AI存在性恐惧的文化根源｜Ben Thompson 主持的《Sharp Tech》

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, Ben Thompson. Ben, how are you doing?

Horrible. Absolutely horrible. I'm sorry, I'm going to have to rant. It is ruining my life.

Hmm. The delivery windows?

UPS and FedEx. Seriously, it's unbelievable. Okay, where do I start? This is a building problem.

Good.

So, number 1: last week, iPhone week, right?

Mm-hmm.

I know they're very busy. I get it. I was home alone, and I had to go pick up my son from school. Of course, they came in the window when I wasn't there. I'd heard they are now just leaving iPhones. I guess not in my neighborhood. It's fine.

Yeah.

No, it's fine. It's fine. I get it. It's an iPhone, one of the best resale items ever. I get it. No problem.

Sure.

So, on the next day... Did I complain about this already? I can't remember.

No.

Anyhow—

You complained about it privately—

So—

—but I'm glad—

Okay.

—you’re taking it public here.

No, no, it's building. No, it's building. So, yes, I was not going to complain about this, but it's building. On Saturday, I wake up early because it said it was going to be there at any time during the day. Then I open the UPS app, and it says, “We'll be there between 12:30 and 4:30.”

Mm-hmm.

So I'm thinking, “Oh, the specificity.” I'm so beaten down by these companies that I'm happy it reduced a 12-hour window to a 4-hour window.

Give me a 4-hour window. I will be available, sure.

No, I canceled an appointment that I had in the afternoon. I made sure that I was going to be there, and because my wife was flying in from Taiwan that evening, I had to drive down to O'Hare. It was a late flight.

Mm-hmm.

I knew I'd be up late driving, so I thought, “I'm going to need to take a nap,” because I woke up early.

Yeah.

I can't take a nap in the afternoon because I don't want to miss my delivery. So I'm going to take a nap now at 10:00 a.m.

Mm-hmm.

So I lie down and take a nap. Great nap. Actually, one of the best naps I've had in a while, right up until I wake up at 11:00, and it says, “Your UPS driver came at 10:30 to try to deliver your package. You were not there. And if you don't get it the next time they come, it's going back to sender.”

Oh, God. Wow.

The sword of Damocles hanging over your head here with this iPhone.

I got ahold of someone. I'm very mad, very upset. They're like, “We'll have the driver circle back around. We'll have someone from the ground. We can't contact the driver directly. Someone there will call you within an hour.”

Mm-hmm.

No call. I call back 3 hours later. They're like, “Oh, it says on here you're scheduled to get a call on Monday morning.”

Oh, God.

Anyhow, I do get the iPhone on Monday. We can get to the iPhone. I have an iPhone review—

Mm-hmm.

—which we can get to later in the show. Foreshadowing.

So, anyhow, this is just dangling. I'm kind of annoyed about this. You might have heard I built a new rack.

Yeah.

It's a nice rack. It doesn't have my new server in it, which is obviously the critical part of the rack, to say the least. It's been a little delayed. It's finally on its way. It was supposed to ship last Monday. They didn't ship it until Thursday. I've been tracking this. Whatever. My server is traipsing around the state of Illinois right now. I'm not sure exactly why.

Okay.

It was supposed to be delivered on Wednesday, then it was delayed to Thursday. Now, apparently, it's delayed to Friday. You have to sign for it. Someone has to be here. Guess what I'm not covered for—

So you're just standing watch—

—deliveries on Friday.

Oh, my God.

Anyhow, no one cares, other than: bring me the AI drones. Bring me the AI customer service.

Automate all of this, for the love of God. Yes.

There is a—everyone is anti-AI right now, broadly speaking, in the general public. There is an aspect, though, where corporate America has so seeded the ground for AI. I saw a tweet about someone bemoaning the end of customer service.

Mm-hmm.

There's an angle to Muse. People are saying, “Well, it's cool Muse is finding you discounts and refunds, but once all this stuff happens, it's just going to become much more adversarial between companies and their customers.”

Mm-hmm.

And I'm thinking, “We're already there.” These are my 2 mortal enemies. I despise them with the fire of a thousand suns. I don't care what happens to them.

I hope Muse finds every single dollar they owe every customer—

—it does have a happy ending, because you did get the iPhone and you hated it, which, as far as I'm concerned—

I did.

—is wonderful.

1. The iPhone Air Wins Again

I have to completely retract everything I said last episode. I said, “Oh, that 1 day you need battery is worth the other 364 days.” That is totally a lie.

By the way, I didn't realize the 18 Pro Max looks exactly the same as the 17 Pro Max.

Mm-hmm.

