating or what the implications were going to be. I think I think that was your point just now. No, no, it is true. But, but I think what the section 230 defenders do get right is, and maybe this is your point, is the modern internet out today would not exist without section 230, without question. Like, basically the entire realm of user-generated content would not be viable without Section 230. Now, I think we have some questions on here.

We'll get to you later about why do you hate the common man and denying his ability to sort post on the internet. But that would not exist otherwise. It just wouldn't be possible. And you could sort of work your way down the stack where, well, it might be possible. We'll get to that in a moment. So Section 230 is foundational for the Internet sort of as it is today. what is interesting about this case specifically, though,

this case is different than that case like a couple of months ago in Texas, where Texas basically said, if you moderate your exercising editorial control, and that's a violation, that's clearly inconsistent with the way Section 230 has been interpreted for years and years and years and years. And if we care anything about precedent, you know, at least technically only the Supreme Court can overturn precedent,

that case should be turned around. But that case is talking about moderation. This case is about promotion. And so it's actually a very interesting legal question where is moderation the same thing as promotion? In some respects, you could argue it's negative promotion, right? By not moderating some stuff, you're promoting it. But in this case, the algorithms absolutely do promote certain things, that just sort of the way that they work.

And so I actually think this case is not as open and shut as some people want to presented as being. The question here is can a company be unreliable not for moderating, but for actually suggesting something for you to watch, putting something forward in front of you. And there is a consistency with saying this should be covered by 230, but that would fit in the historical trend of 230 being expanded far beyond what it was intended to be, which again, I think all those decisions in isolation make a lot of sense. it's interesting to it's one of those things where you look at where we're at today and where we were in 1996 and it's we're definitely at a very different place and on the flip side I think

there's actually a strong case to say look no moderation is clearly exempted by section 230 that has nothing to do with promotion and it's actually completely opposite so I understand the argument that the internet that we have today simply would not exist if not for section 230 where I start to raise an eyebrow is when people defend the integrity of Section 230 to the extent that it basically creates immunity for all of these tech companies to do whatever they want and not really be held accountable. And I understand it would impair different aspects of their business. But in terms of like the theory, it makes sense to me that we're going to need a way to strike a balance between where we've been and where we're going. It's a good point. I do.

it is worth pointing out that all these companies have operated under the assumption that all this is covered by 230. And so on one hand, you can say, oh, that's very bad then to impair that because you'll hurt these companies. On the other hand, you make the case that, well, there haven't been any sort of controls or incentives in place to even be careful, right? I wrote an article a few years ago about the, like, the Polyanish assumption about how so many of these companies' decisions are sort of predicated on the assumption that like everyone is well acting and good. And that was, you know, this was six or seven years ago. I think there's been a real wake up and awareness of, oh, there's a lot of bad actors sort

of on the internet. But for sure, like section 230 and the assumption it covers everything is tied into this broader culture that, hey, just build this stuff and don't worry about what happens otherwise. And, you know, there's a real imperative to move forward. Now, should that be the case? I do think there is an argument that the entire thrust of 230 was mistaken. Now, I'm not just saying I think I think the Texas 23 was a mistake.

Like I said, it's the foundation of the internet today. But there is a way to think about what's a sustainable route going forward that deals with the issues in like that Texas case or deals with the issues in Gonzalez versus Google, particularly from a sort of legal liability standpoint. And that is, if you go back to the first case that prompted Section 230, which was against CompuServe. So the second case was Prodigy, where Prodigy did moderate and Section 230 was basically passed to say, that's okay. It's okay to moderate. Again, it's called a Good Samaritan provision. In the case of CompuServe, they were ultimately held not guilty because they didn't do anything.

It's like, hey, like they're just providing a platform. People can sort of post whatever they want. Now, in practice, this would be terrible, right? Like just look in your email spam folder. Like it's kind of like the go-to example of why some sort of moderation is necessary. Like the fundamental issue on the internet is that producing content is free. If producing content is free, you're going to get a whole bunch of crap content that's produced. It'd be crap as far as see economic favor.

It'd be crap in terms of trying to so chaos. It could be crap in terms of horrible, you know, violence and racism and all those sorts of things that we talk about. that's a fundamental economic structural problem with the internet because it's free to create content. You know, one of the just as an aside, one of the hilarious things over the weekend is like, well, why doesn't Russia just pay $8 per bot? It's like, because that'd be really freaking expensive, right? The whole thing is that it's free, right? So that's one side.

Now, the bigger problem in general is how do you manage and sort of police the internet? And I wrote an article early in the pandemic called Zero Trust Information. And the concept there was making an analogy to the way security is moving. Like how do you secure like a company's assets? It used to be you have like a castle in mode system. You'd build a firewall around your company. Everything inside the firewall was free to move stuff around like no problems. And then if you were like a traveling salesman, you had to use a VPN so you could basically

build a tunnel into the castle and then you could converse and do whatever you would. wanted. It turned out this was a bad idea for lots of reasons. One reason was the explosion of devices meant there was so much more activity happening outside of the physical internet of the office. So that's number one. Number two is like this is software. It's very hard to patch everything. And if someone broke the walls, they had free reign to get access to everything in your company. Like it was a very, very sort of poor situation. So the answer sort of moving forward is this idea of zero trust where you assume that everything accessing any, is not trustworthy and has to be verified and verifiable. And it basically shifts the burden

of security onto every sort of individual device and actor. And you do this through, you know, there's entire services that are built around this. You have to authenticate and you have to do XYZ. So number one, it's like much more distributed. So it just from a functionality standpoint, it works better for all the people using devices outside of the home office. And in COVID has actually accelerated this with all these people working from home. But then number two, if, any one single device is compromised, it's not fatal to the entire structure because it can sort be contained because you have to be continually be authenticating and showing who you are, et cetera, et cetera. That's a sort of a bastardized description, but that's the idea. My argument is this is

the long-term solution to misinformation. It's never going to be viable unless you want to go full Chinese firewall style with all the problems that that entails to control misinformation from a top-down sort of perspective. What we need to do is empower individuals, to sort of figure out how to manage and navigate that along the way. And part of this is just a cultural thing, right? I made the argument on this podcast. I think that younger generation is far more suited to the internet because they've grown up, assuming it's all BS, right?

