(Preview) AI and the Winner’s Curse, Google’s Genie 3 Breakthrough, Questions on Intel, Fertility Rates, and Banning Advertising

Hello, and welcome to a free preview of Sharp Tech. Hello, and welcome back to another episode of Sharp Tech. I’m Andrew Sharp, and on the other line, back in the United States, is Ben Thompson. Ben, how are you doing?

Struggling, Andrew. The worst thing about getting old and having an increasingly difficult time with jet lag is—

Mm-hmm.

—the fact that I realize I just complain about it on every podcast. It’s a really bad combination: being old, dealing with jet lag, and podcasting regularly. I’m sure everyone’s sick of hearing about it, so I will not dwell on it.

No, it’s a great summer theme here. I had a feeling—because we’re recording later in the week and you’ve been back for a couple of days—that the jet lag would hit you like a ton of bricks after about 72 to 96 hours. Now here we are. Who knows whether we’ll make it through the next hour and a half, but I’m excited for the journey with you.

I’m glad you’re excited. Someone needs to be. Let’s do it.

Also, good news. This might be the last day of work either of us needs to do because, as we’re recording, OpenAI is set to release GPT-5. Sam Altman posted a photo of the Death Star looming over Earth last night, and I don’t know. The whole world could be about to change, and our future may just be you and I sitting around, smoking cigars and hanging out while GPT-5 does the work for everyone.

That’s a great point. We are obviously not going to be covering ChatGPT-5 on this episode.

Mm-hmm.

Who knows? Maybe ChatGPT-5 will produce this episode and replace everything we said. It’s a great observation. Who knows what might happen?

Exactly.

Yes. One long cigar night for the next several decades, at least as long as we can afford cigars.

1. The AI Winner’s Curse

In any event, we’ll hit GPT-5 next week. For now, we’re going bigger-picture on this episode. Last week, when we recorded, we had a question about big tech companies in the AI era. Bob asked us whether any big tech incumbents would lose in an AI world. On Tuesday this week, you wrote an article headlined “Paradigm Shifts and the Winner’s Curse,” which I thought was an interesting coda to the conversation we had a week ago about all the different advantages that the big tech incumbents are enjoying as the AI era takes shape.

Your article focuses on Apple and Amazon, and particularly AWS within Amazon, after last week’s earnings calls. So my question at the outset is: Why are we focusing on those 2 businesses? Apple just reported monster earnings, its fastest quarterly revenue growth in more than 3 years. AWS is still the most dominant cloud business, albeit with slower growth than competitors like Azure and Google Cloud. What drew you to the AWS story and the Apple story?

2. Google’s First Mover Failure

It’s not just an Apple and Amazon story. It’s also a Google story. The email contributed, so thumbs up to Bob.

I think when I reflect on where we are and where we’re going, I would say first and foremost that I brought up “The End of the Beginning” article in my response to Bob. I do think that generally holds. I’m not at all predicting the doom of Apple and Amazon. What they do still matters and will continue to matter.

To take Apple as an example, we’re going to need devices. To take AWS, all the companies on AWS aren’t going anywhere. They’ve established a really important and solid position.

Yeah.

But I just had to be honest about the fact that the other article I went back to was “AI and the Big Five.” All along in this whole AI paradigm, the company we’ve been wringing our hands over—and I’ve been a bit of a ringleader in doing this—has been Google.

Mm-hmm.

As we discussed a few episodes ago, I didn’t bring this up in the article, but maybe I should have. The real condemnation of Google is that they invented the transformer. They invented this technology, and they were totally caught flat-footed by OpenAI and ChatGPT.

Right.

ChatGPT is the product that, theoretically, Google should have created, at least from a technical perspective. The fact that they didn’t is disruption. Whatever happens going forward, ChatGPT is going to be a thing.

If OpenAI went bankrupt tomorrow, the bidding war to acquire ChatGPT would be astronomical. This is a product that is important. The whole structure, the equity structure, all of that with OpenAI—not just because it’s a nonprofit, but also because of all the cost issues we’ve talked about—is a separate question from whether ChatGPT is going to be a super-important and meaningful product going forward, which it is.

Right.

It’s dominant in the consumer space because they were the leaders in the consumer space, and now they have an advantage that just compounds over time with word of mouth, user preferences, and things like that.

Yeah, exactly.

All of those advantages could have gone to Google had they moved first, in 2021 or something like that, on the strength of the transformer technology.

3. Apple Was Not Always First

Apple loves to say, “We’re never first to a category. We just do it right,” blah, blah, blah.

I don’t know. That was definitely the case with the iPod. That’s probably the preeminent example. There were digital music players in the space. Apple came in with a completely different approach: a thousand songs in your pocket, based on this 1.8-inch drive they found in Japan from Toshiba, and turned around this product in 6 or 7 months. It was one of the most incredible product-development stories of all time.

But when you go to the iPhone, yes, smartphones existed before the iPhone, but smartphones as they existed before the iPhone were more capable dumb phones.

Mm-hmm.

