(Preview) LLM Utilities, Google Disruption, Remote Work Angst, Family Tech Support, and a Thanksgiving Mailbag

Celebrating the holiday week with emails from subscribers. Topics include: integrating LLMs into the modern workflow, human reactions to AI content, Google and its disruption risks, a follow-up to last week’s trade discussion, advice for a young engineer working remotely, creating new reality shows, an unpaid shoe testimonial, family tech support, and more. Happy Thanksgiving to all!

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Andrew Sharp?

  Hello and welcome to a free preview of Sharp Tech.

• 0:08Andrew Sharp?

  Continuing on here, Ajike says, guys, fantastic episode on America and the trade status quo. At one point during the show, Ben said in passing that it's easier to quote, "Develop up than it is to develop down." In other words, someone lower in the value chain, like China, can develop capabilities higher in the value chain, whereas it will be harder for Silicon Valley to learn how to build the lower value chain stuff efficiently. This sort of makes sense and seems to track historically, but I can't put my finger on why that is the case. Are there structural reasons why this is a rule? If so, I'd be interested to hear them. Any thoughts, Ben?

• 0:53Ben Thompson?

  Well, I think part of it, it kind of goes back to that Google bit before. There is just a financial and margin component to this, where if you're used to having high margins, it's very hard to go back to being low margins. And you know, one of the reasons why no one responded to AWS for a long time is everyone just assumed it was crappy margins. Like, yeah, nice play by Amazon. They can handle the low margins. They don't make any money on their store either. I mean, you know, it's like, and then when the results came out and it was like 40% margins, suddenly boom, everyone's like, I can handle that. I could do 40%. Yeah. So there is, there is an aspect there. There is a, I don't want to overly speak to the rest of the economy.

• 1:35Ben Thompson?

  Like, we went fairly out on a tree branch with that episode. And so I'm wary of it snapping. We've played our luck fairly well so far. But there is a bit about tech where, and this was a bit I was trying to capture in that podcast, that article, is tech is addicted to scalability. And, and this is a big thing where the R and D costs for these companies are astronomical. And yes, they have like all the servers and stuff like that. It goes to the cost of goods sold, but it's all about, we can invest so much upfront because it scales to the entire world. And, and, and our addressable market size is seven, eight billion people, or sort of whatever it might be. And yeah, okay, you're a smaller company. You can only scale to 2 billion people, right?

• 2:20Ben Thompson?

  Like, it's just, this is the mindset and approach and software sort of makes this possible. And Again, this everything's interconnected. This ties into Silicon Valley salaries being so high and attracting talent. And that talent goes to these places where there's maximum leverage. And it's like crazy that your average starting engineer can can sort of go to these companies and their total compensation, once you include the stock, is in the hundreds of thousands of dollars. But it works out because the leverage on these businesses is so massive. And that doesn't necessarily exist when it comes, when you're in a marginal cost good,

• 3:01Ben Thompson?

  like making hardware or making things, the scalability is not the same. And then you just have the the basic sort of like, like the capability of manufacturing a computer or putting together what might be. It starts out, the whole point of cheap labor is the labor is a commodity. And it and it's sort of predicated on that. And you make decisions going down that road that that assumes that that is the case. And then there's a lot of path dependency. Once you've built up these capabilities, if you're building the, if you're in the marginal cost business, if you're making the computer or whatever it might be, it in many respects makes sense to hold on to all those other capabilities because that keeps your costs lower because as you've internalized this sort of cost, you're not, you're not, you're

• 3:45Ben Thompson?

  taking away the margins in between. And, you know, Samsung always talks about comparison to the iPhone. They were a pretty integrated company too, but they were integrating more in sort of the, they made the phone, they also made the screen and they made a lot of the components. And sure, they had, they were modular in the aspect, they took Android as the operating system, but their model was very, remains very effective as far as internalizing sort of other costs along the way. Apple pays all its components suppliers. They're not making the bits and pieces, they're buying the screens from Samsung or they're buying the manufacturing is done by Foxconn or Qantas or whatever it might be. There are different models and just in general, it's hard to change models. But if you're going to a model that makes you more margin, it's easier.