A lot of the people who are susceptible to misinformation of the internet are like boomers. And we know this from like surveys and stuff. Like they just, they're so used to a world where producing content costs money that they can't fathom that they're, that like, well, if someone bothered to print it, it must be true. No, that's definitely not the case anymore, right? So in this world, you want to sort of empower the individual. And so one of the arguments I made about Twitter years and years ago is that Twitter's fundamental failing as far as abuse and misinformation and all the sorts of things is that it failed to give people the tools to not see the garbage, right? Instead, Twitter tied to eradicate the garbage.

That's an impossible task. It's impossible for fundamental economic reasons because the cost of production is zero. You can never sort of keep up. But if you can enable people, and I argued years ago for things like, look, you should, like one of my proposals, like, and again, we're doing back like 2014, 2015, like back of the Gamergate days. Number one, everyone should be able to be verified. And number two, you should be able to see only responsible people who are verified. Now, as a user, you should be able to decide I don't care. I want anyone to contact me, open up my message to everyone.

And what this does is to combat internet problems, you need internet style solutions, which means you have to leverage scale. You have to leverage zero costs, marginal costs. And by deputizing all of its hundreds of millions of users, Twitter could build a much more sensible system. But there's this mindset that, no, it's not enough to stop people from seeing it. We have to go and kill it. That's just not going to work.

And what's interesting is this actually is close to the CompuServe solution where we're just not going to moderate anything, but we are going to give users tools to decide on their own what to moderate. And so Twitter could say, you start up, say, look, you can access everything. But we've also built a filter for you that is our hand-derived algorithm that we think favors good stuff and not bad stuff, that you can click or you can maybe it comes by. default, but you can clip to turn it off, for example, right? And what's interesting is, is that sort of like solves a lot of these problems where all these disputes are about centralized control, centralized, you know, some people call it censorship, some people call it moderation. Of course, both sides pick an example that fits their preferred narrative. But you have all these court issues. So that's just in broad strokes. You didn't ask it, but you have it on the rundown.

Like if I were to regulate algorithm the content, I think the only sustainable solution is something that works at internet scale. And what works at internet scale is actually giving people tools in mass so you get hundreds of millions of individual decisions that do pull in like a certain direction. I hear what you're saying and it does make sense to me

that that is frankly just more practical. Like if you were to repeal Section 230. This is a practical argument. My whole thing about controlling misinformation, weaving aside, this is not a partisan thing. It's just not, it's not. it's not going to work. Right.

It's not going to work. And also, if you repeal 230 and allow people to sue based on misinformation or terrible things that happen as a result of the internet, there's going to be an avalanche of litigation. And it's going to be really difficult for courts to even manage that issue. In almost all cases, the companies would win anyway because of the First Amendment. What Section 230 does is just basically like cuts off the entire like, let's, let's forego the lawsuit process, all the cost, because we already know what's going to happen, right? They're going to win.

Like, the First Amendment is the ultimate backstop against a lot of this sort of stuff. Companies have the right to make editorial decisions. Like, that's just, that's just the fact of the matter. And so even these 230, like, so the people who are advocating against repealing 230 are basically hoping for control via lawsuit paralysis, which is just terrible. Like, like, on the face of it is bad. Yeah, it's terrible. I do understand the impulse, though, because otherwise, again, it's just impossible to hold these companies accountable for what's happening on the platform. And that seems like it's a problem that we're going to need to figure out how to solve. I will say, though, I'm uncomfortable with ceding this issue to individual courts as well. And even if it's the Supreme Court, I'll be very curious to see how they handle it, because ultimately, it's up to Congress. And it's Congress's fault that none of this has been addressed since 1996.

And so they need to grapple with these issues rather than leaving it up to partisan courts here and there to figure out what the right answer is. And that goes back to a bigger grievance. We just don't really have a functioning legislature. Well, I think there's two points on that. So one, like that is a pushback. And this is a pushback on like, I think a lot of the antitrust stuff versus tech is misguided, not because there isn't a problem with marketing.

power, but because the statues as currently sort of written don't really address the issue of what I call aggregators. They are fundamentally different than platforms. And then, again, you can absolutely make the case that the current interpretation of antitrust statues like the Chicago school makes the issue even worse. But the reason why I think it's full-hearty to necessarily try to change that via the courts is you're going to end up in a situation where you just lose all the time, like the FTC is doing in a lot of cases here.

and then you're actually in a worse place because you've sort of delegitimized any sort of oversight power. And the obvious solution is Congress. And I don't completely accept the, look, it's hard to pass stuff in Congress. Like that was the design is hard to pass stuff. Has it gotten harder in the last 30 years, almost certainly? But like that really is the only way forward. And that's not the worst thing in the world to be having some sort of democratically accountable body making these sort of arcane decisions. Now, number two, there is the First Amendment, though, to.

