Apple 与平台的权力、AI 进击 Google Search、ChatGPT 进击高等教育

Hello, and welcome back to another episode of Sharp Tech. I'm Andrew Sharp, and on the other line, Ben Thompson. Ben, how are you doing?

I'm doing well. Better than you.

Hmm.

I feel like you blew all your being-right points picking the Lakers to end April.

You can't win them all.

You've been wrong about everything ever since. But, you know, so it goes.

You know what? It was worth it. If it meant that I had to forsake all my pick karma for the remainder of the playoffs, it was worth it because that's how satisfying the Lakers' loss was in round one to my Timberwolves. Really rough outing for the Wolves earlier this week against Steph Curry in Game 1. Honestly, I'm bummed. Monday night, we had both the Celtics and the Thunder taking it on the chin as massive favorites, and it was just so much better than tonight, when the Celtics are rolling as we come on to record. I'm sure the Thunder are going to kick Denver's ass later tonight.

I do have an announcement, and apologies to everyone who's not into sports, but I have decided that I am cheering for Denver to win the title.

Hmm.

The reason is I'm psychologically tortured by the fact that everyone universally acclaims Jokic as the best player in the NBA, despite the fact that there's absolutely nothing in his résumé that is better than Giannis's. They should both have the equivalent number of MVPs. Giannis does have a Defensive Player of the Year, which Jokic doesn't play defense—

Good luck—

—in my estimation.

—maybe one day.

And you know what? Let him have another title just for my own peace of mind. Then there will be no questioning or disputing it, and I'll be happy to get on board. That way, I can stop raging against the world defending Giannis.

I mean, it sounds like all the followers of @notechben should be rooting for Jokic as well to get this done. Let this particular crusade just pass you by, you know?

So it goes. Let it be. Whatever phrase is appropriate in this circumstance. Enough basketball, though. Let's move on.

Indeed. Enough basketball, enough with the reality check night around the NBA. We're going to bounce around on this episode, and we will start with a bigger-picture piece you wrote on Apple and the App Store and the nature of platforms. As a preliminary question, I want to focus on one sentence from that piece. You wrote, “Before I started at Stratechery, I worked at Microsoft recruiting developers for the Windows App Store.” Can you give me any color from that era of your career? What year was it? What sorts of developers were you pitching? What became of the Windows App Store?

1. Windows Store Lessons

Well, there are actually a few undocumented aspects of this—once-documented aspects of this. I am not one to delete tweets. By and large, I think that what I say is out there, for better or worse, even though I've always argued that Twitter should delete tweets by default.

Mm-hmm.

I think Twitter would be a much better platform if it just auto-deleted everything after, say, a week.

I fully support deleting tweets. There are a lot of people who look at it as an act of cowardice. I think it's an act of sanity. It makes sense because, divorced from context, there are all sorts of tweets that can come back to bite you. We're all on Twitter trying to have a good time. It doesn't need to be anything more than that.

No. The only reason to go back and dig through people's tweets, by and large, is to try to do bad things to them.

Hurt people.

Yeah, exactly.

Exactly.

I am in favor of deleting tweets also, but I don't delete my own tweets most of the time. Back when I was in business school and looking for a job, I ultimately ended up at Microsoft, and I decided discretion was the better part of valor here. I had a lot of tweets extremely critical of Windows 8 when it was first announced, and the entire concept of it—this idea that Microsoft was going to leverage and completely transform the Windows experience into a touch-first experience, and it wasn't going to work. They were too late. All these sorts of pieces, which, in some respects, I wish I had held onto because they were totally correct.

Hmm.

Windows 10 was the apology tour for Windows. It still nominally had the additional blocky Start screen and stuff, but you could just get rid of it and use Windows as normal. It undid a lot of the things with Windows 8. But all the same, I needed a job, so I went to Microsoft.

I was going to say, so you're deleting this in the middle of the interview process? Is that right?

I don't remember exactly when I wiped it all out, but I thought, “Maybe better to not have it out there.” For the record, I went in there and gave it the good college try, and it was a ton of fun. It actually worked out really well for me personally because I joined a team that was nominally focused on OneDrive or something like that. But within a month, our team got switched to working on the Windows App Store.

Actually, Windows 8 might have happened after I already had the offer letter. I hadn't joined yet. Anyhow, I think I actually exercised discretion within about 2 hours of posting my tweets.

Mm-hmm.

Regardless, our team got switched to doing the Windows Store and developer relations. Our team dealt directly with the biggest developers, and then with enabling what's known as the scale motion. Microsoft has this huge team of people all over the world who interact with developers. I've been writing about the last couple of Microsoft earnings calls because one of the reasons they missed on earnings, or slightly missed, a quarter ago was that their scale motion was messed up.

In this case, their scale motion was too focused on AI. Basically, they had a lot of their field team pushing small and medium-sized businesses to adopt AI, and the small and medium-sized businesses were saying, “We're not ready to do that.”

We're not there.

And so they were saying, “No, we don't...” They had to get them back to just doing normal things, like getting them on the cloud basics.

I created the presentation. They had a presentation for the scale motion that was terrible. One time on a plane, I thought, “This is such a bad presentation. I hate using it. I can't imagine what I was trying to sell when I was in on this.” So I rebuilt the whole thing from scratch.

Mm-hmm.

It ended up being used all over the world by thousands of people.

That's awesome.

It was awesome. Our team stepped in, and I actually ended up with a role that was like a Level 64 role. This is Microsoft nomenclature, even though I was only Level 61 or whatever. I had way more responsibility and was doing way more things than I would have normally. It worked out very well for me. It was a ton of fun, and I worked super hard.

I got the Kindle team on board. I got Automattic on board with the WordPress app. They were in the initial announcement of the App Store. Those were 2 of the featured apps. I worked with a lot of the publishing industry. I learned a lot about publishing and the challenges and issues that they had. Lots of trips to New York. It was all fun, and then Windows 8 actually launched, and it wasn't very much fun anymore.

And so when you say not much fun anymore, are you still interfacing with those companies and having to explain, “Don't worry, one day the users will get here”?

Right. You go out there and say, “Windows has a billion users. Everyone's going to use this.” You're selling this story that this is—get in ahead of the curve. Remember, don't you wish you'd been ahead of the curve with the iPhone or with Android or whatever it might be?

The reality was there was no curve to get ahead of. What are you ahead of? You're ahead and you're stuck in the desert, right? As part of this, some of the companies built the apps themselves. Some of them were like, “Fine, we'll do it,” but there was definitely pain going on. We worked with third-party developers, these different shops that would actually build the apps for them.

It was a very useful experience. I learned how to manage software projects and understand when you're being bullshitted, when there are actual legitimate problems, like working with a still-in-development API that was not stable. It was a phenomenal learning experience. The learning experience was more about managing dashed expectations—

Which was, again, not nearly as fun.

Well, it seems like it would be good experience to get windows into how all the—pardon the pun with Windows—but to be given a look at how all those other companies are doing business and what they want out of a platform. Does any of that resonate?

For sure. It gave me an in-depth understanding of how everyone is thinking about the space, and a lot of the specifics that go into it. What do companies actually think through when it comes to launching an app or a new service?

Mm-hmm.

It’s not just about building the app; it’s the maintenance. Even if you have a small number of people using it, if you cancel it or deliver a bad experience, that small number of people can be very vocal and can tank your reputation broadly.

Mm.

That’s a big risk.

Yeah.