• 4:30Andrew Sharp?

  I was gonna say, I mean, that's, that's my prevailing takeaway from everything you just said is it's a case of show me the incentive and I'll show you the outcome. It's like, it makes sense that it's easier to go up the value chain because there's great incentive to develop some of those skills. And going down the value chain and investing a lot of resources and stuff that's going to make you less money is a harder sell for a lot of companies. Unless it's an existential concern.

• 4:57Ben Thompson?

  Yeah, exactly. Cause there's another option out there that is already doing it. It's not differentiated. There, and well, actually to the extent it's differentiated, that differentiation is based on efficiency and experience. Which you're not going to get efficiency experience when you're starting out. So you're, you're just, there's always the option to just buy the cheaper one. That's also better. Like it's not that, that's, that's not great. I mean, it reminds me of people used to wax poetic about Apple and the iPhone. Apple's so amazing. This is like talking about Roger Federer, right?

• 5:27Andrew Sharp?

  Yeah.

• 5:28Ben Thompson?

  Yeah. They're like, they disrupted.

• 5:30Andrew Sharp?

  David Foster Wallace essay about the grace of Apple phones. Absolutely.

• 5:34Ben Thompson?

  Exactly. Yeah. They, they disrupted the iPod business. The iPhone up, got rid of the iPod. They're such forward business thinkers. It's like, the iPhone sold for three times as much and had like double the margin. Not a tough decision, right? There's an internal politics bit where the, the iPod team or whatever it was had to be sidelined. Apple would tell us, because we have a functional organization. So people are always serving the, the, the better of the company, whatever it might be. At the end of the day, not a difficult management decision, uh, to like, like now, so that is sort of US tech in a, in a nutshell. It's, it's, it's not hard to go up and get a better business that makes you more money with higher margins.

• 6:15Ben Thompson?

  It's harder to go in the opposite direction, particularly when to be competitive is about efficiency, is about your cost structure. All these tech companies, their cost structures are ridiculous. They went through your Mark Zuckerberg year of efficiency, X, Y, Z. Yes, they. Slimmed down, it's like me celebrating my weight loss from like 600 pounds to 550. It's like, I mean, you know, it's like, yeah, still living like a king. Let's be clear.

• 6:40Andrew Sharp?

  Well, the one other thought I had, the most fascinating aspect of the trade discussion is one that you hit on explicitly in the article you wrote. And we alluded to it in a couple of different places on the podcast last week, but China's own industrial policies have largely worked, you know? Like they blocked Western companies from entering the market and wound up developing their own massive platform companies and software companies. And now software expertise in the EV space. And it's just, it's a sort of an interesting object lesson as America considers what the next 10 years might look like. And what some of the US industrial policies might be able to achieve or might not, as we covered last week.

• 7:25Ben Thompson?

  I feel like this happened, we've had this discussion in the context of antitrust where the two most important antitrust bits or outcomes ever, at least in the context of technology, was number one, where AT&T basically, both formal and informally, they were always sued for antitrust, like every 25, you like, you want a natural monopoly, like all the phone, like all the phone wires are.

• 7:51Andrew Sharp?

  It was basically a never ending case throughout the 20th century.

• 7:55Ben Thompson?

  That's right. And basically it's part of it. They're like, yeah, spend a lot of money on research and like Bell Labs, but just make it freely available. And you got the transistor, we got C, we the C programming language, we got Unix, so just these incredible outcomes of Bell Labs. That was basically AT&T doing R and D for the nation and giving it away so that they could keep their monopoly sort of going on. Great outcome as far as like sort of antitrust goes.

• 8:19Andrew Sharp?

  Yeah.

• 8:20Ben Thompson?