Apple really did create and define the category. A lot of my early writing about Apple came at a time when people were assuming that Apple was going to be overtaken by Android—not just overtaken from a unit-share perspective, but that Apple was going to be doomed like the Mac.

Right.

No, Apple was first, and they established this huge ecosystem. They’re in a very strong position and aren’t going anywhere.

In contrast, the Mac was not first. DOS was first, built on IBM. The Mac came later. It was first to have a graphical user interface.

Mm-hmm.

But the fact that the Mac basically languished and carved out a little bit of space in the design market, particularly desktop publishing, meant that they had a nice little business in the late 1980s and early 1990s. By and large, though, it was a failed product. It cost Steve Jobs his job and his position at the company.

And so that’s actually a counterexample. The 2 biggest products, the Mac and the iPhone, show that being first is, in fact, really valuable. Again, it all comes down to definitions. You could say the iPhone was not the first smartphone, and so on. It was clearly the first smartphone as we understand smartphones today.

Right.

Had Android—

It established Apple’s advantage in the app ecosystem. Then Apple came through with iPhone-adjacent businesses like the Apple Watch and AirPods, for instance, along with a number of other examples that Tim Cook could point to and say, “We weren’t first in these categories, and then we came to dominate and define these categories.”

But it’s all downstream of the iPhone, which was first and then established an ecosystem that turned into a massive moat and made the iPhone the most successful product of all time.

Right. I think there’s an alternate reality where—again, I don’t think Android would be what it is today without the iPhone. Google was trying to copy the BlackBerry—

Mm-hmm.

—and then they totally pivoted. But imagine Google comes out with Android as it is today: all-touch-based icons, the whole App Store, the whole kit and caboodle. Would Apple be doing as well as it is if it had been 3 or 4 years behind Android, like Android was 3 or 4 years behind the iPhone?

Mm-hmm.

I’m a little skeptical. The extension of this observation is that what Tim Cook likes to say about Apple—“We’re not always first, but we come in and dominate,” blah, blah, blah—is actually kind of Google, right?

Google has Search, this money-generating thing, and then these other businesses. They come in with Chrome and just—

I was going to say, yeah.

—take over the whole market. They come in with Android. Apple’s iPhone is a really strong product, so it maintains its market share and a huge portion of the profit share, but Google cleans up everything else.

4. Google’s Zombie Strategy

And I think if you look at AI, it kind of feels like that's what's happening. ChatGPT is there. Anthropic is there, really focused on B2B and coding, things along those lines. But you have Google, which is some combination of a Loch Ness Monster and a zombie that is just irrepressible.

Mm-hmm.

And it's like, here, I guess we're going into AI products now. You have this entity that feels like it can't do anything right. It feels like it's so unfocused. It can't make good products. It's canceling things left and right.

I've recounted again and again how I have a hard time understanding this company. I was talking to someone who used to work at Google, and he's like, “I totally sympathize with your difficulty. You always talk about how you have a hard time understanding Google.”

Frustrations. Yeah.

He's like—this was someone who worked at Google and was a high-level executive for many, many years—“I don't understand Google. No one at Google understands Google.” And I do sometimes wonder: Do I have a bias against Google precisely because they are so frustrating for me as an analyst?

Right. Mm-hmm.

I want there to be a top-down strategy, something that I can parse and explain, and I can see what's going on. The problem is, the critique of Google is that the right hand doesn't know what the left hand is doing.

Yeah.

Or in the case of Google, it's like an octopus or whatever it is, and none of the tentacles know what the other ones are doing. But maybe there's a little bit of a method to the madness, and maybe it's not even a purposeful method. It's like Google is this organism, just sort of an amoeba that envelops—

I was going to say—

—everything as it floats around. And actually, it's really impressive what they did in mobile.

Yeah.

And if you squint, it feels like that's what's happening in the world with AI. Again, in this view, you have ChatGPT as an iPhone-like product that comes in, defines the market, and dominates the consumer space. But then you have the infrastructure king—

Mm-hmm.

Google has a competitive model. They have the infrastructure that allows them to offer it very cheaply. I'm very bullish about their cloud prospects, by and large.

And so, yes, this was an article about AWS and Amazon. But almost at the end it was—not an apology tour, per se—

A salute.

For me toward Google, but it was like a self-critique. This article was very introspective in a way.

Yeah. What I enjoyed about it was your description of Google as this amorphous blob. I don't know what the right animal analogy might be. Maybe a giant jellyfish, as opposed to—

It's like a zombie jellyfish.

—the giant Loch Ness Monster.

Yes.

Yeah, I guess the Loch Ness Monster is giant by its nature, but a giant zombie jellyfish—unkillable—because Google is dedicated to making cool products and shape-shifting over time. We've already seen them shape-shift to dominate the mobile era, the way they dominated the desktop era. And that amorphous nature should favor them as the AI paradigm arrives and takes shape here.