consider. And I think that is what makes all these discussions very challenging is particularly once the First Amendment was sort of extended to corporations where that's maybe the issue where people should be pushing on. Because if a corporation can do whatever it wants, whether it be asserting editorial control or algorithmically promoting XYZ, then Congress's hands are completely tied. Well, the one thing I would add is this makes it imperative for Congress if they did want to do something that would potentially erode corporations and their First Amendment rights, they would need to articulate like a compelling government interest that makes that reasonable. And a lot of this stems from people's general uneasiness with where things have been trending. And I'm not sure we're in a place where there's been some horrible tragedy. that compels people to really articulate exactly what is going on and what these platforms are

responsible for and how we should try to regulate it. Like everybody's just sort of fuzzy on this. Well, because that's the whole thing. What people are truly uncomfortable about is they're being exposed to a much greater extent than they ever were previously of the opinions of people they hate and disagree with. And that's the sort of stuff that's not going to be addressed by Congress. because it really like political speeches was actually making people upset. Now, even you go to something like Gonzalez versus Google, in this case, this is a 230 case

because the case was dismissed. If even if the court rules against Google here, Google is almost certainly going to win because there was not an imminent encouragement to violence or whatever the sort of like very tight limitations are on what is sort of, you know, when it comes to the speech that was in this case. Like, so Google's going to win regardless. The question is they win because of 230. or they win because of the First Amendment.

But to your point, what makes people so upset in my estimation about the current environment is not ISIS, it's Trump or it's the woke or it's XYC, which is political speech. And this is why just from a sanity perspective, I hope the companies go in the empowering users' perspective. Right? Like, wouldn't it be great if I could get the default Twitter algorithm or I could suburb? or I could subscribe to X, Y, Z and get their algorithm, or do ABC and get their algorithm. Like, this whole idea of mixing people who completely disagree with each other,

like the whole filter bubble idea was a total mistake, right? Like, no, we were much better off when we were in our own bubbles. Like, that's part of the issue. Yeah, well, and I think you hit the nail on the head. Like, when people are complaining about misinformation and the consequences of different moderation policies, what they're actually complaining about is polarization that's more a byproduct of echo chambers.

And I do think algorithms play a role. No, I think I disagree. All evidence is that encountering regularly people you hate and disagree with hardens your opinions, not the other way around. It's not an echo chamber thing. It's like there's an echo chamber on like, you go to like cable news and you get your echo chamber. feels very comforting.

You come online to Twitter. Like, Twitter has done a phenomenal job of exposing people to opinions they disagree with. That's why that's why people get very upset. Well, Twitter also incentivizes consensus and escalating opinions to be more and more extreme. I think there's a radicalizing function on both sides. No, for sure. For sure. Yeah.

But part of that is you're radicalized into your side. Even if I disagree with my side, it's like, well, I could. see that other side. That's nasty stuff, right? And of course, your side's elevated the other side's nasty stuff. So you can all see it and realize, oh, no, no wait. Yeah. Well, all I know is that 10 years ago, there were like 5 to 10 percent of each party that were just totally uninterested in any compromise. And now it feels like closer to like 20 percent on each side. And they're just, again, not interested in working things out with anyone who even moderately disagrees with them. And that's a not great place to be.

Like, if 40% of the political public are just totally unreasonable. And I think we are not going to be able to solve social media's role in creating that reality. And who knows how significant the role has been. We're not going to solve it on this episode. I mean, not to sound like a Marxist here, but I think there's structural issues. that are underappreciated here. And I talked about it a couple of podcasts ago where the issue, like my argument is

actually no, the mainstream media was much more responsible for collecting Trump than social media was. But social media made the mainstream media desperate enough for attention that they didn't care who they promoted and put forward X, Y, Z. And so I think there's a multi-level argument here where it's really easy to look at the surface things that are going on. And then I can take the devil's advocate side, but I can back up and take like a very sort of deterministic, structuralist view and say, well, in a media environment where everything

is equivalent, everything is equal. Like, that's why we have like the celebrity politician, right? Because like breaking through and getting attention is the most difficult thing. And being extreme, being highly differentiated is a way to break through, right? Differentiation does not necessarily mean good. differentiation means polarizing. Like that's kind of the point of it. And again, that polarization could be something benign, like high quality versus low

quality. But in the case of politics, it's going to be, you know, values A versus values B. And that's, that's tough. Yeah. Well, we will continue to monitor this going forward because there's just we could do a four hour podcast by dropping a 28 minute answer on you. No. It was really good.

And I am curious, your solution to misinformation where you shift the burden to the posters does make a ton of sense to me. To be clear, I'm not asking people to go out and curate their feed. I think the way this would actually play out is Twitter would have a recommended algorithm that they would produce, right? And there would be like, you could buy algorithms or something along those lines. So you would still be outsourcing this to other folks, which, by the way, is how politics works. There's no single person, no matter how well informed they are, that actually knows every aspect of XYZ. Like political parties, as much as people detest them, exist for a reason because to function, we outsource a lot of our, like, it goes back to values and you outsource sort of like policy implementations to whichever side you agree, you know, in general reflects your values or detest, which you detest their values, right?

And so that's likely how this would play out. And I'm not saying this would necessarily be a panacea that's going to make things better. But at a minimum, I think it would at least help a lot of the legal issues here. Like I do think this idea of having a centralized entity making these moderation decisions is like having a corporation thinking they can stop all intruders by building a firewall. Like it just in the long run, it's not a sustainable approach. Right. And your proposal is a lot more practical. And my final question before we move on was going to be, how do you handle the algorithm aspect of all this as opposed to just the misinformation?