  The other one is IBM was worried about being sued for antitrust. And so in an attempt to preempt that, they voluntarily split apart the hardware and software businesses, which basically created the modern software market where you could write a program for another company's hardware. And they got sued anyway. The Justice Department wasn't, wasn't impressed. Incredible outcome. Both cases, kind of by accident, not really the point. That's how I kind of feel about China, the Great Firewall. Great industrial policy. Absolutely. Not sure that industrial policy was the driving factor in doing that, even though it's been one of the, again, of course, I hate the concept of the Great Firewall

• 9:04Ben Thompson?

  and all those sorts of pieces from sort of a political or personal perspective. From industrial policy, fantastic outcome, and I think completely accidental.

• 9:13Andrew Sharp?

  Mm hmm. Great job weaving there. I wasn't sure how you were going to tie the antitrust outcomes to China's Great Firewall, but it landed 100% for me. All right, we are now moving to the non-tech questions. We promised non-tech questions. This is actually sort of a hybrid. Max says, Ben and Andrew, I graduated this May from UW Madison and just started to watch, what's that?

• 9:39Ben Thompson?

  I said, go Badgers.

• 9:41Andrew Sharp?

  Go Badgers, absolutely.

• 9:42Ben Thompson?

  Go Badger basketball, football team's a disaster.

• 9:43Andrew Sharp?

  I was gonna say, a rough, a rough year on the gridiron for the Badgers. And just started working as a software engineer this fall for a defense company in Washington DC. Welcome to the city, Max. I moved for this job as I was under the impression I would be in person, but I got put on a remote team and I might go insane if I have to keep working remote. I knew this job was most likely a stepping stone to another company, and it was the best I could do with this job market. However, I'm now having second thoughts about a career as a software engineer as remote work culture becomes more prevalent. I'm sure this is all a bit premature since there are opportunities for software engineers in person,

• 10:25Andrew Sharp?

  but nonetheless, in-person work is a priority I hadn't considered until now. My plan is to move back to Madison. Oh, boy. Farewell, Max, and then try getting an in-person job.

• 10:36Ben Thompson?

  It's cold. You know this.

• 10:39Andrew Sharp?

  Be warm. Well, yeah, that's right. He was there. And then try getting an in-person job with maybe an early-stage company somewhere in San Francisco or Seattle. Should I be this concerned about working in person and/or with people I like? Or more focused on the actual work I'm doing? And more broadly speaking, how do you see work culture shifting for new grads? And what might be some of the second and third order consequences we see? So the last question there is a pretty big one. I don't know if we're going to be able to wrap our arms around that at this point. But do you have any advice for Max in terms of his calculus going forward?

• 11:17Ben Thompson?

  The remote thing is I'm just the worst person to talk to because I've been remote for ages. Obviously, there's a lot of talk about this online and sort of back and forth. I do feel very bad for people like Max, new college grads in particular. It's super obvious why working remote is amazing when you're established in your career and you have a family and there's errands you have to run. When you're starting out and you're just trying to gain experience and make friends and whatever it might be, it's pretty brutal. It definitely sucks. And I do worry in the long run, you know, as I mentioned, I joked about tech companies getting more efficient. There is a bit tech companies are using a lot more AI for sure. They're on the cutting edge of a lot of this sort of stuff.

• 12:02Ben Thompson?

  And, you know, is there a concern? You would start out as a sophomore, you'd get a job at a big tech company and then you would meet people and then you could get into a startup, right? How do you get into a startup from Madison when you weren't able to get a job at a Google or Facebook first and then meet people, make, build a network and then sort of go into something interesting? Number one, it sucks for Max. Is it going to suck in the long run for the ecosystem in terms of sort of getting people in, finding the next generation, all those pieces? I think those are all legitimate concerns. And there's also the bit where

• 12:41Ben Thompson?