It’s why you would want to build an app even if someone else is paying for it. But the overall question that I wanted to mention in passing, which I think is important, is often framed as a chicken-and-egg question. You need apps to get users because users want to use the apps, but you need users to get the apps because developers want to serve users.

My contention—a point I made 9 years ago, I think, in the context of the iPad, Apple, and Microsoft, so it’s not a fresh point—is that it was worth mentioning again. Sometimes I’ve made a point somewhere in the past. I have a somewhat encyclopedic memory of what I’ve written. Most people don’t.

Mm-hmm.

I thought this case was a good opportunity to mention again that the chicken-and-egg fallacy is a fallacy in this case. It is a fallacy; it’s not a conundrum. The reality is that users come first. And this is—

Always.

—obviously in line with Stratechery in general: my whole point is that demand is what matters, not supply. But the other point I wanted to emphasize with this article is that, just like the chicken-and-egg thing is ultimately a dishonest—“dishonest” isn’t the right word—but an incorrect description of the dynamics at play here, there’s been a lot of dishonesty, incorrect depiction, and wishful thinking about App Store dynamics all along. Some of it came from me, particularly in the early years.

Yeah.

I backed off this a long time ago because what I wrote in this article has become clearer and clearer to me. But I thought it was important to make explicitly clear that the reality of the App Store is that this is just a fight about money.

Mm-hmm.

Because the App Store is so important, and the iPhone is such a big platform with so many users who are worth so much money, there’s basically nothing Apple can do that will stop developers from building for it. At the end of the day, it is a sympathetic frame for Apple’s positioning: you’re just asking Apple to be generous. All the cautions—“Apple, you’re going to have developers flee your platform”—no, they aren’t. That’s not going to happen.

Mm-hmm.

The only valid critique for Apple all along, which has stood the test of time and was validated in this case, is that you’re inviting government regulation—

Totally.

—or a judicial decision.

Yeah.

That’s always been the only hang-up, and I think it’s worth being crystal clear about that.

I didn’t think it was all that sympathetic. I thought it was more just a realistic, clear-eyed look at Apple’s motivations and Apple’s incentive structure in the midst of all this.

I do like puncturing tech myths, though. I always love talking about how everyone says, “Windows beat the Mac because it disrupted the integrated product with a modular player.” And I’m like, “No, that’s actually completely incorrect.”

Mm-hmm.

Windows won because it built on DOS, which was distributed by IBM. DOS had won before the Mac ever came along. That summary of Windows versus Mac leads to really bad analysis from people I really like and respect. This is a Clayton Christensen reference, who, in my estimation, completely misunderstood the computing market precisely because he misunderstood the Windows-versus-Mac dynamic. He misunderstood it because he just repeated what was totally accepted wisdom.

The totally accepted wisdom is totally wrong and led to bad analysis, and I think there’s a similar thing here. Everyone is all about developers: you need developers to build platforms. The answer is no. You need a good product that collects billions of users, and then you get a platform sort of by default. That’s how it actually works.

2. Platforms Versus Aggregators

So the headline of your piece was “Platforms Are Underrated.” Can you define for any normies in the audience what distinguishes a platform from an aggregator?

A platform, in the specific case that I’m using it here, is like an operating system. It’s something that has APIs that third-party developers build on top of.

Hmm.

If you’re Photoshop, you depend on macOS existing. You depend on Windows existing.

Windows, yeah.

You depend on iPadOS existing. It’s a foundation on which things can be built. The conundrum here is that platforms are one of the largest value-generating things ever.

Mm-hmm.

Basically, all of the value in technology, by and large, is built on top of platforms. They’re super important. We should want them to exist. They should have a lot of leeway, by and large, because their benefits are so significant.

But the reality is that this also gives them basically untrammeled power, and that’s why I didn’t mention it, but I put the Lord Acton quote at the beginning: “Power corrupts.” At the end of the day, all appeals to Apple about the App Store have been an appeal to Apple’s better angels, as it were. And the reality of—I’m going to sound like a Marxist here—shareholder capitalism in 2025 is—

I was going to say—

—that’s not going to mean anything.

The baby bird is growing up to be a proud anti-monopolist falcon here before our eyes. I love to see it.

Well, this has been my whole point about it. Aggregators, on the other hand, have no lock-in. You don’t have to use Google. You don’t have to use—

Click away.

That’s totally true. Google always says another search engine or another site is a click away.

Mm-hmm.

That is the fundamental difference between aggregators and platforms. Aggregators depend on users choosing to go there, but that is a choice users make that is not forced on them. Platforms are much more like railroads. Someone built it—

Infrastructure.

That’s right. They’re infrastructure. That’s why I’ve always been so critical of the anti-monopolists focusing on aggregators. I’m like, “Solve the platform problem.” That is an actual problem: they are making choices that are restricting innovation and restricting competition, and that is addressable and can be fixed.

Mm-hmm.

With aggregators, you’re trying to change the preferences of billions of people, which is impossible. There’s a bit about the whole anti-monopolist take against aggregators that is similar to the worst utopian visions. There’s a saying, right? If anyone says, “If people just did X…” If your political program entails people doing X, your political program is doomed. And if you get power, you’re going to be authoritarian because you’re going to get frustrated that people don’t do what you want, and you’re going to make them do it.

And make life worse for everybody, yeah.

That’s right. All the anti-monopolists focused on the aggregators are adjacent to this because they’re actually frustrated by the individual choices of individual consumers.

Platforms, on the other hand, are a point of leverage that you can address and that can fundamentally transform the economic value they provide. It’s like an environmentalist who was anti-nuclear for years and now they’re a clean-energy advocate, or whatever it might be. We can have a debate about nuclear costs and all those sorts of things, but I’m not going to take you seriously if you spent years objecting to fundamentally clean energy in terms of its impact on global warming, the atmosphere, or all those sorts of things. I don’t think you’re serious because you don’t actually think through trade-offs here.

It’s a similar thing here. If you’re worried about monopoly in tech, and you’re obsessed with Google and Facebook and have turned a blind eye to Apple and the App Store, it’s hard for me to take you seriously.

Yeah. Well, I mean, I think those people don’t really exist. I think everybody who’s concerned about Google and some of the aggregators is also concerned about platform abuse.

No. The priority order matters. You don't get to say, “Oh, I care about all of it.” What do you actually focus your time and energy on?

Yeah. Well, your article more generally was a great meditation on 15 years of this conversation. Your articulation of the App Store platform specifically—you wrote:

“Once there are users, there is demand for applications, and that is the only thing that incentivizes developers to build. Moreover, that incentive is so strong that it really doesn't matter how many obstacles need to be overcome to reach those users. That is why Apple's longstanding App Store policies, egregious though they may have been, ultimately did nothing to prevent the iPhone from having a full complement of apps and, by extension, did nothing to diminish the attractiveness of the iPhone to end users.”

So it ultimately didn't matter how good a salesman you could have been for the Windows App Store, because Apple's users were the ultimate—

I was a great salesman.

—moat for—

Yeah, exactly. And it didn't matter. You gave it the college try, a good-faith effort.

I thought in general it was just a great, sober look at the realities of Apple's market power, Apple's policies, and why those policies haven't changed over the last 15 years. I will say, for me, it was interesting to read your article as an outsider to tech, because the conclusions were all a little bit self-evident at this point. I missed the era when people were arguing that the App Store being locked down was going to really cost Apple's business. I missed the 7 years of that argument. Today, it's clear that it hasn't really affected Apple's—

Well, it—

—bottom line.

