Pod Insights Sharp Tech with Ben Thompson (Preview) Today’s Internet and Tomorrow’s AI Innovation, Progress Reportedly Slowing for New LLMs, The Vision Pro as a Productivity Device About this episode An emailer wonders whether 30 years of Internet investments and data were the bootloader for an AI transformation in the real world. Ben offers his take on recent reports that OpenAI and Google are seeing diminishing returns from their latest LLMs, and the arrival of Ultrawide capabilities leads to refined takes on the Vision Pro and advice for Apple moving forward.
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Andrew Sharp? Hello and welcome to a free preview of Sharp Tech.
0:08 Andrew Sharp? Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp and on the other line, Ben Thompson. Ben, how you doing?
0:19 Ben Thompson? Doing okay, Andrew. How are you?
0:20 Andrew Sharp? I'm doing all right. It's nice to be back, you know? No mailbag on Monday. It feels like it's been too long, so it's good to see you again.
0:27 Ben Thompson? Yeah, there is actually an oddity. I mean, first off, Happy Veterans Day to any veterans in the audience. But the Shtetkery calendar follows the stock exchange calendar. So when the stock exchange is closed, that's when I take a day off. And Veterans Day is odd in that I think it's the only federal holiday that is not where the stock market is open. And so, I mean, you're in DC, you care about the federal government. I'm online, I care about the financial markets. And so, because you said, oh, it's gonna be a holiday. You know, we're gonna do something. Is it okay if we take off? I'm like, yeah, of course. And I'm like, why am I writing? Like, ah, that's why.
1:08 Andrew Sharp? I wondered the same. Yeah, no, I mean, there are certain people who are working on Monday, certain people who weren't. That's a fun bit of trivia at the top of the show.
1:17 Ben Thompson? I know, I know, because I was actually kind of confused. So I had to go back and look it up. I'm like, why did I not have this scheduled as a day off? Because I have to, I have to do the days off in the federal holidays because I'm a maniac and then I would just feel guilty. Like, it's like a forcing function. Like, look, you have to actually take some days off. And so I'm like, why am I working today? Ah, that's why, that's why. So there you go.
1:38 Andrew Sharp? There you go. Learn something new every day. Well, as for today's show...
1:44 Ben Thompson? The big question, Andrew, that we need to discuss is, why does the stock market not honor our veterans? That's the question people are asking.
1:49 Andrew Sharp? A great question. Something for the audience to ponder for the remainder of the week here. You know, maybe write your representatives and let's be the change that we want to see in the stock market going forward. But for now, we have three very different topics as we return today. We're gonna start with AI. We're fixing the Vision Pro. In the middle of the show and then at the end, we'll hit a few Elon emails. But we'll start with Benny who writes, "What if the internet itself between 1990 and 2020, with Web 1.0, Web 2.0 and mobile, was the bubble or frenzy phase in Carlotta Perez's model?
2:29 Andrew Sharp? And what it brought us for the future was not clean or cheap energy in the form of new nuclear plants, but training data? Honestly, most B2B SaaS is kind of irrelevant with questionable profit outlook even after a decade of Zerp, etc. But what it brought us was cloud infrastructure at scale, which led to AI factories on top of that. Most YouTube content is irrelevant, but it brought us video data at scale. Most social media is irrelevant, but again, training data. Most blogs are irrelevant, but training data. You get the idea. So what do you think of the internet economy as merely a bootloader, infrastructure for AI? That now lays the foundation for a jump back into tech transforming
3:14 Andrew Sharp? the quote unquote "real" economy with Tesla AI, robotics intelligence, spatial AI, materials design, biological manipulation of proteins, etc. So Ben, I like that email and I recognize that line of argument because it's an idea that you threw out in the midst of a digression at some point in the last month or two. What do you think of that possibility?
3:40 Ben Thompson? Yeah, I mean, yes, I have thrown that idea out. I think that's something me and Nat and Daniel have talked about in our sort of AI series at points about the internet being a bootloader for AI. I think this email gets a little too cute about the concept. I mean, there's a difference between lots of companies leveraging SaaS applications and the question of like SaaS company valuations in the Zerp era versus now. Like the reality is, is Stratechery, a relatively small company, we pay lots of SaaS providers because they make our job easier and they're useful services.
4:19 Ben Thompson? And YouTube, I think YouTube's pretty compelling and pretty relevant and like this is, there's shades of the old like.
4:28 Ben Thompson? You like the long run, like one of the reasons why it's been, Facebook has been such a useful company for stratechery is not only have I been able to write about it, but people are just have, I feel like one of the reasons people have always been inherently skeptical about the company is because there's just a, what's the actual utility here? Well, it turns out the utility is talking to other humans, which is like the most important thing to us. And it's actually extremely important and valuable. So there's a sort of deductiveness to this argument that I think it's, it's kind of like, you know, I do need to know, is Benny a college student and how potent is the weed, you know, sort of perspective to this. So, no, I'm just kidding, Benny. But you could like get these theories and spin
5:12 Ben Thompson? them up and everything's pointless. You know, it was all just about AI. So I sort of reject that point. And I don't think that computers or the internet have been a bubble, right? Has the car industry been a bubble just until we got to like self-driving cars? Like when we get to teleportation, was everything just a waste of time until then? So yeah, so I, I sort of pushed back against the premise of the question. I, the reality is, is there's been real value created. And even with this AI stuff and the stuff that's being built and funded, it's being built off of real cash flows, off of real businesses. That's not a bubble. And even the productivity measurements, there are productivity gains. They've tended to lag. Like there was a big question in the 80s, 90s, where's
5:57 Ben Thompson? the IT productivity gains? Well, they showed up and it just took a while. And it's sort of a similar thing with the internet. And there's all these sorts of things about the excess consumer value that's sort of hard to measure. Entertainment by and large and access to all this information is free. How does that actually show up in GDP? And yet it does sort of exist. There's, there's lots of like technical questions you can get into about this, but I think it's going too far to say that it's all, you know, fake. So.
