B (2:52)

I'm actually kind of curious what your thoughts are and what other readers are going to say. I did put it to. Put it to ChatGPT and a list of companies that came up pretty quickly actually were all the defense contractors and I think it's sort of Boeing sort of falls into that. You could make the case that actually you could think about intel as falling into that, but to what extent? And maybe this is something we won't truly see until God forbid there is some sort of large scale conflict that we find out that these entities that we're relying on that were basically granted de facto monopoly status by the US Government are actually completely rotten and incapable of scaling up or doing things beyond sort of what is possible. I mean, there's lots of criticisms about sort of current weapons platforms, how long they take, you know, and all the software problems that there are. What happens in an actual fight, do they perform well? I mean there's, there are reports that actually they do. Um, it turns out that some of the most criticized platforms turn out to be pretty effective at, at their jobs. Hopefully that's the case for lots of things.

A (4:00)

That's encouraging to hear. They certainly seem like a pretty sclerotic base right now, but maybe they're better than they seem. That would be nice.

B (4:09)

Right, well. And is there any capability of sort of scaling up? I mean, you know, it used to be. We talked about it on a previous episode. Our defense contractors back in the day were our regular big companies. And now there's been sort of this specialization and I do think sort of broadly speaking, it's just sort of, you know, I sort of did a little Year in review sort of Daily Update and it just sort of strikes. You know, I've been wondering when this moment would happen for a while. I've actually written some variation of this a couple times over the past few years. But we've gone through this era of all decisions are economic decisions. And I think that's a way to think about globalization sort of broadly. At sort of the biggest level it applies to things like aggregation theory, but in a world where all the sort of constraints are loosened, it's good for me as an analyst, you just sort of think through the economics, think through the way things like work and you can predict what's going to happen. You know, the example I use.

A (5:09)

That's right. If it's all rational, it's all fairly predictable. Right, right.

B (5:14)

And rational in a. Not in a. Like an individual's making decisions, but in a. The Inexorable sort of advantages that come to scale. For example, or like in a world we might have this beautiful romantic notion of the decentralized Internet, but the reality is without any constraints and zero marginal cost of sort of distribution, you're actually going to get more centralization, not less because the returns to being centralized and discovery are so massive and, and they sort of build on themselves. And you know, the, there's a bit that, that describes what went wrong at these companies. If you keep making sort of the quote unquote economic decision, which may be sort of deprioritizing R and D or elevating sort of your cost structures in certain ways, or your pricing or whatever it might be, you make decisions that you only come to regret 15, 20 years down the road because you weren't making the long term sort of investments. And so I would, I'm curious if our readers are going to have like killer examples of this, but I would imagine there are more than we know of.

A (6:23)

Okay, so please submit your nominations for the next Boeing and Intel over the holiday break. We can return from the holiday break with just sort of a haters rundown of different companies that people should be shorting as we head toward 2025.

B (6:39)

That's the hard thing though. That's the hard thing. I talk about the context of Intel. You would have gotten killed by shorting intel if you went back to shorting them. When I started pointing out that they have big problems ahead and that gets it right. Like I felt in 2013 when I wrote that intel was in big trouble and they needed to shift to being a foundry instead of being an integrated device manufacturer. I'm like, I'm years late writing this. Unfortunately. I just started Checker this year. So whatever, I'll just get it out now. Turned out I was, if you listened to me, you took a bath 10 years early. But I was also totally right. And that is the real challenge with a lot of these, with these, that captures the issue right from a stockholder perspective. What complaints are there? If you actually understood this, you would say, no, I'm getting in. I'm going to capture all this upside from these short term decisions and whatever. I'll just get out when the top hits and it'll be fine.

A (7:38)

Well, and also, I mean, you mentioned the defense industrial base and even intel was part of this story as well. A lot of these companies, if they make the wrong decision, they don't feel it for a while because they have a lot of structural advantages that have locked in massively Successful businesses. And that can last a decade, two decades. And so it takes a little while for sort of the cycles to catch up with them. My nomination on this question. Just reading the email from Nicholas would be almost too easy. But Nike has to be mentioned here, as we chronicled over the summer.