It's funny because this is adjacent to where I really cut my teeth early on with the App Store. And I mentioned the angle I discussed. I've said this before on the podcast, but the dominant sentiment was that Apple and the iPhone were screwed because everyone was going to flee to Android. My whole point all along was, no, they aren't.

Yep.

You don't—it's funny because we're both laughing at the possibility that this was an argument, but literally I spent 3 years on Stratechery arguing this point and built my entire—

Ben, you built a great business—

—business off of this.

It's worked out well for all of us.

I know. So it goes to show how stuff becomes super obvious over time. But that was another example of these myths and fallacies that are spun up that actually are not correct. And yes, the same thing with the App Store and Apple's power. It was absolutely assumed that they needed to do something. It's interesting because the counterpoint is to look at the Vision Pro.

I wrote before the Vision Pro came out that it was in trouble because entities like Netflix and YouTube were not going to be on there. And I thought, “Why would they go on there? Why would they want to help out, take a chance with Apple, given how poor Apple's treatment of them has been?” It's interesting because that article was written before the Vision Pro came out—

Mm-hmm.

—which was similar to my time at Windows 8. And I can tell you from that time, people are more receptive. The potential is a bigger selling point than reality. So I do think that was true then. The fact that you couldn't get Netflix on board is a really bad look for Apple and its state of developer relations.

Yeah.

The reason you can't get Netflix on board today is because there are no Vision Pros in the market. If Apple sold millions of Vision Pros tomorrow, Netflix would get an app for it. Now, it's purely market dynamics. To the extent that critique held, it holds in the potential era before something happens, when it's all kind of fuzzy and no one has proof points. It's like, “I've scratched your back all these years. Will you scratch my back and put an app on here for you?” And Apple's like, “I've stabbed you in the back for 10 years. Will you—

Yeah.

—will you scratch my back?”

Well, for 5 years now, the risk has been a regulatory risk, and that's the incentive to potentially adapt and evolve for Apple. But they haven't done it.

3. The App Store Tax

Two more emails I want to read before we shift gears. The first is from Vishal. He says, “Why shouldn't Apple charge whatever they want for their IP? They don't have 100% market share over app platforms. If Spotify or others don't like paying 30%, they can sign people up on other platforms. I don't understand why this is an issue.”

So I could answer this, Ben, but I get the sense—

No, you do it.

Well—

You answer it. I'm curious about your answer.

I mean, first of all, Apple can and will charge whatever they want in the App Store, and Apple's probably going to continue charging 30% fees on in-app purchases. None of this will really change until Congress intervenes, because in-app purchases are still going to be a massive business.

The reason Congress should intervene on the 30% fee, and the reason there should be a ban on Apple's anti-steering provisions, is that over time the economy will produce better outcomes for more people if everyone who wants to run some kind of business on a mobile phone isn't paying a 30% tax to Apple or Google.

So are you saying those are 2 separate things: the in-app purchase and anti-steering? Because those are 2 distinct—

They're 2 distinct—

—questions.

—things, yes. And I think in-app purchases are now being overlooked as everyone says, “All right. No more anti-steering. It's all good. This is a whole new era.” I think in-app purchases are still going to be an issue going forward unless Congress addresses that issue. Perhaps we'll cross that bridge in a couple of years. I just think—I understand where the writer is coming from.

Well, no, we're going to cross that bridge in this podcast right now, let me warn you.

Well, I understand where the writer is coming from. Apple has a right to capitalize on having a great product, and Apple has capitalized on having a great product. But I would say adhering to that principle starts to look ridiculous when you're then defending Apple's right to collect perpetual taxes on millions of digital businesses that just have no practical alternative but to pay Apple merely for the right to exist. What do you think?

So I have a few responses. I'm going to start by responding to you, and then I'll respond to Vishal.

Okay.

My response to you is that the reason I called out that distinction is, to your point, which you acknowledged, that you should be able to harvest the benefits of your creation, right?

Mm-hmm.

So we agree on that at a high level. I think, by and large, the user experience of in-app purchase is a great thing that Apple created and should be able to benefit from. We can accomplish our goals by letting apps go out to the web. That's what they control. Why do they get access to Apple's infrastructure?

The issue with the steering thing is this: it's not a free-speech argument, but it's philosophically a free—

Sure.

—speech thing. They're not letting developers—

It's freedom of commerce.

—talk to their users. This is Apple's problem: setting aside all the legal stuff, this specific aspect is so gross it just kind of disgusts everyone, including me.

Mm-hmm.

My view is, is it a little clunkier to go out to a webpage and have to make sure you're signed in? Again, tokenization helps fix this, but you do a purchase, then go back to your app. Is it a little clunkier? It absolutely is. Guess what? Apple gets to have a better experience because they created the better experience.

I have a much more difficult time with—and in fact pretty strenuously object to—developers having a right to Apple's integrated purchase when they can do their own purchase on their own website. If you want to use Apple's purchase flow, Apple gets to have a take, and I think that's totally reasonable.

I think you are going too far to say that it's not just that you get to collect payments from users without giving money to Apple; you also have not just a right to communicate to users, which I think is fine—

Mm-hmm.

—you have a right to Apple's creations to do so.

Yeah.

So I would just disagree with you, number 1.

That's fair, and I will just add that Apple's anti-steering provision—I'm only a couple of episodes into Andor season 2, but it's real imperial behavior from Apple, trying to collect fees from external websites days on from the initial link click.

But I actually agree that in-app purchase is a wonderful customer experience, and that's one of the reasons I'm confident that this issue isn't going to go away, because a lot of people are just going to prefer to do business in the app.

Yeah, but now we’re getting to entitlement. Why do developers have an entitlement to a better experience? They can own websites.

No.

This is where I actually agree with Vishal: they could make their own platform with a beautiful, smooth experience. They want access to Apple’s platform.

To seamless, frictionless purchases. Yes, I just think the rates are super competitive.

Well, let’s talk about the rates. What’s interesting about the rates is I can spin an argument for why the rates should be 99% and an argument for why the rates should be 0%. It just depends on how you look at it.

The 99% is what I wrote in this article. The reality is that people want apps on their phones. Their phones are the most important devices to them. Apple controls that, and because of that, Apple is like Facebook in its advertising, right? If Facebook finds you a customer you would never have found otherwise, what’s the limit on what they can charge for that? Every cent that you make is additive, right?

It’s not quite the case with apps because you could go otherwise, so it’s not a perfect scenario. But by and large—and this was the point of the article—the power of owning all the users on a platform that is essential is that you can do what you want, and developers will abide by it. They’ll complain, but they’ll abide by it because that’s how powerful platforms are. So that’s the argument for the one side.

The argument for the opposite side—why it would be 0%—is that, suppose Apple were not allowed to charge anything for its intellectual property. The market-clearing price would be 0%.

They would still offer it. Because it’s so important to have apps collectively, even if you have power over any one individual app, having apps is such an essential ingredient for a platform that if Apple were forced to, it would still make APIs and documentation and all these sorts of things. They would be like Windows back in the day.

It’s funny: everyone points to Windows as, “See, that’s how you should do it.” Honestly, if anyone knows anything about Windows, I think it’s probably the case that they just didn’t think of being able to charge this much, or it wasn’t possible in a world when you bought applications—

Disks.

—on CD-ROMs—

Yeah.

—or disks, back in the day.

So the reality is the market-clearing price is 0%. This is my whole problem. This is the fundamental logical flaw in the judge’s case: she’s just asserting it’s super competitive based on nothing.