Then, on the other hand, you were juxtaposing them with Apple and Amazon, who have very clear, strategically sound strategies as this new technology arrives. And what was remarkable about your article is that you're tracing the comments from Tim Cook and Andy Jassy on the earnings calls, and beat for beat, they're making the exact same points about why Apple is well-positioned in AI and why AWS is well-positioned in AI.

They talk about, “It's early. We will serve actual use cases. Our chips are best. We have the data.” Both CEOs made comments to that effect, and those were the umbrella categories that you used to describe the points they continued to emphasize on these earnings calls.

And so all of that is true, and it goes back to the beginning, like two and a half years ago, when we looked at Apple and said, “They have this advantage, they have this advantage,” and then Apple Intelligence arrives and is announced, and it's like, well, this makes sense in theory for a way that they can capitalize on all the first-party data they have, all the advantages they have. What are the risks that you see of doubling down on those strengths as AI arrives here?

5. Apple’s AI Control Problem

I think it just goes back to the paradigm question. The Apple one, I think, is easier and more straightforward to understand. But we talk about a world where the phone is not the most important thing or, to the extent the phone is the most important thing, it is a conduit for AI.

Mm-hmm.

And if the plane of competition is not the user experience in terms of actually touching icons and moving them around or whatever it might be, and it's not necessarily the App Store or the players in your platform, but how good the AI that is accompanying you is, is that a competition that Apple should feel comfortable about?

Mm-hmm.

Yeah, this was where the Nokia tie-in came in. It's a little risky to use Nokia because they're not in the phone business anymore, and I'm not saying that Apple is going to be in that position. But—

Right.

My whole thing with Nokia is, if they had just adopted Android, they'd probably still be around and be fine. Would they necessarily be as profitable, as dominant? Probably not. But then again, they were operating in a world where they were making phones. They were making phones to carrier specifications, delivering carriers what they wanted. They weren't differentiated.

Yes, Symbian—I had a Symbian phone. It had some niceties, but at the end of the day, you bought it for the hardware, and Nokia hardware—

Mm-hmm.

—is amazing, by the way. Still some of my favorite phones. But if they adopted Android from the get-go and got all this stuff that they were bad at for free, would they still be under the thumb of Google to a certain extent because the operating system matters more? Yes. That's also a better existence than the one they ended up with. All right?

Completely.

The real cherry on the sundae is when they finally realize that Symbian's not going to make it. Instead of going to Android then, they go to Windows Phone, and they had such a need to be different.

Mm.

It's like some status-obsessed individual who just can't eat at the right restaurant or can't get the right car. They have to be different, and they just keep getting themselves into worlds of hurt because all that matters is that they have to be distinct. That was Nokia by the end, and in the end they're not around anymore.

The analogy to Apple is maybe they should just partner with someone and let them do Siri, right? Anthropic or OpenAI or whoever it might be. And do they have to actually control it? You could say, yes, they have to control it because that's the future. But what if you suck at it?

Right.

Right?

I mean—

You have to make it to the—

That's a real concern.

You have to make it to the future to worry about control.

And again, sitting here today, the iPhone's not going anywhere. Come on, Ben.

Yeah.

But then again, how good is this AI going to get? How important is it going to be? And if you're stuck with Siri and your ragtag group of AI engineers who weren't good enough to get an offer from Meta, how competitive are you going to be?

6. Apple’s Contrarian AI Bet

It's a fair question, and I'll read this note we got from Jason because I think he sums up the other side of the argument. He writes:

“I'm noticing a lot of similarities between the AI spending races and the streaming wars six to seven years ago. Back then, content makers saw the growth of Netflix and convinced themselves they needed to compete with their own streaming services. The only studio that sat out that spending spree was Sony. Cut to today, where Sony benefits from having a large number of distribution platforms bidding on their content while saving money from not running and marketing its own streaming platform.

“Similarly, while the incumbent tech giants and pure AI labs spend hundreds of billions of dollars combined annually on AI CapEx, the latest earnings call showed that Apple is sitting this trend out, spending a meager $4 billion in the next year. While many will be disappointed, I think this was a bold and courageous decision to go against its investors' expectations.

“While the animal spirits are saying, ‘Spend now or miss the next generation’ (bubble alert), Apple is realizing that their superior hardware and distribution leverage mean they can give their customers the option to run the best models on their devices without having to spend money of their own, in Apple's case.”

I understand that argument. I would feel better about embracing that take and celebrating Apple's contrarian impulses.

Apple is zagging while everyone else zigs here, but we're also watching Apple's core competencies erode in the midst of all this, in addition to falling behind in AI. So I'm wary of trusting Apple to figure it out. But did you have any thoughts on Jason's take there?

No, I've made the same take, and it's very compelling from a business perspective. But I think implicit in Jason's take, I would assume, is they're partnering with someone else to help provide this.

All right, and that is the end of the free preview. If you'd like to hear more from Ben and I, there are links to subscribe in the show notes, or you can also go to sharptech.fm. Either option will get you access to a personalized feed that has all the shows we do every week, plus lots more great content from Stratechery and the Stratechery Plus bundle. Check it out, and if you've got feedback, please email us at email@sharptech.fm.