Substantially heavier.

Interesting.

Yeah. And actually, I don't have the numbers in front of me, but every iPhone has gotten heavier and heavier since the iPhone 15 Pro, which I think was actually the best one of this generation. It's a chunky phone.

Thiccc, with 3 Cs.

Well, the Air exists. If that's what you care about, Apple sells you an alternative. I think we'll get the Air 2. I don't know if we'll ever get an Air 3. You know what I'm going to do with the Air 2?

What? Buy it in person?

Not get it delivered.

Drive to Chicago?

I'm going to be standing in line. No, no, there's an Apple Store here in Madison.

Okay.

Look, we're not that backwards. Give me a break.

I wasn't sure. Okay.

I will be in line to get my Air 2. I cannot wait. Also, look, the burden is mine.

Mm-hmm.

If you bought an Air, you're doomed to forever want that phone—

Yeah.

—not anything else. But I gotta tell you, 18 Pro and 18 Pro Max users, your phones feel cheap.

Ooh.

Throwing shade at the 18 Pro Max community.

I'm just saying, the Air—it's not just that it's beautiful and so light and so thin.

Uh-huh.

You get the sense it can take a beating, and it can take a beating. I've dropped my phone multiple times. It's solid.

Yeah.

The 18 Pro is like this flimsy aluminum—

Mm.

—blah, blah, blah.

Well—

I mean, it's fine if you want a cheap phone.

The 17 Pro is probably my favorite iPhone I've ever owned, in part because the 16 was such a piece of crap, so it felt miraculous to upgrade to the 17. But I'm just thrilled that your love affair—

Wait—

—with the Air is alive and—

The 16 was a piece of crap, so I upgraded to the 17. What a privileged statement that is.

Well, all I can say is the battery life is better, and the camera is fantastic. It's really the camera for me. I could never go to the Air because I need the best possible camera.

No, 100%. If you have little kids, you should be getting the Pro or the Pro Max. Like, the—

Yeah.

—I'm sure the camera's spectacular.

The Duo piqued my interest, but I'm not settling for a worse camera at this—

That's right.

—point in my life.

Not only should you have the absolute best camera, you should also upgrade every year, because that's the one piece that generally gets better every year. And you get one opportunity to take pictures of your kids at the particular age they are. So this is the beautiful rationalization: you can justify getting a new iPhone every year, and I think it is a legitimate justification. It's one of the few that actually exists.

There you go. I'll take it.

Look, no one is better at coming up with justifications for spending money on electronics than me, but this is one that I think is actually a real one.

Makes sense for everybody. Absolutely.

That's right.

Well, all is right in the world. You're back in the fold as an Air evangelist. As for the show today, we will begin with Meta Connect. Meta Connect was held Wednesday night in Menlo Park.

I've already made my request: stop having events on Wednesday. Please accommodate my publishing schedule. Wednesday night? Come on, Meta. At least do it during the day.

Mm-hmm.

Reasonable request, I think.

Yeah.

Hey, listen, I had a good time reviewing the Meta Connect event at 1:00 a.m. as I was lying in bed. There was a lot going on out there, okay?

Well, what was great about it was that it was 50 minutes.

Yeah, nice and tight.

In, out. Tight.

Live audience.

Yeah.

Zuck doing it the old-fashioned way. Love to see it.

Yep.

Dude, live is so much better. I was going to watch it in the morning, but I just started watching at night and was locked in. I watched the whole thing. Why? Because live is compelling. Apple puts out these marketing videos, and I'm drifting off into space within 30 seconds. Live is just better.

They actually pulled it off with all successful demos, so congratulations to them. I think that's the first time they made it all the way through a keynote without a failed demo. But yeah, 50 minutes. Love it. Great. Live.

Not Wednesday night

Well, and Muse mania is also sweeping the United States as we speak, so we have a lot to wrap our arms around with respect to Meta. Let me start with a general question, though. Facebook changed its name to Meta on October 28, 2021. If we consider the past 3 weeks as one moment in time, is this the first time since the name change that Meta has officially been cool? What do you think?

Ooh. There's a point where you can lean too much into it, and the cool thing becomes cringe.

Mm-hmm.

That's a good observation. I mean, I think if you're an investor, Meta was pretty cool from 2022 or 2023 onward. Probably. I mean, it's the first time they've probably been cool in a very long time. It's a good question. I don't know. What do you think? You're the normie here.

I probably do have a better sense for what's cool. Zooming out, the metaverse was a creepy disaster that almost instantly became a nationwide punchline. The headsets never really moved the needle. Llama was a cool idea that never gained any real traction. The glasses are a hit in relative terms. It's not a Google Glass situation, but I wouldn't say they've swept the nation or captured anyone's imagination, really.