B (8:14)

That's a good one.

A (8:15)

Yeah, they've been slipping on tech innovation. They're not really pushing forward in design anymore. Like, I can't remember the last pair of new Nike shoes where I was like, man, those look cool. Granted, I'm not really the target audience for Nike anymore, but even young people are not like, lusting after Nike shoes

B (8:32)

the way no other shoes are. Like, are like color retro old shoes.

A (8:36)

Yeah, like designs from the mid-90s. And I am the target audience for that. But they're also seeing some of their core business disrupted by structural changes to the market where the cost structure in E commerce is just way less favorable to their business and it requires a lot of investment from them. And the board went out and hired John Donahoe a few years ago, specifically

B (9:00)

former intel board member, I believe.

A (9:02)

Oh, interesting. So there are a lot of overlapping trends with Nike and Intel. Nike also overreacted to the COVID spending and the success that they had during the COVID years and wound up going down the wrong path as a result. So they're too easy. But alongside that nomination, I do have a question for you. If we're talking like real flagship tech companies, is iOS 18 bad enough for you to wonder about Apple's long term vulnerability? Are we invoking the Charles Barkley meme? Do we need to have an Apple dialogue here?

B (9:44)

I mean. Well, I have another. I was going to actually bring this up later on. You have an email about Android and Google and stuff along those lines. And I feel like my instincts were right. It was time to go to Android and I bailed too early. Yeah, iOS 18 is horrible. I mean, I don't know that this applies to the question other than just a couple. Everyone's talking about the photos app. They did weird stuff. The control center, like the whole connectivity panel. It's like promoting Airdrop and you don't have actual access to things like Bluetooth. I had to like do all the individual ones.

A (10:18)

I saw it. You and several other people were really upset about that particular tweak from Apple.

B (10:24)

I find the connect. I mean, this is one of the ironic things. I had problems with the modem on the Google Pixel.

A (10:29)

The.

B (10:29)

The connectivity on. I don't know if this is an iPhone 16 thing or iOS 18 is terrible. Like the wifi is bad. The cellular connection, when you, like if you lose a connection in an elevator, it just doesn't reconnect. Like I have to turn it off and turn it back on. Another reason to be annoyed at the control panel interaction thing. There's some other feature or some other. Oh, the emoji situation, absolute disaster. They're cramming all this other stuff in there. I don't know. I hesitate to draw a lot. Like, Apple's had bad iOS releases before. There's a lot of talk that they pulled basically everything to focus on this Genai stuff, which is a separate discussion. And it does feel like lots of stuff sort of got collected.

A (11:12)

B team or C team putting together iOS18.

B (11:16)

Right. I mean, I don't know. Is Apple focused on. I mean, what's the evidence though? What decisions would they do differently that wasn't focused on short term results? I mean, I guess you could make the case about the generative AI stuff, like their sort of approach and not doing their own large model in particular. But that's not, it's. That's not sort of an insane way to think about it. It is, but I guess a lot of stuff isn't always insane, right?

A (11:43)

It's like arguably rational for the short and medium term. I mean, that's.

B (11:49)

They did put out the Apple Vision Pro. It might be a mistake, but it's not like they, you know, it's not like they're just not doing anything. So. But all the pieces are in place certainly, like you have a long running run of dominance that's basically undergirded by risks and decisions that were made a long time ago and sort of cemented them in place. I'm not ready to go there, but.

A (12:16)

Neither am I. Neither am I. But it came to mind and I think you could make the case that, that they're not like Apple hasn't produced anything new and great in a long time. And it's not about iOS 18.

B (12:29)

Well, you're the one who doesn't use AirPods to be fair. But.

A (12:31)

Well, well, I would say AirPods. I mean, like, that's a counterexample that people always invoke. But the AirPods and the Apple Watch are both impressive, but they're ultimately just iPhone accessories.

B (12:44)

Yeah, they're iPhone accessories.