Because the answer is, actually, the correct rate is either 99.9% or 0%. Any point in between is arbitrary. 30% is arbitrary. She’s mad at Apple doing a post hoc valuation of its IP, but any evaluation is post hoc because that’s just the nature of it. There’s some aspect of this decision that is not in touch with reality, and that speaks to the lack of an appropriate law in this case.

This is my point all along. The laws we have don’t address this as clearly as they could. That’s why I conclude by saying, look, it’s actually pretty straightforward: if you have a platform—and I think you could define a platform pretty tightly—you have a high number, which means it’s consumer, not enterprise, because in the enterprise it’s a much more sophisticated set of buyers and users, and it’s much more viable to do other things.

If you have a platform, you break 2 fundamental American truisms, and I threw in this extra one along the way. The platform doesn’t get to interfere with your commerce—

Mm-hmm.

—and it doesn’t get to interfere with your speech. Who can be opposed to that, right? Apple doesn’t get to kick apps off because of what content they have or don’t have, and Apple doesn’t get to kick apps off because of what they do on their websites between 2 consenting adults, right? Who’s going to be opposed to that?

I was going to say. You may run into some issues where Apple has traditionally not been a fan of adult content on Apple.

No, there’s a case to be made that the judge’s opinion is 100% morally right, but it’s legal nonsense and economic nonsense. You just stated blindly that it’s super competitive. Super competitive based on what?

Well, if there were other app stores available, they wouldn’t be charging 30%. Also, I think that the rational economic outcome—like Spotify, for instance—is ultimately a redistribution of wealth from Apple to Spotify. But Spotify is the one providing the value to users. Apple is not providing 30% of the value to users for every single subscription.

I completely agree. Just to be super clear, I feel like I’ve said it’s so important, which is why I say it again and again, because there’s such emotional resonance to this bit, particularly among developers. I completely agree. I’ve been objecting to this for years and years and years. I hope my stripes in this case are well-established.

Mm-hmm.

But at the end of the day, Spotify’s not going to pull its app from Apple and say, “Actually, Spotify is only available on Android,” or, “We’re going to build our own Spotify phone, so that’s how we’re going to compete.” It’s not going to, right?

No. And that’s fair. I only raise that to say that would be my response if a judge asked me, “So, is this super competitive?” We’re saying “supra.”

Supra.

Not super.

Yeah, supra. Yes.

I would invoke it. Yes, I need to enunciate better. But yes, that’s what I would invoke. Ultimately, I think this is not the optimal economic arrangement, but Apple is a platform, makes great products, and is able to dictate terms in perpetuity until Congress intervenes.

Yeah. Well, and by the way, the judge’s solution is the solution: allow links out to the web. The problem is, again, I think Apple did file both an appeal and a request for a stay. We’ll see what happens. I think there’s a 50/50 chance, maybe 55/45, that they’ll actually win.

Mm-hmm.

But if they win, or even if they don’t win, I think a law here is actually—

I’m not one to call for laws. I’m not one to call for regulations. There are aspects of the DMA that do get at this. The DMA is just way too complicated and way too in-depth, and it tries to police too many things.

The fact of the matter is, everything is a trade-off. I am someone who is hesitant about regulation. I want to open up the way for innovation. But there’s a bit where infrastructure is important, and platforms are infrastructure.

At the end of the day, if we want innovation, if we want creative outcomes, then sorry, Apple. Congratulations on your success. I again completely disagree with you that you should be forced to open up anything on your phone that you don’t want to, including in-app purchase. But the idea that you can police what developers say as a price to be on an essential platform is bad economically, and it’s fundamentally un-American.

As you noted, on Wednesday night, Mark Gurman reported that Apple has filed a motion for a pause on last week’s ruling on the App Store that it must stop collecting payments made on the outside web. This journey is going to continue for years to come. Good news for everybody, we're never going to stop talking about the App Store here on Sharp Tech.

4. Apple's AI Search Gamble

On that note, I want to stick with Bloomberg and a story that is also from Mark Gurman and also from Wednesday. Busy day for Gurman. He wrote that Apple is “actively looking at” revamping the Safari web browser on its devices to focus on AI-powered search engines, “a seismic shift for the industry hastened by the potential end of a longtime partnership with Google.”

Eddy Cue, Apple’s senior vice president of services, made the disclosure Wednesday during his testimony in the remedies phase of U.S. v. Google. Cue said he believes that AI search providers, including OpenAI, Perplexity AI Inc., and Anthropic PBC, will eventually replace standard search engines like Alphabet’s Google. He said he believes Apple will bring those options to Safari in the future.

“Prior to AI, my feeling around this was none of the others were valid choices. I think today,” he said, “there is much greater potential because there are new entrants attacking the problem in a different way.”

Not included in that excerpt was a note from Cue that, in April, searches on Safari dipped for the first time in 2 decades. So clearly some user behavior has already begun to shift. What do you make of this discussion and what it could mean for Google?

Well, the incentives here are important to note, because Apple gets $20 billion to $25 billion—or whatever the number is—of pure profit from Google based on searches in Safari. It’s pretty remarkable. Apple somehow, as an integrated device manufacturer, has made itself susceptible to AI disruption.

There’s a bit here where Apple has squeezed the juice out of every aspect of its business to such an extreme degree that it’s actually way more brittle and at risk than it should be. Why should Apple be having a quarter of its profit coming from Google?

Why should Apple be so dependent on China? Why should Apple be so locked down on this App Store stuff and not willing to give an inch? This is maybe another way of looking at the Cook critique, which is that this company has resolutely optimized to an unhealthy degree.

I wrote an article years ago talking about this. I think I quoted Aesop's parable about the reed and the oak tree.

Mm-hmm.

The oak tree mocks the reed that bows before the wind, and then the storm comes, the oak tree is uprooted, and the reed is still there. Tim Cook has destroyed Apple's resilience in multiple areas, and this is a good example. Of course, Eddy Cue is up there saying, “Oh yeah, Google's in big trouble,” because we need Google.

We need this money, man.

That's right. We're extremely dependent on this money for our profit and ultimately our stock price and our compensation.

And we should clarify: the incentive there is that Google is trying to argue that the search market is more competitive than ever.

That's right.

Therefore, the claims of its monopolistic dominance are overstated. Apple is an interested party trying to help Google sell that story.

And the way to help them is to take Google's stock price down by, like, 10% or whatever by—

Yep.

—revealing what we've been predicting and talking about for years is happening. Search volume going down is a big deal, because search is a service. It's something you do. So even if, say, iPhone sales flatline because the upgrade cycle is elongated—the favorite word of analysts—

Mm-hmm.

—because people don't buy phones more frequently, or they buy used phones—

It's also true. Yeah.

—or they're still using their phones every day and so should be searching every day. So the fact that search volume is diminishing is not because people are using their phones less. It's because they're using something else on their phones.

The thing with AI that should be so compelling as a product is that anyone who seriously uses AI only uses it more and more. It's a one-way ratchet. It's not like Instagram, where you get super into it, then one day you're like, “Why am I wasting hours of my life?”

Making better choices.

And you just stop.

Yeah.

You don't just stop using AI. AI is much closer to the personal computer or the phone than it is to an application. It's like a level down in the Maslow hierarchy stack, or a level up, or whatever direction it is.

And so this is only going to go in one direction. Now, for now, Google still retains commercial queries, so the value of those searches is still high. I would bet that Apple's payments are not diminishing because they're paid on a performance basis, and so they're probably still fine financially.

But the reason the market reacts to this is that, theoretically, this should be happening, and boom, here's a proof point. There you go.

Mm-hmm.