I think of our past podcasts after Meta keynotes, where you would come on and say, "You know, I really admire this risk that they're taking. They're committed to it, and some of the innovations in the glasses space are pretty cool."

Wait, did I say that?

Well, you've generally been pro what they're doing with that initiative.

Look, I don't want to—

This is mischaracterized.

Look, I've had a bad week, okay? I was on TVPN, making the point that—you might not know this, Andrew, but I'm very enthusiastic about agents.

Mm-hmm. I've heard.

You may have heard about this—

And racks.

Yeah, on and offline. And racks, and hardware to run my agents and my own purpose-built agent for my thing that I've spent weeks and literally blown past my subscription multiple times. I've spent real money on this. I believe in agents, to say the least.

When Muse came out, I said, "When Muse came out, this is the same day as OpenAI solved that math equation." Or, sorry, that sounds so trivial.

Navier–Stokes.

One of the—

Yeah.

Yeah.

One of the Millennium Prize Problems? Sure.

No, one of the math questions that is so significant. I knew about it well before this AI moment. It's a super big deal, to be clear. But I made the case, like, look—

This is a bigger deal. Manus is a bigger deal.

This is a bigger deal.

Yeah.

2. Agents Eliminate Everyday Hassles

Just what it represented: giving literally everyone a computer. That's the core of it. It's not just AI; it's actually giving people a computer and an entity that can use that computer on your behalf.

Mm-hmm.

It's actually incredible. I have told everyone in my regular life that they should try Muse, just because I believe in agents so much. I think they're so helpful and so useful.

What I don't think is useful is trying to sell people on productivity. People don't want to be productive. Enterprises want to be productive. What you sell them on is making annoying stuff in their life go away, making life more convenient. This is about a marketing message.

Eliminating hassles is what people will buy.

So I make this point on TVPN, and I'm talking about the framing of this question.

Mm-hmm.

Somehow I get clipped, and it's all over the internet that I think agents are stupid and pointless, and that Muse is a failure. It is so aggravating. The clipping thing is out of control. It's fun to go on these other podcasts, but it makes me not want to go on them. I don't blame the TVPN guys, but it's really annoying. This happened when I went on the Pat Roach Honesty thing too. The clipping is just—

Every couple of hours, I see you responding to someone new on Twitter, saying, "Actually, I've written the exact opposite of what you just said I said, so stop mischaracterizing me." I don't want to mischaracterize you. As far as Meta Connect, all I'm saying is—

Anyhow. What I'm saying is—

After—

What I'm saying is, no, stop.

Okay.

What I'm saying is, as annoying as this week of clipping has been, do you know the basis on which I met Mark Zuckerberg for the first time?

I don't. Enlighten me.

This was in 2020, the week before everything shut down. He invited me in, and the reason he wanted to talk to me was that he wanted to understand why, for 6 years—however long it had been—I had been relentlessly opposed to what they were doing with the metaverse and Reality Labs.

It actually increased my respect for him a ton. He didn't want to argue with me. He did argue with me, but it was a Socratic sort of, "I want to really understand what you think."

Mm-hmm.

I obviously didn't change his mind about it at all, but he said, "I just want you to come and meet because I want to really understand your thinking about this. I want an outside perspective. Why do you think I'm doing the wrong thing?" We went back and forth and had a good discussion. Then he's like, "Okay, great to meet you," and kept doing the same thing. Which, by the way, is very Mark.

Yeah.

Which is fine. All that to say, my discomfort with and issue with Meta doing this whole thing, spending all this money, and doing hardware when they're a horizontal services company that should be serving everyone—

Mm-hmm.

—and hardware diverts you and sets you into competition with folks that should be your partners. All of which happened, by the way.

And then you get on this podcast and say, "You, who have been defending their glasses effort all along." It's unbelievable. I didn't even get clipped.

No.

You clipped me on the other podcast.

Look, all I'm saying is that on those podcasts after past Connect events, I've felt like the wet blanket who has to come out and say, "Look, I don't think anybody's actually going to use these," and that there's not anything real here.

Right. I've been enthusiastic about the technology.

Correct.

I've always had a problem with Meta specifically doing it. In fact, I think there's an argument that Meta actually killed the technology by virtue of bringing all the talent in-house. It didn't develop organically. It didn't develop in the broader ecosystem. Anyhow.

But after this year's Connect—

FedEx.

UPS.

Andrew Sharp. TVPN.

Great.

I'm like Aria Stark right now. I'm going to take all you guys out.