A (12:45)

That's right. Apple was smart to release them and they're both good products. At least some people think the AirPods are good products, but they also.

B (12:53)

That's why I understand that you like the Apple Watch. I'VE just never. I've not. Every time I put it on for a few weeks I'm like, yeah, this sucks. I don't want why?

A (13:01)

And people point to their success and it's like Apple doesn't allow third parties to have access to certain functionality on the iPhone. So there's a real competitive advantage. They're not playing on a level playing field in some of those markets. And you know, where has Apple pushed forward over the last 10 or 15 years? And why should we believe that Apple has like the cultural and technical chops to capitalize if and when there's a paradigm shift in the next five to 10 years? That would be sort of the skeptics case. Again, I'm not ready to really make that case. It's terrifying to bet against Apple, but I think there are fair questions to raise after the last 15 years.

B (13:41)

Well, so is Google now in our good graces that they've blown everyone's mind this week?

A (13:47)

They have blown everyone's mind. Should we just skip to the Google question here because we got this from Saurabh he says hello Andrew and Ben. I recently attended a physiotherapy assessment and received a hard copy printout detailing my bi weekly sessions for the next five weeks. On the drive home, the thought of manually creating each calendar entry felt daunting. Once I got back, however, I decided to see if one of these new advanced LLM tools could help. Using Gemini advanced on my Pixel 9 Pro, which Ben returned, I snapped a photo of the schedule and asked it to create calendar appointments for each session. To my surprise, it not only recognized all the session details accurately, but also added them directly to my Google Calendar. A process that would have taken me 15 minutes was completed in seconds. Out of curiosity, I tried the same with ChatGPT. While it correctly parsed all the session details, it understandably couldn't automate the calendar entries and instead provided instructions to add them manually, essentially no different from referencing the paper schedule myself. This experience underscored Google's current edge in certain areas with its combination of TPUs, optimized AI infrastructure in house models, consumer apps used by billions, and hardware like the Pixel 9 Pro, which Ben returned Google stands apart. He's such an Android guy. In contrast, Apple lacks comparable AI infrastructure and large scale models. OpenAI doesn't have consumer hardware or widely used apps. Microsoft, Amazon, Anthropic and others each face gaps in areas where Google Excelsior this raises an important question. Given Google's unparalleled position in AI right now, how Might they mismanage this opportunity and risk losing their technical leadership once again? Do you have any thoughts?

B (15:45)

Yeah, I mean the, you know, I mentioned one of my favorite use cases for my brief foray with the Pixel Pro, which everyone, I think this happened last time with Android. Everyone's like, just buy a Samsung phone. Like all that edge stuff, like the modem stuff and all the things. They're like, it works better. The Twitter app works fine. So I promise I'm gonna get another Android phone. Maybe not right away. I just bought another one.

A (16:08)

Don't make a difference.

B (16:08)

I'm getting a Samsung. No, I'm getting a Samsung next time. No, honestly, my iPhone experience this time has been terrible. Like, it's got that. So I'm open. But I also, I mentioned I translate stuff all the time and the Google sort of lens sort of like take a screenshot and then you can interact with stuff on the screen. Right there is great. So yes, they have all the pieces. I think the, the big issue for Google was actually sort of captured this week, which is they have all these incredible announcements. Their sort of new video generation model is amazing. It blows Sora out of the water. It's kind of embarrassing how much better it is than Sora and Sora. Meanwhile, it's this separate site. It's not even part of ChatGPT, which is probably sort of capacity issue. But like, then what are you actually bringing to the table? Like, you don't have the distribution advantage, you don't have the consumer sort of zeitgeist. Like, everyone knows ChatGPT. That's a thing. I've made the case that chatgpt.com or which I think they bought chat.com so it might change soon. But it is, it is like that is the value in ChatGPT. You can make the case like it's. Is it actually the models or is it actually the consumer sort of mind share again, which is sort of a counterintuitive take and a little controversial, but I think there's something there. They don't have that for Sora. And meanwhile Google just has this new product that's sort of blowing them out of the water and everyone's like, oh, this is so much more interesting than ChatGPT's boring announcement of like ChatGPT search is now available to everyone. But ChatGPT search being free to everyone is a big problem for Google. Like, like the, you know, I think there's that, you know, we're doing a good way of segueing to other question, but yet My default search is. Is in the vast majority of cases, maybe not default search. That's putting a bit strongly. I think I gave a challenge to you. Try to catalog every time you use Google Search in a day. It's such an astronomical number, you don't even appreciate it. But there's a lot of times that I go to ChatGPT. First I took the Google app off of my home screen so I could add another chat app. So I have Both Quad and ChatGPT are on the home screen. I figured I needed to give Quad sort of more of a shake since everyone keeps raving about it. But that's the problem for Google is that's where they make all their money. Right. They're like the.