Yeah, it also makes me wonder: is the end of the Google–Apple search payments default-placement arrangement a bigger problem for Google or Apple? Because removing Google as the default on iPhones could really accelerate this process and make it a lot more real a lot sooner.

I think that's part of what people were reacting to on Wednesday: “Wow, okay, so if Apple is going to make it easier to just search through Safari on ChatGPT or Perplexity or any of these other options, this could change overnight.” It feels like an asteroid that is now much, much closer to Google Search.

You've always been able to use a different search engine. There are 5 or so—now, the selection list is limited by Apple—but you've been able to change it. Most people don't because Google is good enough.

I think your point is that if you can change it to one that's utterly better, maybe more users will do it, but this whole thing has always been about defaults. My initial answer to your question is that it hurts Apple more. A quarter of your profit disappearing is a quarter of your profit disappearing. That's just kind of hard to paper over.

You can make the case that Google, at any other time until now, this would've been a non-event because people would just reselect Google. Fine, we have to—

Right.

—have a search engine choice screen. Now, if we didn't, it might be different, but the fact that it might be different speaks to the possibility that this judgment is not necessary because there actually is competition.

It's like Facebook dominates social media. It's like, well, social media doesn't really matter anymore. It's about user-generated content, and TikTok's taking a big bite out of them. If you zoom out even larger, this is the general critique: is the heavy-handed hand of the court necessary for a 5-year interregnum of dominance when the evidence is that it keeps changing and keeps going away?

Again, I think for platforms, probably yes. But when you get to some of these area questions, it's a little fuzzier.

Yeah. It's interesting because I would go the other direction and say that this is evidence of why the verdict was important and why ending these default payments is important. Right now, Apple may be forced to integrate with ChatGPT or Claude or whatever it may be—Perplexity.

But if not for this verdict, Apple would have every incentive to continue working with Google because Google monetizes much better than any of the AI options in the market, and it would just continue on with most iPhone users.

I feel bad admitting this on the podcast because you were very judgmental of another friend who wasn't really availing himself of AI search capabilities. But I don't have that much firsthand experience with using ChatGPT for search because I use Google and Safari for 95% of my search needs, and that's how most people behave.

Apple would be happy to keep that going in perpetuity and continue to make $20 billion a year, and in a couple years I'm sure it'll be $25 billion a year, if not for the court intervening here and saying, “This payment is illegal given Google's market power, and it's time to go a different direction.” I think we're getting exactly the sort of intervention at the perfect time that the body of antitrust law envisions.

Like I said, your movement ends in authoritarianism because people don't do what you want them to do. You can set another default browser, by the way, which you couldn't do previously, so that's an important change.

You can set a different search engine. You can install the ChatGPT app. You're just mad that people don't do what you want them to do. So—

Well—

We're talking about defaults here. No, I'm being a little antagonistic. I'm sort of—

No, but I'm going back. The reason that conversation about the default is more compelling now than it was a couple years ago is because there actually are alternatives to Google that are more compelling in a lot of people's eyes.

And you're mad that people won't go out and find them and use them.

Mm-hmm.

Again, this is less of a legal argument and more of a “this just rubs me the wrong way” argument. People being mad about other people not doing the thing they want them to do—that's honestly a motivation that I'm just instinctually opposed to.

That's fair. Google is headed for some turbulent waters over the next couple years if, in fact, Apple ends up making this shift at some point. And it's interesting because—

Well, but that's the thing. Apple has no interest in making this shift because they make so much money. This is a whole other point. It's not in Apple's interest to—I mean, Apple is kind of a paternalistic company also. They like to say, “We know what's better for you than otherwise.”

They're the empire, man. Yeah.

They're not going to give you the best experience. They're going to give you the experience that monetizes the best.

Again, they've over-optimized themselves into this corner. They're up there saying, “Please, Google, hold on. Figure out AI Overviews. Do a good job. We need the money. We're dependent on the money now.”

Please, Judge, the market is more competitive than ever in 2025. Yes.

Well, one final question on Google. Would it be fair to say that Google is still well-positioned for the future even if Search is disrupted? They still have Google Cloud. They've got YouTube. They've got infrastructure advantages in AI and Gemini. How panicked should people actually be in the wake of these developments?

I don't know. Google is really hard to evaluate. Like I said, they still dominate commercial queries. They still...

There’s a lot of infrastructure that goes into ads that’s hard and difficult to build. It works very well. Their CPMs are going up. There’s a flight-to-safety aspect to them. We critiqued the sort of “users have to click on the ads” route, but there’s also some value in that. Can you ever actually deliver a compelling ad experience in a chatbot?

Mm-hmm.

AI Overviews, by the way, are the most-used ad product in the world, which is kind of like—

Yeah.

You don’t think about it that way, but more people are using generative AI on Google than on any other platform. They claim that they have ads in there. They claim they monetize as well as Search. So there’s reason to believe that they could. Just as a rule of thumb, these things hold on longer than you think, and new things tend to layer on top as opposed to completely replace them.

On the other hand, Search on Google sucks compared to ChatGPT for a lot of things. It is interesting, though, because ChatGPT is kind of like the Wirecutter phenomenon. Wirecutter is the review site that The New York Times bought, and it’s like, is this actually the best for me, or is it sufficiently convenient to just go there that I’m just going to go with it?

Yeah.

You—

Well-presented.

Yeah. No, exactly.

Totally.

Maybe ChatGPT could be better if you really let it understand your preferences. I hate review sites by and large because—how do I say this? This is going to sound terrible—they over-index on price.

Hmm.

If I’m buying this particular thing—

Show me the best product. I don’t care how much it costs.

Yeah. Let me do a value perspective: is this worth XYZ? I actually want an absolute quality evaluation. So maybe I should put that in my prompt for ChatGPT: “No, I want to know the best with a list of prices,” and I will say, “Yeah, that’s way too expensive.”

You know what?

“I’ll go for XYZ.”

I wish I could make fun of you for being Marie Antoinette on the other line here, but I actually wholeheartedly agree because there’s nothing worse than going to a Wirecutter-type article and getting the cheaper option because they essentially oversell it, and then realizing, “This kind of sucks. I should have gotten the more expensive option,” and having to spend that much more money at the end of the whole process. Just give me the best product.

Yeah. Or just live with the fact that you don’t have what you actually wanted.

Yes.

It’s so particular.

Exactly.

Right? There are some products I don’t really care about; I just need something that’s functional. There are other products. I have thousands of dollars of audio equipment here. I have $2,000 mics—

So do I.

A travel one and one here. I have a very high-end mic interface. And you know what? It’s cheap relative to its importance to my work and career.

Mm-hmm.

If I go to a site and it’s like, “Well, this is actually your best option because of some sort of value,” that doesn’t resonate with me. That’s not what I need. So, yes, but this is an aside.

Just a couple of podcasters here talking shop, talking hardware on the podcast. Maybe we could be reviewers at some point.

Well, wait. Let’s throw Bill Bishop under the bus, right? Bill Bishop is very expressive and moves around. So we ended up getting him a headset mic, so the mic’s always in front of his mouth.

It follows him, right?

You get some diminishment in quality, but it’s better than it was. I remember going through this whole thing. All the recommendations are for these low-end ones—no, we got the same mics they use for NFL broadcasts, right? It’s a podcast you’re selling for money. I’m going to invest to get the best thing.

What I had to do was go to an audio professional, explain the situation, and then he’s like, “Yeah, this is what you need, and you need this interface box, and you need X, Y, Z,” and it’s great. It works great. No complaints from the editor or anyone else about the sound quality now.