6:25 Andrew Sharp? Okay. Fair. I mean, I liked it for two reasons. One, I think it's a good coda to the conversation a couple of weeks ago about meta AI and the potential downside. For the human experience and what the next fifteen or twenty years might look like. Because there are also all these possibilities in the quote unquote real economy as he describes it with AI Teslas and robotics and. You know, biological manipulation of proteins like all of that stuff excites me more than the possibility of AI content integrated into my feed without me even realizing it.
7:02 Andrew Sharp? So I like that his focus was there. And then also, I mean, I do look around at where we are now. And it's like, there are some pluses and minuses to the transformation that we've all undergone over the last thirty years. It's plausible to me that we could look up in another thirty years and look at this recent phase as being sort of a transitional phase. And the real leaps forward that like improved the world and improved everyone's quality of life. Happened with AI being adapted for the real world and transforming how we live and interact on a daily basis.
7:42 Andrew Sharp? Whereas I think a lot of the transformations in terms of how we live and interact over the last fifteen or twenty years. Things have gotten not necessarily like definitively worse, but certainly more complicated in terms of how humans interact with technology. So I like this glass half full outlook for us.
8:01 Ben Thompson? Well, I mean, well, even just if you want to take a glass half full sort of looking backwards. Yes, stuff is different. I think it's overstating it to say that, you know, there's just been, you know, and we we've talked plenty on this podcast about the sort of complications and issues of online. And we both, I think, kind of agree that right now is not the ideal state of affairs like, and so it's not. Index on the way things are right now. But at the same time, there's lots of things right now that are that are pretty great now. I always hesitate to go here personally because I'm obviously in a unique situation being halfway across the world.
8:38 Ben Thompson? But I would say us sitting here twelve time zones apart and recording a podcast, which people all over the world are going to listen to and hopefully feel that. Their life is better, even if it's just enjoyment.
8:50 Andrew Sharp? Us becoming friends, us becoming business partners, us talking all day long. All of this is facilitated by technology that didn't exist 20 years ago.
9:01 Ben Thompson? For sure. And I think a lot of the intervening stuff is, a lot of it is around sort of mores, right? Uh, we, we were in a, a group chat conversation yesterday and you were sort of bemoaning the loss of quote unquote, old Twitter where people would have actual sort of conver- maybe it was you, maybe it was someone else. So they're like, you have conversations.
9:19 Andrew Sharp? And we agreed that that version of Twitter is, is overrated and may have never really existed. It may have existed for like a year or two, but yes.
9:28 Ben Thompson? If I could borrow a phrase from Ben Goliver, is there a mouse in your pocket? Uh, it did exist. Of course it, no, it absolutely existed. Like early Twitter, like to 2007 to 2013, 14 or so. Actually, I even put it to 2015. There were absolutely real conversations online. You would go out there and you'd go back and forth and people would be observing. People would chime in and it was like the world's largest group chat. And that was absolutely the case and it was a great experience. The reality of the internet is the scalability of everything because of the zero marginal cost nature of sort of accessing these services. Means that doesn't last. You have a tragedy of the common sort of effect where you're going to get all the trolls.
10:10 Ben Thompson? You're going to get all the people interpreting stuff in bad faith. You're going to get people that want to leverage it for, you know, to attack people for other reasons, maybe real world reasons, wherever it might be. And yeah, that version of Twitter definitely does not any longer exist, but no, I will push back 100%. It did exist and it was great. And the reason.
10:28 Andrew Sharp? I'm going to push back on your pushback. It, it was overstated. It was fine. Uh, but people have been complaining about Twitter for literally like 15 years at this point.
10:37 Ben Thompson? When did you join, when did you join Twitter?
10:38 Andrew Sharp? 2009.
10:40 Ben Thompson? Okay. I just had to make sure this wasn't a, I joined in 2015.
10:43 Andrew Sharp? I was in there on the ground.
10:44 Ben Thompson? I was a little dark.
10:45 Andrew Sharp? Darren Robel was like a celebrity back in the day, but yes.
10:49 Ben Thompson? Well, I mean.
10:49 Andrew Sharp? No, Twitter was fine. I understand what you're saying. I'm just, I think there's a nostalgia to the discussion that uh, sort of allied some of the toxic aspects of 2012 Twitter.
11:00 Ben Thompson? Yeah, that might be the case. It also might be the case that this was different areas were in different cases. I think there's a bit where the sports arena has always been pretty toxic. Like even whereas I'm thinking more of like tech Twitter back in, back in the day. But I, I think we could come to common ground here. Sports online commentary has always been extremely toxic. So I think that was probably, it was the case on message boards back in the day. That's right. Yeah. So, no, okay, I see where you're coming from. That was, that was, that was your primary ground. So, fair enough. And, uh, but I, I think that was definitely the case on tech Twitter anyway, that it was very collegial, very interesting. And I can absolutely understand people missing that. And it's inevitable that it sort of went away.
11:46 Andrew Sharp? 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 Strategery and the Strategery Plus bundle. Check it out. And if you've got feedback, please email us at email@sharptech.fm.
Previous (Preview) What Trump 2.0 Could Mean for Tech, Questions on Antitrust, EU and Growth, Elon Musk and the Election Next (Preview) Monday Mailbag: Netflix and Its Mike Tyson Adventures; WBD Settles with the NBA; Lessons from Building Passport Page updated November 6, 2026 · transcript and summary are machine-generated