And lots of random people on Twitter who've been driving you insane for the last 72 hours.

3. Meta AI Makes Glasses Matter

I was sitting there last night thinking, "Holy crap, they nailed it with the AR glasses." What did you think of Meta's VR headset? I look at this now and I'm like, "They're nailing it on all these different vectors," and I'm wildly impressed with where the company is.

Only a couple months after everybody was freaking out, saying, “What the hell is Meta doing?” you wrote a script for Mark Zuckerberg to get his messaging on point, and now here we are a couple months later: it’s the hottest company in technology.

There’s a sport you might have heard of called basketball.

Mm-hmm.

Yeah, in basketball there’s this interesting dynamic that happens when a player is injured or when a team acquires a player, say, at the deadline. You can have a team that seems woefully deficient, and you add one more player and they get way better—not just because they added that player, but because everyone slots into a position—

Slots into their proper role.

—that’s much more suited to them.

Yep.

Right? Of course, you know what a number one needs to be. Everyone knows it’s coming, and they still score. They draw the defense and kick it out. You need a number two who can carry the offense with the number one on the bench, who can be a secondary ball-handler. Depending on whether number one is a distributor or not, if they’re not, number two needs to be a distributor. If number one is more of a distributor, number two needs to be a scorer.

Mm-hmm.

Number three is basically the guy you can put on the best player, who will knock down every single open shot. Then you have the dirty guy who does all the dirty work—

Sure.

—grunt work.

It’s like, imagine the Knicks without Carl Anthony Towns. Carl Anthony Towns is their second-best player, and then he allows their role players, like OG Anunoby and Mikal Bridges, to slot into their proper roles, as opposed to asking those guys to do too much and inevitably being disappointed.

Exactly.

So the problem is that if you’re missing your number two, then your number three seems worse because he’s a bad number two. Your number four seems worse because he’s a bad number three. It goes all the way down the chain.

Yeah.

You can slot just one guy in, and a team gets massively better because everyone is now in their proper position.

Mm-hmm.

Meta’s problem is they were missing a number one.

Okay.

The reason the glasses seem compelling is not just because they’re small and thin, or even because of the VR ones. I thought it was a great reveal. Actually, they’ve been sitting out here the whole time, and you didn’t notice—

Amazing.

—because they—

Yeah.

—they look kind of like— That was the best part, just from a showmanship perspective, because he kept saying, “Oh, look, here’s all our glasses. Here’s all our glasses.” He was setting it up the whole time. Once you have Meta AI, and Meta AI is the thing, now you’re not selling VR glasses, you’re not selling AR glasses, you’re selling Meta AI delivery devices.

Mm-hmm.

It’s so obvious. Muse is something that wants to break out of the phone. You see it in your little app, up in the little corner, typing away and doing work for you. It’s like, “Oh, poor little Muse is stuck in my phone.” You’re not thinking this clearly, but you’re getting a sense of, “I would like Meta AI in more places. How can I get it in more places?” Oh, Meta has devices on which I can interact with Muse. It isn’t theoretical AI glasses.

Mm-hmm.

These are now Muse glasses. Your little Tamagotchi is a Muse Tamagotchi, and that is what— That’s why the keynote comes along and feels so focused. It’s why you feel so excited. It’s because they had AI as their number-one player. Also not a number-one player, right?

Mm-hmm.

Your number-one player is a product that resonates, and the product doesn’t need to be pushed on people. People pull it because they want it. Once you have that pull effect, everything else slots into place. Muse is their number one. Their AI models are number two.

Mm-hmm.

The VR glasses are a great number three. Klay Thompson is not a number two. Terrible number two.

Yeah.

Can’t dribble the ball. The best number three in history. Locks down opposing guards, doesn’t miss a shot. That’s why the glasses feel better now.

4. The Vision Pro Already Lost

And as tech, the glasses also look better. They weigh 100 grams, the same weight as a deck of cards, Zuckerberg pointed out, and they’re set to retail for $1,300. I would call them a Vision Pro killer, but I think the Vision Pro is already effectively dead.

Killed itself.

But they’re a third of the price and basically have all of the performance that you’re getting with the Vision Pro. Vision Pro is real 4K, and these aren’t quite as high-resolution, but it’s just a super impressive product and a lot of what we’ve imagined should exist in our conversations about the Vision Pro over the last couple years. It’s not nearly as heavy, and it’s an entertainment device. They’ve secured sports rights. They’ve secured football and UFC. I’m curious—

Yeah.

—to see what it turns into.

I think it’s a little oversold. I think ESPN has 180-degree cameras at some of these games.

Mm-hmm.