It’s interesting. I’m actually thinking about this: maybe this is actually bad for Google, because this is something the AI should be better at. Google should, in theory—this gets back to the Meta AI and the personalized-AI bit—

Mm-hmm.

I would like Google to gather more information on me. I would like it to track my purchasing habits and start to build this sense of what my priorities are, what I care about. If anything, we’re entering an era where privacy is just a massive detriment to the user experience. I want my AI to know more about me.

This idea I was just pitching—go into my custom prompt and tell ChatGPT to focus on X, Y—why do I have to do that? You should be—

It should just know.

Yeah.

Yeah.

No, exactly.

Yes. Well, I have breaking news.

Anyhow, off to the guillotine next.

I have breaking news here on the podcast. Have you seen the Celtics score?

I have, and it’s been hard for me to focus for the last 30 minutes—

Oh, my God.

I was going to tell you at the end.

Just magnificent. I love to see it. Even though I didn’t literally see it, I will be watching it the second we finish recording here.

I did turn on a recording because I find the Celtics winning—

Yeah.

Depressing.

I don’t normally watch that.

Yeah, so fortunately, I have means to access your recordings. I’ll be downloading it from my recording shortly.

Oh, my God. Just spectacular.

I definitely enjoy watching the Celtics lose, I can tell you that.

And I’m going to be taking credit—

I lied.

—for the Lakers jinx.

You know what? Total lie. I’m not cheering for Denver. I’m cheering for a Celtics loss first and foremost.

Whoever the Celtics are playing.

Past that.

Go Knicks.

That’s right.

Absolutely. Let’s go Knicks. All right. Final segment here. I’m going to read you 6 quotes from a New York Magazine article on ChatGPT’s impact on colleges. The headline was, “Everyone Is Cheating Their Way Through College.”

Quote number 1: “AI’s takeover is a full-blown existential crisis for higher education.” Quote number 2: “College is just how well I can use ChatGPT at this point.” Quote number 3: “I think we are years, or months probably, away from a world where nobody thinks using AI for homework is considered cheating.”

Quote number 4: “It isn’t as if cheating is new, but now, as one student put it, the ceiling has been blown off. Who could resist a tool that makes every assignment easier with seemingly no consequences?” Quote number 5: “Massive numbers of students are going to emerge from university with degrees and into the workforce who are essentially illiterate. The humanities, and writing in particular, are quickly becoming an anachronistic art elective like basket weaving.”

And then 6: “Every time I talk to a colleague about this”—this is a professor speaking—“the same thing comes up: retirement. When can I retire? When can I get out of this? That’s what we’re all thinking right now.”

So, Ben, I sent you this article before we came on to record. It went minorly viral on Wednesday.

Hilarious article.

As a parent of a child who’s going to be attending college in the fall, what do you think of all this? Is it a problem? And if so, what do the solutions look like?

Well, if I could pat myself on the back, I’m looking at this article on Monday, December 5, 2022. I believe that ChatGPT came out on a Wednesday, so I didn’t violate my general “don’t write on Friday”; I waited until the following Monday. And my article is entitled “AI Homework.”

My way to write about ChatGPT from day one was to start with homework and to expand from there about how this is going to change things and change everything.

I completely forgot that.

Being-right points for me. Yes, thank you.

Big victory lap. 2.5-year victory lap. I appreciate it.

I mean—well, this is the challenge, right? This was eminently foreseeable, and it’s funny to see people worrying about it now in 2025.

Mm-hmm.

I think, zooming out, by and large, things like this are less causal than they are revelatory. What I mean is, ChatGPT is not causing college to be worthless. It's revealing how worthless the vast majority of college is.

Mm. Good take.

And I think you just read this article, so much of this comes through in it. The unfortunate reality is—I assume the millennial author worried about Gen Z not knowing what's true and what's not, when this entire article was caked in this sort of embrace of these myths. For example, everyone should go to college; everyone is equally capable. This is just dripping with a blank-slate sort of ideology, which has tremendous consequences. That's the problem with myths: when you act on them.

Why do we end up with this hilariously distorting market of student loans and all these bits and pieces where we're going to immiserate all these people? And, oh, by the way, now we're going to make it so they can't declare bankruptcy to get out of it? All these knock-on terrible effects are downstream from saying, "If we just get people college degrees, they'll magically make more money."

Yeah.

No, actually, the premium for a valuable college degree is still significant, so of course I'm going to send my kids to college. I think people go too far with this. They start highlighting some of these issues with college, like, "I would never send my kid to college." No, you're being a moron. Of course you send your kid to college.

The reality is, it is an essential credential. Now, it's bad that we've extended the credential. The credential used to be, like, 12, right? Then it was like 18: you need to have a high school degree. Now it's 22: you have to have a college degree. And people are like, "Oh, what about fertility?" Well, maybe if you're extending the credential to be a functioning person in society—or not a functioning person, but to make X amount of money in society—you're pushing it up to 22. And now you need a graduate degree to be a librarian or something, right? Now it's up to 24 or 25.

There are all these knock-on effects that are downstream from mixing up cause and effect. The reality is that college, by and large, is a filtering mechanism to see if people sufficiently have their shit together to get a degree that actually isn't that hard, because we've dumbed down and introduced all these majors. It's funny; I've embraced all these myths. I mentioned last episode that I was a political science major. Dumb major, to be clear. I was in it because I loved it. I thought it was super interesting.

It's a really fun class.

No—

I really enjoyed my political science classes.

Well, the great thing about it was I realized it was like a cheat code. Almost every class I took in my last 2 years was one of these seminars. I actually organized a few seminars. I went to professors I knew and said, "I got a group of people who would like to do this. Can we spin up a seminar?" There would be 5 people reading. I did one about the philosophy of education, and there were 5 of us reading Plato or whatever and having these 2-hour-long debates about it. I love talking and debating. It was freaking awesome.

Mm-hmm.

My last 2 years were amazing in college. The reality is that, while I maximized it, I'm not sure what political science is one of the largest majors at the University of Wisconsin—and not because everyone wants to sit around and debate the philosophy of education and read Plato. It's because it's easy, okay?

Yeah.

I freely admit that. But that kind of speaks to it, right? College back in the day—not hundreds of years ago—was for a very small number of people who were intrinsically motivated to discover and find all these things. And, by the way, if you still have that motivation, AI is amazing. It actually lets you know more.

5. AI Rewards Agency

I started 15 AI chats yesterday about basically a gazillion different things, because why not? I was the heaviest Google user in the world. When I'm in a conversation with someone and some question comes up about some fact in the world and no one knows the answer, I can't handle it. I immediately go look it up. Then people are like, "Wow, Ben, why do you know so many things?" Because I'm a freaking maniac: every time I encounter something I don't know, I look it up, and then I have a good memory, so I sort of retain it and keep it in. And, by the way, I love to argue and debate about everything. College for me was amazing. It was this incredible experience, and AI only enhances that. It gives me agency to know more, to dive deeper, to see all these sorts of things.

Mm.

And the reality is that most people aren't at college for that reason, and this isn't a value judgment. There are lots of things to be done in the world and lots of roles and things to play that don't entail debating the philosophy of education and reading Plato. But the problem is, as a society, we've instituted this idea that you actually need a college degree as a test of conscientiousness and basic intelligence.

The way we got here involves some Supreme Court decisions. There's this case called Griggs v. Duke Power Co., involving disparate impact. Employers can't give tests to employees, or else they'll get sued into oblivion, basically. So they've outsourced filtering to colleges. But because people who pass the filter make more money, it got interpreted as, "Just get more people shoved through the filter."