But it sounds like you just answered it. Like you're going to be able to choose your own algorithm and there will be explanations for what you're choosing. Because I just feel like algorithms play a role in all of this and in shaping what social media is. And empowering people to choose what they're seeing and what they're getting makes a touch. ton of sense to me. But correct me if I'm wrong. Yeah, I mean, this could be a little, that would be, I think, more difficult to implement in some regards.

And I have personally less of an issue, even if it's very powerful with like the algorithmic bits than the like moderation bits. I mean, one of the whole problems with the moderation question is it's a very Motten Bailey sort of thing where you choose the example to support your thing that's very extreme, you know, like if you're, if you're anti too much moderation, you choose people who are right about aspects of COVID and you're like, look, this is going too far, right? If you're pro moderation, you choose the most vilest people on the, on Twitter and say, look, you want to allow this stuff to exist? And like, you know, it's, it's productive to realize what an intractable

and difficult problem it is. And it's so difficult and tractable that I think there is some value in companies just sort of washing their hands of it and say, look, here's our suggested thing. Like, like, this is how email works, right? Now, email's an open protocol, so it's more viable. But if you sign up for Google, you know you're getting the Gmail sort of protocol, right? The challenge for Twitter, and that's a reason, by the way, lots of people like Gmail, because they think their spam filtering is really good, right? And so, so that in broad strokes, I think that makes more sense.

From an algorithmic perspective, I mean, it's tough. Like as long as stuff's out there, like there's lots of ways to go around the algorithm. I mean, in general, I get a little. Well, look, you can choose right now. You can say I want a chronological feed. Instagram. I'm sure there are controls that exist. I'm just wondering, like, to the extent this case shines a light on the problems with algorithms,

how do we account for YouTube promoting harmful shit to people who have, you know, their own to people who are potentially prone to embracing some of that extremist content. Yeah, I'm kind of thinking out loud here. So maybe I'll instruct this. But maybe this is a case where liability is actually a good solution because your concern with liability is that companies overcorrect, right? Like they self-police. But if they're self-policing as far as the algorithm is concerned, as long as there's

a way around the algorithm, then I think it's less of a big. deal, right? Like, this is where I, like, I get very annoyed at people who say, no, actually, content on the internet's very expensive. You have to like, you know, because you have to pay for Google ads. Like, well, no, people can still go to your site. They can type the URL, their address bar. They can download your app. Like the, I mean, if Apple allows it, which gets that why I'm more worried about Apple. I'm much more worried about stuff being erased and eliminated just because there's no, there's no alternative. Now, that's not quite right with Twitter, because you can start your own website, but I am completely 100% opposed to ISPs, for example, or at service providers

limiting what you can access because there's no, there's no alternative, there's no way around it. I'm less concerned about regulating algorithms because you can, there's ways around the algorithm or I'm less concerned about companies being very active with their algorithms. And I appreciate that an algorithm is essential to an ad-driven business. And I appreciate, that these free services exist generating tremendous positive consumer welfare that's very easy to take for granted and that that's how the model works.

So on both sides, I'm less worried about algorithms in general and I'm less worried about regulating them than I am about sort of moderation. That's sort of why I focused on that point. Yeah, well, and it sounded like you would be open to liability if there are extreme cases. Yeah, it's one of those things where in theory,

I'm in favor of it. I'm very worried pragmatically about how, would play out. It would be very difficult to enforce. Well, 230 made the rules of the road very clear. And like that's a phenomenal way to generate a whole lot of innovation is when you know what the rules are. And the concern with opening the door to more and more lawsuits is the, it's not just

that it's not the liability issue. It's the uncertainty issue. That's sort of a, that's something I would be concerned about. Yep. Well, we'll see what happens over the next 30 to 40 years. Yeah, so it's going to be fun. It's a failure of a Belvig episode. Keep it rolling.

Another question about algorithms, actually, from Graham. You guys talked at length about how AI requires meta and alphabet to step up their investments in GPUs and the like. The price tag is now in the tens of billions. So how come we don't hear the same about TikTok slash bite dance? Is there something special or different about its algorithm? Or are we just in the dark because the company doesn't disclose? any information. Ben, what's your answer? Yeah, I think we're a bit in the dark.

Also, it's not like, like, they're not necessarily doing magic here. Like, one of their big breakthroughs was really focusing on, like, implicit action. Like, the reason why TikTok collects so much data in part is that's how it works. Like, if you linger on a video, you don't have to, like, you don't necessarily have to hit like or whatever. Just like the fact you watch the video to completion or you watch the video twice is an indicator that you should get something sort of similar. Now, analyzing all the videos out there, I'm not completely clear to what extent they're doing that.

To what extent they're relying on other signals from users that they determine to be similar to you. I think they're kind of more advanced than what Facebook was doing, but less advanced than where meta wants to go. And so I think there's an aspect where meta just wants to be even better at this and even more customized and tailored. So that's number one. Number two, what I regret about not stating when I was talking about all this GPU stuff, the real payoff for the AI stuff is ads. just to be super, super, super clear. Ads need to be so much more targeted. It's one thing, like, there's no cost to showing you the wrong video.

There is a potential cost in engagement where you find it less compelling. But if you're kind of broadly coast, it's okay. When you're trying to sell an ad, you're trying to really pick a needle out of a haystack. Find the right user at the right time with the right product. Because you're asking them to not spend a couple seconds. You're asking them to spend a few dollars. And that's a much greater barrier. and the necessity of precision is much, much higher that if you want to sort of make money.

And so that's where the investment in AI is going to truly pay off. People are like, oh, like these sparse models are almost as good as the real models. It's like almost as good when it comes to measuring ROI doesn't cut it. And like, and so I should have emphasized that the reason why the capital spent for a meta makes the most sense is this is the way around ATT. It's getting super good at targeting in a probabilistic way. Sorry, go ahead. Don't feel guilty. You did emphasize that.