A (18:30)

You get the VO cash from Google Search. That's how. And you get the Willow cash from Google Search and so.

B (18:37)

Exactly.

A (18:38)

That's a problem.

B (18:39)

That's right. And this isn't a management issue. This is more like a real disruption challenge which is a different technology comes along and, and it's not competitive with what you do on some aspects, but it is in others and it sort of screws with your business model. And that is, that's a problem. And you know, OpenAI is going to work to build some sort of advertising model to fund this sort of free level and increase the rate limits and things over time. And it's going to be pretty crappy at monetization relative to a Google search. But for a ChatGPT or for an OpenAI, it's fine because they don't make any money from advertising today. It's all sort of additive and they can figure it out. For Google, it's not fine if they're cannibalizing their core business. That is a problem. But that doesn't fit in like the engineering culture bits I think we were talking about before because that's just a, that's a business problem to be managed. It might be. It's still unclear if they're going to fully manage it over time. But you know, I'm pretty sympathetic to companies going through that challenge. I do have to give Google credit as someone who's been very critical of their management. They are actually shipping this stuff and the sort of morass and slowness with which they sort of responded and the sort of layers of bureaucracy that you sort of heard about that these things had to go through, that that didn't just make them slow, but actually made the products bad. Maybe there's been actual real changes there. I mean, we'll see how this gets when it rolls out. I mean hopefully we don't have to sort of rehash the problems with their image models and their text models that came up a year ago. But to the extent they fixed that, hopefully they did get sort of the kick in the rear end because we do need great engineering companies and Googles are great engineering companies and you gotta let the engineers ship. And to their credit they are shipping. That is good news. They still face this business challenge. But yeah, like if you went back to 2020 and said oh yeah, Google has all these advantages and they're crushing it, everyone would be like yep, that's exactly what I expected. And a lot of the angst and perhaps going too far in criticism but then again maybe not was because they weren't fulfilling their destiny. They weren't doing what you expected of them for several years. They're sort of in the middle and Android should be the best choice like in an AI world because of all these advantages. And what does that mean sort of in the long run, are they better placed than Apple? Yeah, they are. Does Apple still have a lot of the advantages in the lock in and the consumer mindshare and all the sorts of things? Yes they do. Can Apple recover? Can they have good enough models? They don't. Apple doesn't have to have the leading edge model necessarily. One of the big themes over the last year has been we haven't had the big GPT5 sort of breakthrough but what we have had is a massive array of models that are GPT4 level but cost a lot less to create and a lot less to run. And it's interesting because there's a bit of a issue I think in the model landscape, which is the big payoff from doing a leading edge model is that model helps you create more efficient models but that model never is actually economic to operate and run and you never make your money back. Like so where, where is the incentive? Like there's a, there's almost a long term concern about who's going to fund this leading edge model that everyone uses to make the efficient models that are actually make money in the marketplace. And you know, OpenAI gripes about this. Oh, everyone trained their models using GPT4 outputs which is as we noted is a great irony given their liberal interpretation of the use of other people's sort of textual outputs. But you know, it is an issue going forward. But that's an advantage to Apple. Like it's actually this might be a case where being a fairly fast follower, if you have the places to leverage it and distribute it is a good place to be. So we'll see to what extent they can sort of systemically take advantage of that.