Mm-hmm.

What AI is doing is exposing this farce, exposing how we've flipped cause and effect and how we've gotten this all backwards. So you have people going in there and cheating—whatever you want to call it—on all the papers. There was one in here about the history of jazz music, and I hate to break it to you, professor, but people are in that class not because they love music. They're in that class because they have some stupid requirement to take some sort of—

And it sounds like—

Class—

An easy A. Let's just call it what it is.

Yeah. I took a history of music or something. I loved it. It was a great class. It was also ridiculously easy. We sat around and listened to music and wrote thoughts about it. It was all good.

Yeah.

This is technology, and you see this again and again. It's revelatory. We were just talking about this—what podcast were we talking about this on just a little bit ago? The competition bit, right? A big problem people have with tech is that it engenders perfect competition, and people don't like that.

Hmm.

What is revealed is that there are only, like, 10 sports columnists in the U.S. worth reading, right? The fact that every city used to have one wasn't because there were a lot of great sports columnists. It was because people were geographically constrained to their local newspaper, and every local newspaper had a columnist. As soon as people could read the whole internet, guess what? They all flocked to Bill Simmons.

Yeah.

A similar dynamic is happening here. AI is not ruining college. College was ruined, and AI is exposing the extent of the rot.

So I like that take. I do think it's a little bit more complicated than that. First of all, what I don't understand is, if cheating is the concern—and this is probably because I'm old and my last academic experience was law school, where basically every class is just a series of lectures, and then you have a written in-class exam at the end—I don't know why that model is supposedly unavailable to professors today.

Yeah. No, the—

If—

I have never understood the whole take-home exam thing in general, even when I was in college 25 years ago.

Yeah.

I'm just coming out guns blazing because I had a hot take and I really wanted to fully deliver it.

And I'm just noting for the record: I think it's more complicated. I enjoy the guns-blazing Ben at the end.

No, I know, I agree. I do. I just want to cook today. But there is an aspect of this that is sad, to be clear. Even going back then, it was basically on the honor system that you were going to do your take-home exam according to the parameters the professor delivered and do it, and it mostly worked. One of the sad things about our society is the extent to which stuff like that doesn't work.

Mm-hmm.

We've talked about this in all sorts of arenas. As you lose societally accepted mores and taboos, you're forced to introduce hard laws.

And when you have hard laws, you're introducing red tape and all these sorts of issues. There's a massive loss that comes from that. So, yes, one of the answers is to go back to in-class exams: write down your answer.

Totally. If the concern is purely rooted in academic integrity—and again, I'm probably dating myself here—I took a lot of exams in blue books, and it doesn't look like it should be that hard.

Well, and by the way, as long as I'm on my soapbox, taking notes by hand—

Better.

—offers much better retention than trying to take notes—

Totally.

—on the computer. I actually think, look, okay, college, I'll let you be paternalistic: ban laptops in the classroom. Make people write it down. You're actually doing them a favor. I'm fine with that to an extent, but there is a bit where, if you think about it from an employer perspective, if employers are looking at colleges as certification machines, guess what employers want their employees to do?

Know how to use and capitalize on AI tools.

Know how to use AI. Colleges are actually doing their job, right?

Yeah.

They're like, “How can you get through all this busywork?” Busywork which accumulates because we need more and more laws because we've lost societal mores and taboos that control people's behavior. As a society, we're eliminating judgment from people anyway, so there's some aspect where actually this is totally normal and arguably even appropriate. And, like—

Yeah.

—the guy they opened with, the guy who's making all the cheating apps and things along those lines—

Hmm.

—he's not going to work for a big company or something. It's like, yeah, of course he's not. But also, big companies are liars too. They're like, “We want you, we want entrepreneurial students. We want people—”

Hmm.

No, they don't. They want cogs in the machine. There's a reason—by the way, I did that job at Microsoft. I took the initiative to build that deck. I didn't get a single promotion at Microsoft, which really irked me at the time, but in retrospect, I was a—

You get it.

Yeah, I get it. I wanted a promotion. I was surly. I had little patience for people who didn't get it.

Really difficult to work with, yeah.

There's a bit where me making that deck was so out of line because it was a deck that someone way higher than me in the stack had made. It was just so good that it got adopted anyway. I had an amazing manager who really, I think, shielded me and protected me from this stuff. But also, to get buy-in for a promotion, he had to convince his other peer groups, and he was not able to do that.

All the qualities that made me a poor corporate cog at Microsoft are the exact qualities that enable me to sit on this podcast, have tens of thousands of listeners, and cook on here and say this is all pointless and worthless, right?

Yeah.

I don't care what people think. I don't care what they say. I'm going to go out and do my own thing. More power to this league. I was totally unsuited for college anyway. There's a bit where the returns to people like that, questionable though their morals may be, are only getting larger.

The internet was already step 1. People like you and I could go out and just put a podcast out there, and because it's perfect competition, our willingness to compete and ability to succeed were rewarded. We've talked about the returns to AGS coming from AI. It's only going to get larger.

Mm-hmm.

And it's going to be tremendously disruptive. There is a bit that's sad about this. We romanticize the college experience. I think fondly back to those seminars that I sat in and debating things back and forth. We talk about, “Oh, this is a dorm-room topic that we're just going to debate back and forth.”

Absolutely, man. Common room.

It's sad to see that go. But one thing that's important to keep in mind—and this is going to be a really hard thing to deal with—I think a problem societally, generally, is we have a romantic view of humanity that doesn't match reality.

Hmm.

For example, there's a lot of crappy jobs in the world that need to be done. People do those jobs not because it's their calling or their dream. It's because they need money. That's a job they can get that puts food on the table, and if they don't have the job, they're not going to eat.

Yeah.

And that's very difficult to think about because it seems vaguely or explicitly wrong. Of course, people should be able to eat. People have a right to eat. They have a right to housing. They have a right to health care, all these sorts of things.

Again, this is a dorm-room discussion to a certain extent, but we do ourselves a disfavor in thinking about the reality of humanity by not appreciating the extent to which people do stuff they don't want to do because they have no choice.

Mm-hmm.

There's an aspect of this college discussion that's along similar lines. There are a lot of professors here who fooled themselves into thinking students were doing work for their classes because they were interested in learning. Actually, you have a faulty view of human nature.

It was always a chase to a credential. Yeah, no, that makes sense. We're not going to close out the podcast by quoting Caddyshack—the world needs ditch diggers too—but I do want to read this professor. He says, “Williams knew most of the students in this general education class were not destined to be writers, but he thought the work of getting from a blank page to a few semi-coherent pages was, above all else, a lesson in effort. In that sense, most of his students utterly failed.”

They're using AI because it's a simple solution, and it's an easy way for them not to put in time writing essays. I get it because I hated writing essays too when I was in school. But now, whenever they encounter a little bit of difficulty, instead of fighting their way through that and growing from it, they retreat to something that makes it a lot easier for them.

I just want to say that that's a real concern, and if people are abdicating their critical thinking and entrusting ChatGPT to do it for them, I don't know where that leaves us as a society in 10 or 15 years.

Well, the other hilarious thing—again, there are a lot of very funny things in this article—

People should read the article.

—if you read between the lines.

It's a really entertaining read.

One was the girl who's like, “Yeah, I have so little time to do my homework because I spend 10 hours a day on TikTok.”