And I came away thinking, okay, so this is the only card these companies have to play in a post-ATT landscape. And it's going to take a couple years maybe, maybe more. But this is how they'll ultimately respond. Well, thank you. I'm glad that I did communicate that point. And it's worth noting that Apple has played its card, right? Like there's nothing they can do about Facebook just getting that.

It's almost like an evolutionary sort of system, right? Where Apple has forced a selection pressure onto the ad ecosystem where you can no longer just do the nuts and bolts of deterministic targeting. You have to actually develop this huge capability. And it's like sort of like if you like take insufficient penicillin, like the bacteria that survive are like superbugs, right? that's that's going to be a situation here where you're going to have companies that may not make it or really struggle like a snap or a Pinterest or whatever might be but the ones that do survive are going to be sort of untouchable yep well we'll see how long it takes for them to get there well that's the other thing about the whole meta thing is they're still they still took it 27 billion dollars last quarter right like people are talking like this is a company going to zero it's it's crazy i i mean it's a great example of how sentiment out runs sort of fundamentals, at least in the short to medium run. Well, and it's also a familiar pattern with them, right? I mean, people have been rooting for their downfall.

No, no, you're right. You're right. No, you're right. It's happened four or five times. Like, people have been sure that Facebook is going out of business literally since it started. So it is part of the pattern. But there's a bit here where I think a lot of people that pushed back on that kind of threw in the towel on this one.

So there's like this like this cascade. Well, it's, the issue is they all watched Meta Connect and came away from that three-hour video. It was the most successful rebranding ever. Like the rebranding works so well that people have now associated meta's business prospects with Legless avatars. Oh, I know. All these defenders are like, you're on your own on this one, Mark. But for now, let's talk Twitter. PD says Twitter's key use case is publicly visible intellectual validation via verified slash trusted

celebrities who you actually have a real shot of engaging with. No other platform offers such a high degree of engagement between quote-unquote regular folk and quote-unquote celebrities. And as a result, Twitter will continue to thrive as long as it remains the place where the likelihood of interacting with thought leaders is the highest, barring perhaps an invite to Davos or the Metgala. In this light, Twitter Blues paid service is essentially akin to buying a backstage pass at a concert, which, while not guaranteeing anything, gives you a real shot and hope of connecting with a celebrity slash thought leader you really want to engage with. Totally worth it in my opinion. So Ben Thompson, celebrity slash thought leader with a blue check

next to your name. What do you think of this? I enjoyed this framing of Twitter blue. It's more compelling to me than the messaging that's been coming out of Twitter. over the last seven to ten days. Like, I mean, low, speak of low bars. I think you dropped a couple of low bars there. One,

delivering me a celebrity and two, talking about being clearer than the messaging coming out of Twitter. I know, I think this is, I think this is right and valid. And I think I, you know, to the extent that Twitter does really elevate the visibility of if you're

sort of Twitter blue or not, I think that will pay off. What I would push back on is that this being Twitter's key use, use case. I think Twitter's key use case is the rapid assimilation of lots of content, like the data fire hose sort of bit, but on sort of like an individual level. Yes, there's the fire hose that company sort of ingests the whole thing, but also just it's so efficient to get a huge amount of information on Twitter. Now,

is that information valuable? Is that innovation just chum for partisan waters? Well, that's a different question. But there's no question five minutes on Twitter is going to get you so much more information than five minutes on any other social network by a sort of massive amount. And in general, we don't like the number of people consuming and never posting is as a rule always way higher. And so when it comes to if just from a sheer number perspective, I think Twitter's key use case is getting information, which two things. Number one, that is why in theory an advertising model makes sense because you want to monetize that big chunk of users who are just consuming information. But also number two, it's why I make the case that if you want to really do a subscription,

I say make everyone pay because you actually are delivering so much value to those people that just consume that they will, despite the state of preference, would definitely be no, be more willing to pay than you would think. Yeah. Well, I do think that's important to account for. Everyone in my life, I'm one of the rare people in media who still has way more friends outside of media than in the media. and all my friends outside of media have Twitter accounts

and check Twitter on a near daily basis. Maybe not all of them are checking on a daily basis, but they check it frequently and just have it posted since 2011. And I think a significant chunk of Twitter users fall into that category. So finding a way to monetize those people does make sense. Now, whether you could really do that at scale, I have no idea.

You can say no way because everyone thinks I'm insane here. But that's the, it is the core of my argument. Well, and if you ask me to pay, I'm, again, I'm King Normie. So I'm a pretty good barometer. I would pay $3 a month. I wouldn't go much higher than that because I already have a love, hate relationship with Twitter as we've discussed. Sebastian asks, what does Ben think about Mastodon as a Twitter replacement?

and does he think this could somehow bring us closer to a decentralized social network? So what do you think, Ben? Are you moving to Mastodon? I feel like that's sort of the Twitter equivalent of I'm moving to Canada now that Elon Musk is president. Well, the better, the tech-related example is I'm going to start running Linux on the desktop, which is a long-running joke. This goes back to like the 90s. I remember when I first installed Linux,

you know, I think it was in college, and, you know, somewhere around then. And, you know, wait 90s. And it's like, well, all they need to do is just figure out a couple things and make it a little easier to use and solve these driver issues. And people will be using Linux. They don't have to pay for Windows and it's more open and XYZ. And the reason it's a joke is because it never happened.