  Is there a pulling up the ladder aspect where you have senior people, yeah, this remote thing's pretty great. Oh, by the way, once we're remote, why don't we just hire developers in India or developers in the Philippines or whatever it might be, or South America if we want to have a closer time zone. And then, oh, we're formalizing all this work, which kind of sucks. And we have to like specify everything out and communicate more clearly. Oh, I can just communicate with an AI and sort of do this, right? Like, there is, I think, there is, there's definitely a question as far as that sort of thing goes. And I don't know, I don't, I don't know the answer to it. Again, I hesitate to pontificate on this. I will say there were times where I didn't realize until afterwards that I was

• 13:26Ben Thompson?

  getting into a bad state personally because I was deficient in sort of in-person interaction. You know, this is a, to get to your broader philosophical point, I don't think it's particularly healthy that work is your life and that's where you find friends and fellowship and whatever it might be. It's also kind of just a reality of our culture in a lot of respects.

• 13:49Andrew Sharp?

  Yeah, especially if you're, if you're in a new city like Max, moving to Washington DC, probably doesn't know that many people here. Like, normal people used to just meet at the office and go out for drinks. And that was like, not an entire social life, but it was a big part of a social life. Yeah. Yeah. I mean, that, that's my takeaway is I, I don't know the software engineering market, but Max is not wrong for prioritizing in-person work experience. Because for me, when I was writing, I benefited a ton from being around other writers in an office, being able to talk to editors, one-on-one, go get coffee,

• 14:29Andrew Sharp?

  like, all of that is really useful and harder to do over email or over, I mean, it was Gchat when I was writing. I don't know how people communicate now, maybe Slack, probably Slack. But when I was a lawyer, I was also, I was trying to learn during COVID. And thank God, I had another partner at my law Firm who was the only other person at the firm who he's sort of served as a mentor to me anyways, but he was willing to come into the office and you can just learn so much more sitting across a conference table than you do replying to an email chain or something like that.

• 15:09Andrew Sharp?

  So you really do develop both hard and soft skills, working alongside other humans in an office. But also it was just, I mean, if I go back to COVID, like it was so much better for my mental health to get out of my own house and talk to humans, even if it's just like this one other partner and people at Starbucks and stuff like that. Interacting with the world is an important thing. And that's what I worry about most as remote culture comes to sort of predominate. But honestly, it seems like things are swinging the other way these days.

• 15:42Ben Thompson?

  Yeah, vibe shift is everywhere. Yeah, just a bit about the next generation and training and helping people sort of, to the extent tacit knowledge matters, is the extent that in-person interaction matters. And I think there's a chance that companies or entities or industries end up shooting themselves in the foot to a certain respect because they're so worried about their short-term comfort and avoiding a commute. They end up sort of formalizing themselves out of a competitive position. It's like, like I could get this exact, like if I'm going to go through all the trouble to be mediated via Zoom calls and Slack messages, why am I going to bother with, you know, this is sort of, oh, I guess a nationalistic patriotic

• 16:27Ben Thompson?

  American take, but like, that's not a great recipe for maintaining competitiveness. It's like, we can formalize the process, we can do all that sort of thing, we can move this manufacturing to another country where the labor costs are cheaper.

• 16:39Andrew Sharp?

  Yeah.

• 16:39Ben Thompson?

  Um, and we, there's. Sort of widespread regret about that. We've been talking about that. And here we go, just creating the conditions to do it all over again.

• 16:48Andrew Sharp?

  Or are we? Because it seems like a lot of people are saying, you know what? There are some things that we can't necessarily measure. Let's bring people back in office, bring them back kicking and screaming if we have to. So time will tell on how all of this shakes out. But Max, I hope you find someplace that will let you do some software engineering in person alongside other humans, because it's good for a lot of different reasons. And a good email. Anthony says, Ben and Andrew, I've long held the belief that if the Strategery bundle costs thousands of dollars, I would still happily subscribe. Given that I've been blessed with this large consumer surplus, I have a pretty free spending attitude toward other paid newsletters. I've really enjoyed the past few Strategery interviews with other newsletter writers,

• 17:33Andrew Sharp?

  and I've added the diff, understanding AI, and semi-analysis to my rotation. Are there any other newsletters or podcasts that you think Strategery Plus subscribers might enjoy?

• 17:46Andrew 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 sharp tech.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 at sharp tech.fm.