“Hours and hours, until my eyes start hurting,” she says. But ChatGPT makes it easy to write an essay in 2 hours. Yeah, that's the concern. And, you know, as someone who—

But that exists, right? Again, I don't know that we have the answer to this. These are real concerns. You go back, to use a literary reference, to the movie Up—or not Up. I always get them mixed up: WALL-E. I always get WALL-E and Up mixed up because WALL-E is a spaceship. It's Up, right?

Mm.

WALL-E—this world of people just sitting in their chairs, endlessly amusing themselves—seems pretty prescient.

And you know what, though? You also talk about the value of these tools. You put in a good 20 years learning in an analog fashion. You were taking notes by hand, and you were reading real books and writing real essays, and so that lays a foundation that I think makes all these tools a lot more valuable. And if people don't have that, I wonder where that leads.

It leads to them sitting on their couch, surfing TikTok, and vaguely depressed about their life, but drowning it in social media.

Well, hey, there was that one kid who launched his own cheating-app business, so it's working out for him.

No, but that kid's fine.

Yeah.

No, I know. But the broader concern is we're barreling toward this reality of the vast majority of people being vaguely dissatisfied with their lives, but having enough to eat, having a roof over their heads, and entertainment that sort of drowns out the vacancy in their lives.

Mm-hmm.

6. The Agency Divide

And while some smaller number of people, because they have the agency and initiative to actually take advantage of this stuff, are going to go out and do things and do far more things than they could otherwise, this is, again, the dorm-room question: is that a good thing?

If you measure world net GDP, probably yes, because the truly talented people with agency are going to get accentuated or amplified to such a dramatically greater effect. In many respects, when I talked about cogs in the corporate machine, corporate machines are there to amplify the person at the top, right?

Yeah.

Apple was a Formula 1 car built around Steve Jobs. NVIDIA is a Formula 1 car built around Jensen Huang. They are there to enact and articulate their vision. What if we could do that for more people without all the difficulty of having tens of thousands of employees?

Mm-hmm.

Right? It’s like the trade debate. You measure GDP because trade makes everyone richer when you measure it as an average, and then the question is, well, what about the experience of someone in the Rust Belt?

Yeah.

Is there something that’s not being measured that’s being lost? And this is the concern about these sorts of things. Maybe on average, from a purely economic lens, we’re going to be better off with all these tools, but the reality of humanity is that a huge number of folks are not just consigned—they’re self-consigning themselves to a world of being cattle.

Mm-hmm.

And that’s pretty depressing. But then again, it’s like the aggregator-versus-platform argument. They’re doing it themselves. They’re choosing to watch TikTok for 10 hours a day and then using ChatGPT to write their answers.

Yeah. Well, look, final thought here: If students aren’t learning how to think because they’re abdicating all of the critical thinking to ChatGPT—

Did they ever know how to think in the first place?

Well, I like to think I learned along the way. I don’t really romanticize my college experience.

I know, but you have initiative.

Yeah.

Right? The reality is—

I didn’t in college, I’ll tell you that much, but it emerged along the way.

I don’t know. I think—

No, I got good grades in college, but college is very easy. That’s one of the reasons I don’t really romanticize college all that much. But I think it is an existential problem for both—

What? Look at you.

What?

College is easy. Very easy. Didn’t even try.

It was pretty easy. I mean, let’s not kid ourselves. I just think, coming from the humanities world—I know a lot of people in that world—I’m channeling my Grantland pedigree here, the cultural elitist on the podcast.

So many people in that world are so smug and dismissive about AI tools, and by failing to engage with these tools, a lot of the people running these institutions are just way less equipped to actually come up with solutions that help students build an immunity to the problem here, work around the problem, and force everybody to use bluebooks. I think that is part of what will compound the crisis that afflicts higher education. But to your point, we started with the best take: That crisis was coming regardless.

7. The Human Nature Problem

Well, it really goes back to this societal malaise, this sickness that I think comes from an overly optimistic, utopian view of human nature.

Hmm.

Right? There’s been a strong drive for decades to diminish social taboos, to diminish mores, to empower individualism, and to say, if you’re not hurting anyone, just do what you want.

It’s very easy at an intellectual level, particularly for people like us and our audience, who are high-functioning adults who can make our own choices and can control ourselves, to say, “Yeah, everyone should be able to smoke weed,” to use an example. But what does it actually mean up and down society when you remove even the imprint of illegality—or the sense of that—from it?

It is. Everywhere smells like weed, and it sucks.

There are so many things that people do for themselves because, “I don’t want…” I’m actually not a weed smoker, but I have plenty of friends who smoke on the weekend once in a while. What they really resented was the vague sense of feeling like they were doing something wrong, because they’d been good kids all their lives—

Hmm.

—and they didn’t want to—

Good little credentialists.

They’re not the problem, right?

Just trying to smoke a joint.

The problem is people who are now smoking a joint every day, and it’s super concentrated. Then they’re basically dropouts from society, and all these sorts of things. It smells everywhere, right?

There’s this general libertarian argument that people—like your bit right here—over-index on themselves and what they want to see happen, and don’t think about the reality of society, up and down, and the collective efforts. Sometimes there are constraints that are put on you because it benefits the whole.

Sometimes those constraints are legal, like the weed example. Sometimes those constraints are mores: No, actually, you’re going to be looked down upon or judged for doing this particular activity. Even though you’re not going to go to jail for it, you’re going to get looked down upon, right?

Hmm.

There’s a drive in the liberal idea generally—and I just appealed to the American ideals of free commerce and free speech and all those sorts of things, so I’m arguing against myself a bit here—which is, take off the chains. Let people do what they want.

There are costs that come from that, and those costs don’t manifest themselves immediately. They take a long time, and sometimes they’re only exposed through things like AI and technology.

And unbelievably expensive college degrees that ultimately are worth less in the—

Yeah.

—economy that awaits us for—

Yeah.

—the next 10 or 20 years. Yeah.

I’m optimistic about my daughter. I think she takes a tremendous amount of initiative.

That’s right. You’re going to be paying—

She’s—

—tuition the next 4 years.

Well, I think it’s, like, the most expensive—I don’t know where she’s going. It’s not cheap.

She takes initiative. She’s very self-directed. She has her own thoughts and point of view, and I think she’s the sort of person who has high agency and will benefit from this. But of course, every parent thinks great things about their kids, so I guess—

There you go.

—we’ll see how it goes.

So we’ll see. For her sake, I hope that they continue to use take-home exams. For your sake, I hope they don’t.

The impressive thing about this is the kids who want to learn and want to push themselves, and they're graded on a curve, and so they're actually competing against the AI.

Mm-hmm.

They’re not competing against other students.

That sucks.

I meant to mention that earlier. It sucks for them. So much of our society is like this, right?

Yeah.

We’re actually punishing the people who want to do it the right way by basically catering to the lowest common denominator. In this case, the lowest common denominator is kids who don’t want to write their own essays. We’re actually punishing the ones who do.

Yeah. Well, on that note—a dour note to end on, but a rollicking podcast—what’s most important is that the Boston Celtics are now down 0–2 in the second round of the NBA playoffs.

0–2. Going to Madison Square Garden.

Let’s go, Knicks. All right, Ben. We only have one episode next week.

I said a lot of controversial things. We’re going to lose more listeners for our anti-Celtics sentiment than anything else, so we’re going to be good.

Right. Jilted Celtics fans, I can’t wait to hear from you in the days to come. And God, I hope they actually lose this series. Ben, we have one episode next week. It’s coming at the end of next week. We’ll be recording—

Oh, yes.

—in person.

No, no episode on Monday. Yes.

Very, very excited. Until then, enjoy your weekend, and I will talk to you soon.

Talk to you later.