And now everyone has realized it is never going to happen. There is a serial undervaluing of convenience and, like, ease of use. and convenience and ease of use does tend to go hand in hand with centralization. If you have centralized stuff, you can control it much more effectively to make it much more accessible and easy to use. And Mastodon is a great example. I have to know, like, which server do I log into? How do I do this?

How do I federate this? Oh, what? I have to trust this server operator with all my private messages? Like, that seems insane. Like, there's all these sorts of bits about it that, uh, just from a pure user experience perspective, I'm exceptionally skeptical and I'd say like 99.999% percent. sure that matchdown will never become a thing, number one. Number two is...

I'm 100% just for the record. I'll put 100 on it. Number two is this bit about what makes Twitter unique is the fact that all the people that hate each other are on the same platform. And again, this is a thing where there's like a stated preference versus revealed preference. People want to be in the same place as everyone,

but no one would ever like voluntarily do that. Like there's a reason Twitter is everyone says, I hate Twitter and also I can't stop looking at it, right? Like, like, no one, into that willingly. You sort of back into that organically, and that's the part of Twitter that I think can never be replaced. Well, and to put a finer point on it, you alluded to it earlier.

A lot of the appeal is people you agree with posting screenshots or different tweets from people you despise and totally disagree with it. Like, that's Twitter's number one use case on a daily basis. It doesn't hit the same when you're posting a picture from truth social versus when you're like quote cheating a tweet. Like it just the different level It's like hard drugs versus versus weed or something like that

Yeah well I took one look at Mastodon And I was like I will see myself out Good luck to everybody who wants to migrate over here But not for me Of course theoretically I'm very much in favor of decentralized Concepts right Like I think it's I just spent a whole time

Talking about the problems of centralization And the issues that it presents I just, I, I, I tend to think that decentralization ends up working best when it's sort of behind this, like Linux, right? When it's, I think it's fantastic that all the world server infrastructure by and large is run on Linux, right? Even Windows Azure, Azure is mostly all Linux. Like, there's very, very few like sort of Windows servers things. And that's great. Like, and I think at the infrastructure level, it's really important to be open.

And that's why, for example, like, I'm on WordPress, right? WordPress is, is it, it's clunkier. It's harder to use. There's five gazillion themes and plugins and all this sort of bits that everyone complains about. They're like, oh, go to XYZ service. It's so much easier. That's true. But given that this is my livelihood and given that I want the freedom to sort of do what I want,

I like the fact that my software that runs my site is open source. And it can sort of be, I can, sort of be. I can pick it up and go. Now, I recognize that comes with certain costs and burdens, and it's not necessarily available to anyone, but it is available to anyone if you can pay someone to help you do it. And I'm not saying that's great. Obviously, it'd be ideal if anyone could do that. But anyone can, if you take the time to learn, right? Like, you can manage.

I ran my own site from grounds, nuts and bolts, like at the beginning. I wasn't very good at it, right? But there's super, it's important as value. I just think because of the inherent challenges from a user experience perspective, that decentralization and openness works best at an infrastructure level. And it's just, it's very hard, if not impossible, at a consumer level. Right. To get mass adoption from regular people. Another question we got, Justin asks, Twitter Blue will almost certainly put in-app purchases in Musk's purview. Do you think Musk joining Epic slash Spotify could be the change agent needed to shift public or government sentiment against Apple? This is fun because I feel like Justin is totally misreading public and government sentiment toward Elon Musk.

I think Elon Musk taking up this cause could very well make people more sympathetic to Apple, but maybe I'm just reacting to what I'm seeing in my Twitter feed. What do you think? is pretty interesting. I think one of the things you see, one of the challenges that Elon Musk is facing with Twitter right now is that because Twitter never built out a real direct response advertising service that worked at scale that was self-serve like Google or Facebook did,

like there was a big push by advertisers to cancel Facebook in like 2020. And it had zero effect on Facebook because it turns out Facebook's bread and butter are all these small and medium-sized businesses that are self-serve and doing their ads and it didn't matter if you would have pulled their ads, right? It just didn't really affect it. Twitter never built that kind of business. So Twitter is remarkably susceptible to centralized advertiser decision making.

And so that's sort of a broader issue there. Now, how does this play out? Number one, I want to get that takeoff. So I'm inserting it in here. Number two, I think there's a bit where it is hard with the must thing in general. when the media sort of like catalyzes into group think and like a particular point of view, it's really hard to know what's actually happening broadly speaking. I mean, look no further than our elections, right?

Like we saw this in 2016 where it was like a lightning bolt. Like it didn't even occur to anyone that Trump could win. But it turned out like the voting was a accumulation of individual decisions. Again, I know the electoral college stuff and all that sort of bit. I'm just saying when individuals are making choices at scale, you sometimes get different outcomes than when there's centralized entities sort of acting. So in this case, would Musk versus Apple, on one hand, I think it's a good point by you and you very well may be right that this will actually help Apple because, you know,

Musk is as acquired this tape. Well, and look, it could be neutral. Musk could just stay out of this fight and not become the public face of the uprising against Apple 30%. No, I think he's already, he's already, he's already tweeted about. about this. Like maybe not recently a few times. But Apple's power is not because the media loves Apple. Apple's power is common consumers love Apple. Like they have power because individuals on the like love Apple and they trust Apple and they believe it. And so it would be interesting

to have someone other than a centralized entity pushing back against Apple, someone with popular appeal pushing back. Now, I think you're probably right, but I'm not sure. It could play out differently. Yeah, well, I would also add that another aspect of Apple's power that has been important is there just hasn't been much scrutiny into the way they're operating their business and there probably will be going forward. And so there are going to be new challenges for them to confront regardless of how much we all love Apple's consumer goods. I mean, I'm an iPad evangelist as of two years ago. And I love my new iPhone, my new MacBook Air, but I think there's basic things they're doing with their business that are objectionable. And we'll see when the government decides to intervene on

that front. We actually got two questions. That's right. I'm going to take charge here because these are not directed at me. They are directed at you. And oh boy, you love this. I do. I like putting you on the spot. So two responses to last week social media. Number one from PD, I'm squarely in the net positive camp about social media purely due to the number of amazing things I found online, including those via sponsored ads. Every boutique shoe and fashion brand, he didn't give details about his body shape, but as we noted, I can relate. You identify it.