So it’s not like the immersive camera stuff that Apple is doing.

Yeah, okay.

It’s going to be more like if you go on YouTube with your Vision Pro and can access some of these 180-degree videos, which are still super cool, to be clear.

It’s cool, yeah.

But, yeah.

Not quite as transformative and transportive as Apple’s.

Well, that’s what’s so frustrating. Apple went to the trouble to make these incredible cameras that actually deliver this fully immersive thing—

And they don’t really use them.

—and then they just suck at it.

Yeah.

So, yeah. We’ll see what— And we’ll see how it works. I’m wary also of it being delivered by a company that is focused on 2D video, because the whole point is delivering a 2D experience when you’re wearing 3D glasses—3D, immersive sort of things. So we’ll see how that works. But, yes, to your point, for what the Vision Pro is actually used for, if it’s used, I do want to see the resolution. You could argue the 4K, or the resolution on the Vision Pro, is amazing.

Mm-hmm.

It’s also still not good enough.

Mm.

You could pick that up with the immersive video things. So we’ll see how these work out. Would a little bit bigger with more resolution be better? I don’t know. Also, they’re only $1,200.

Yeah.

Only $1,200, all things considered. Those are great, but then everything else makes sense, like the glasses with no camera.

Mm-hmm.

Number one, way less creepy, thank you. That’s good. The fact that you can talk—they’re Meta AI devices. They’re Meta AI delivery vehicles.

It’s all coming together.

Muse delivery vehicles make so much more sense.

Yeah. The Muse charm is going to be ready as early as the holidays. I believe that’s when they’re trying to have it ready. Speaking of Meta AI—

Oh, yeah, by the way—

Okay.

Yeah, the Muse charm—there was an article in Bloomberg about how this is what Alan Dye did. Complete BS, by the way.

Okay, good note.

There is some drama about how he’s already doing at Meta everything that everyone hated him at Apple for.

Well, out of curiosity, do you think that’s Alan Dye driving that confusion there? How does that happen?

That’s what Alan Dye does. That’s how Alan Dye became the VP of Industrial Design at Apple: by being terrible at his job.

So, good reputation manager Alan Dye. Good to know.

Yeah.

Do you think failing with the other product launches, or failing in relative terms with past product launches, has helped Meta refine its product instincts with respect to Muse? I’ve been pretty stunned at how well they’re executing across basically every vector with this product. The ads are great. The character is memeable and wonderful and perfect for the internet right now, and they’re shipping new features literally every day. They’re developing new partnerships literally every day, and it’s all come together and looked pretty seamless, which is not how it’s felt in past eras of Meta, at least from my perspective.

You get what you pay for.

Okay. Just bringing in talent—

Um—

—that knows how to do this stuff and has good instincts?

Yeah. I’m biased because this is a previous Stratechery subject, but I don’t think it’s an accident that Mark Zuckerberg called out Nat Friedman by name. I think anyone who’s worked with him knows he’s one of the greatest operating executives in tech.

Yeah.

He swore he would never do it again.

I don't know what it took to get him back, but whatever Mark paid to get him back was the bargain of a lifetime. Honestly, there's just an aspect to that that is real. If you're a big company like Meta that has to deal with all the internal politics, like the Alan Dyes of the world, like Meta does, you don't have people like that. They left long ago, if they were there.

Everyone who's still there has their own little kingdoms and fiefdoms in their own right. And so this is all ultimately down to Mark Zuckerberg, for sure.

Yeah.

The pivot, like a year ago, a year and a half ago—I did that interview with him, which I've referenced multiple times, that was bizarre. He was out to sea.

Mm-hmm.

And to his credit, he didn't know how to solve it, but he realized he was out to sea. And in the end, he basically fired their whole AI team, brought in a completely new AI team, and didn't just bring in an AI team; he brought in a product team.

Mm-hmm.

And they delivered the first new product from Meta since—

That's worked.

—the original—

I know.

—that they didn't buy.

We're only 2 weeks in. Let's not overstate it. Let's put in all the caveats, but the reason to be hyped up is, number 1, the momentum is obviously real.

Mm-hmm.

You could see it, feel it. Number 2, this category is real. I promise you it's real. The hard thing is selling it and getting people to understand what they can do, and Meta AI delivers on the promise of this product by and large. Of course, there are rough edges, for sure. We're at the very nascent edge of it, but there's a reason for it to be taking off.

Yeah.

And then number 3, the implications of winning this space are astronomical. This is the winner-take-all of consumer everything. This is the aggregator of aggregators. You reduce the connection to the human to one cute little character that fits in a Tamagotchi and mediates everything online.

Mm-hmm.

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.