That's right. Every Ben Thompson, Johnny Harris, or niche news publication, every boutique hotel, amazing bar, food delivery startup or artisanal coffee shop, et cetera, that I love today primarily found me on social media. Additionally, just within my family itself, there are about four to five businesses that can attribute their success to visibility and customer reach via Instagram. PD wants to know, Andrew, why do you want to kill small business that makes your life better? Oh, man. Well, this is the thing. This is the argument that Ben Thompson's of the world always trot out.

Like, if you are opposed to social media, you're really opposed to democratizing commerce. And that's 100% true, honestly. Like, when you look at what Facebook's done and the net positive that they can claim or the most, the most, meaningful benefits that they can claim. I do think a huge part of it is that they have allowed small businesses to find a way to thrive, even as corporate power is unprecedented and there's concentration in dozens of different industries. I think it is cool that there are a bunch of businesses that can thrive on Instagram. And my wife has an Etsy shop that relies on targeted advertising as well. So I'm not opposed to PD or artisanal chocolate shops, a large community event space, a boutique homestay in the Himalayas. He listed some of the businesses. I'm not opposed

to any of that. I would still say there are a lot of negative externalities that we've experienced as a result of social media. I've given the argument to PD and Alice because they agree with me. I'll go to the next question from Peter. Hi, Andrew. I really do enjoy the podcast. So I apologize. I mean, you always know a banger's coming when they have to like, it's like the shit sandwich, right? He has to butter you up slightly first and apologize before he drops the take.

This was the nicest angry email I've ever received. So a sincere thank you. Sports, right? Like sports and they just go in full guns blazing. Exactly. In sports, it's like you're an effing. idiot, blah, blah, blah, Peter, thank you so much,

sincerely for being such a good guy. I really do think the show is great. I love your role in the podcast. Just really nice stuff from Peter, but now, now to get to the good part. I was pretty shocked by the comment you made about TikTok, where you said that you did not care

that it would get shut down, as there was little cultural value to it. I think this would be the equivalent of me saying that there's no value in sports because I don't care for it. Excellent point. Peter, you're going to have a hard time recover from that one, Andrew. Continuing on, in reality,

millions of people enjoy sports every day, and I don't have to personally see the point of it to trust that all of those millions of people get something out of it. For TikTok, in particular, I could list 25 unbelievably good creators that I know and love that completely built themselves on TikTok. They come from all sorts of countries, backgrounds, and create content that nobody would have thought would be interesting or engaging,

but it really is. They connect people across culture, societies, and ages, which isn't to say there are no problems, but it makes the world a much richer place. Platforms that democratize creation and then select the best to share with the audience are incredibly empowering, and it is extremely frustrating

that people who haven't engaged with it at all are so dismissive of it. Am I owned, Ben? You're the judge, jury, and executioner here. Am I owned? You should not grab me that privilege because I completely agree with him. I mean, again, as made the point before,

Like, when I started Shatacri, I actually went to several news publications because I so much wanted to do this job. Like, of course, I had to paint myself with some great visionary billion subscription model XYZ, which was the goal. I was so burnt out from trying to work a real job in build Shetri. I'm like, look, just hire me. And my salary requests were very, very low. I just wanted to do this work. And none of them would give me a job. And this is going, it's the same thing as like the whole Google thing, chip on the shoulder thing, right?

But you know what? It didn't matter because of the internet. I could hang out my shingle and through hard work and having stuff that people found compelling. And thanks to a boost for sure to social media, because that's where stuff was shared. Like, that's why number one, Chatechry's been successful. And number two, why you are sitting here today. So I'll just sit back to find a point on it. Let me ask you something.

Do you think Stratory would have succeeded if social media did not exist? No. not at all I mean Chatechere there's maybe a path where it would have taken years and years to build up

but what's so powerful about social media is not that it gives the creator a megaphone but it gives all their fans a megaphone and so they get to go out and broadcast wow there's this great new site

out there you should go check it out and it's the same it's like the moderation thing flipped on its head it's like when you empower lots of people to do something, you get all the benefits of internet scale by being one person. Like, it's incredible. It's incredible.

And so when I started out, people would be on Twitter and you had this bit where, you know, the status bit of Twitter, it's like, oh, I found this site you've probably never heard of. But this is a really great article. And you get this reflection on it. And strategically, I actually went paid within nine months or, sorry, 11 months of starting it, which was, I had a five-year plan. I pulled it off in 11 months.

Now, it got a little hairy. but like 100% because of social media for sure. I think social media may have been an accelerant, but you're also discounting the blog culture that existed before social media. I know. Everyone told me that blogs were dead. Well, listen, I'm just saying if Twitter had never existed, a lot of that sharing and discovery was happening on blog spot and WordPress and Tumblr or whatever.

But all of this is to say, totally valid counterpoints from Peter and PD. And Ben. And Ben, maybe. I'm not advocating for the elimination of the internet writ large, okay? Stratere is safe. Sharp Tech is safe. I just have a problem with these centralized platforms,

which as there's more and more adoption, I think we've seen with Facebook and Twitter. And to my credit, if I can give myself credit, I- I shifted the social media discussion away from TikTok, which I'm not on to Twitter and Facebook, which I've experienced firsthand over the last 15 years, how those platforms have changed. But I think Peter does have a point here, right? Like, again, I have issues with TikTok, particularly its ownership and like, as we've discussed, right? The Chinese angle on it.

But I do think it's right and fair to say, look, there's incredible stuff there because. there is. And not just that, but it's incredible stuff from people that would have never gotten a chance otherwise. And that seems like a pretty great thing. No one was hiring like people who can put together a 15 second skit or, or invent a dance, you know, absolutely. But look, if there are benefits to some that shouldn't dissuade us from considering whether there are costs to more. But it's not just benefits to them. It's benefits to everyone who enjoys them, right? Yeah. Look, and I don't want to, I was not making the argument that TikTok has no benefits. And when I said there would be very little cultural cost to just banning TikTok tomorrow and we never think about it again, I should have been a little bit clearer about what I meant. You and I have talked to the past. The era of social media, it's ushered in an era of abundance. But the tradeoff is that virtually everything on social media lacks any real permanent.

permanence. Yeah, it's disposable. And so that's the bucket that I put TikTok in. It's like, this is not the sort of culture that's going to endure for generations to come. And that's fine. Like, not everything has to last for 100 years, but that's what I meant. If TikTok ended tomorrow, it wouldn't be like we were losing the library of Alexandria or something. Like, we would be okay. And I'm basing my opinion on like 200 TikTok videos. I've watched over time and random controversies, like the West Elm Caleb thing. Did that ever make it to Stratory, West Elm Caleb? Do you even know what that is?

Yeah, that was the guy who was like dating multiple people at the same time and then... Yeah, and some of the most annoying people on the planet tried to cancel him. It was pretty remarkable. It was not a great advertising for TikTok. But I do know that other people use it differently and find it incredibly rewarding and empowering. And so it's no shot. at any of those people. I would say just in your defense.

Number one, there is the obvious sort of take here that look, like the havoc wreaked by giving everybody the ability. It's very easy to see like the first order effects of everyone being able to publish. There are second order and third order effects that it's fair to argue that our, the fundamental structures of society are unprepared to sort of handle this. And again, this is where I always make the analogy of the printing press where, oh, making it easier to print stuff sounds great. Second order, third order, fourth order effects, like a little more problematic, right?

That's number one. Number two, I want to credit you because I think your bit about permanence versus disposability is a good one. I mean, I think it's fair to wonder. I worry about this was just trajectory, right? Like there is like the great business books of the past that are still sort of reference, like Porter's Five Forces or, you know, Clayton's sort of disruption, you know, innovator's dilemma. And it's like there's a sadness to a certain respect that writing a blog is

actually way more profitable than writing a book. It's something I've written about. But there is an impermanence to that, that's sort of part and parcel of the deal. And I sort of have to be, I sort of have to be okay with that. Maybe I will write a book today just because there's, so there's something out there, right? Yeah. So I think that's a really great point from your point. Number three, I do want to say that even in this case where I do on level, disagree with you. And I think that it's so easy to quickly acclimate yourself to the good sides of things and become fixated on the bad things. That's just your your balance is maybe off. Right. It's right and good that you're pushing back on this because I am, thanks in part to my

job and my fundamental outlook, heavily biased to be in favor of this stuff. And I think it's fair and right to push back, particularly as we see the world that we're in. So, so I will I will come to your defense, stand in front of you as Peter is charging. The Ben Thompson's of the world, ending on a magnanimous note, I very much appreciate it. And I appreciate the feedback. We got others that I wasn't able to read here. But the bottom line is, I think the conversation is really interesting when you frame it in, you know, absolute terms, as difficult as that be. It's like the moderation thing, right?

Each side can point to examples that perfectly make their point. right? Peter can list these amazing creators that would never be discovered. It's like, yeah, that's a good point. The other side I can point to like terrible misinformation or the cultural division society and say like, this is bad. It's like, yeah, you're all right, right? Like that's kind of the challenge here.

Yeah, well, and there's no question that advertising, it has been very beneficial to small businesses. But if the, the best argument for something like Facebook is it's like a great advertising tool. I'm not sure that's persuading me. And at least it's not persuading me to just discount some of the costs. But in any event, I really did enjoy the feedback from everybody. And Ben, when you go on book leave, all I ask is that we continue recording Sharp Tech throughout your writing time. Theoretical writing time. The truth is I'm terrified of not having a daily update. That's the only way anything gets done. Well, let me tell you something. My friend and your friend, Ben Goliver,

wrote a book. And it did not sound like a fun process. You seem to have a pretty good rhythm with Tretaquery. So I wouldn't upset the apple cart for the masochism that is the book writing process. Yeah, no, I'm going to wait until I retire and write a very indulgent sort of biography, or autobiography. That sounds a much more fun and interesting because then I just, you know, who cares me when buys it? I can just, say whatever I watch. There you go. Well,

speaking of saying whatever we want and shooting from the hip, we will come back later this week with an episode for subscribers only. We'll also hit some mail that we weren't able to get to today.

We've got some good antitrust questions. Yeah, I think we might have to do a proper real Melvang episode, hit a bunch of these at the end of the week. Exactly. But for now, Ben,

I look forward to talking to you later in the week. and congratulations on the 9-0 Milwaukee Bucks, first place in the entire NBA. Yeah. Well, thank you. I appreciate you acknowledging that.