A (22:57)

Yeah, I mean, strategically speaking on the Apple side, I think that they are well placed to sort of draft off of the leading edge market participants and just sort of capitalize on the progress without having to invest a ton up front alongside Google and OpenAI. I would also say, though, and it isn't about iOS 18, but just it seems like people hate the Apple intelligence product thus far, or at least are just wildly underwhelmed by the Apple intelligence product thus far. And I think the efforts there could be indicative of just a more general malaise and lack of dynamism.

B (23:36)

Well, I mean, I think, I think one example is like their whole image playground sort of thing, which, you know, Gruber made the case on dithering. Maybe this is just, it's almost more of a demo app. But it does speak to, oh, we'll do like stylized output of things you generate. And it like, it speaks to like a place of fear. Like we don't want to accidentally generate something. That's good.

A (23:55)

Let's not upset the Apple cart.

B (23:57)

That's right.

A (23:58)

Keep it simple.

B (23:58)

That's not a good, that's not a good sort of like sense you want from a company. Like, well, let's not give you the cool stuff because you might do something bad with it.

A (24:05)

It's like, well and especially everything is about to change. I mean, that much is clear. And so as far as skating where the puck's going to be, it just feels like a question as to how capable and willing Apple is going to be in the new era here. But they do have tons of advantages. They have me locked in. I will not be getting a Samsung phone to run Google products over the next couple of years because I'm an Apple guy.

B (24:29)

Yeah, I'm gonna, no, I am gonna do it again sort of at some point in part because, and honestly I think this is even relative to a couple months ago when again, and I think some of the issues I had, maybe they were pixel specific, maybe they were just user specific. Like the Twitter issue was a real one. Like it just wouldn't. Which lots of people have said, I don't have that issue. So I don't know what was going on there. But in the modem one was a real one. Again, might have been me specific. But you know, just the, the it was what it was that there was good reasons for me to bail on that phone specifically. But as My usage of these AI products increases and my desire to use them more increases and I have to deal with annoying workarounds to interact with stuff. It increases the salience of man. I would like this to be more deeply built in.

A (25:25)

It's seamless and easy.

B (25:26)

Yeah. And my assumption is I would like it to be ChatGPT or Quad. But I am just gonna assume that the Google alternatives are going to be better than the Apple ones. And that's starting to be a meaningful differentiator for me, something that I need to think about. Because if I'm depending on these and using them and want to use them, that's a reason to choose a different platform. Now, again, I'm hardly representative of the normal user, but this has always been the theory of Android, sort of. And a pixel in particular in the long run is whenever we shift away to apps being the key paradigm and Apple sort of advantages in that space to services being the more important paradigm. It wasn't enough to have like native Gmail incorporation. You needed something more than that. Oh yeah, the email situation was kind of a mess too. But whatever. The AI though is good enough and useful enough. That could be the thing. But this stuff, it does take time. It doesn't happen overnight, that's for sure.

A (26:35)

Yeah, you're not the average user in 2024, but your preferences might be a lot more common by the time we get to like 2028, 2029, and so we'll see how it all shakes out. Matt, while we're talking, what a challenge

B (26:49)

from an Apple incentive perspective is. Their goal just needs to be good enough that, oh yeah, I can do that too. And you don't realize that it's way better elsewhere. It just, it's good enough that you don't, you feel like you don't need to change. That's good from like a business perspective. They don't need to be great. It's kind of corrosive from a cultural perspective where our goal is not to be great. Our goal is to be sufficiently good that people can tell their friends, oh yeah, I have that too.

A (27:15)

Right. And they're not sufficiently good right now. They're mediocre on the Apple intelligence side. But it's still early. Matt says now that ChatGPT is becoming one of my go to apps, it's causing me to rethink how I use my phone. Phone, email, browser and messaging have always been in my iPhone. Doc. I want to swap in the ChatGPT app, but I can't decide which app to demote any suggestions From Ben? Has ChatGPT made it into his iPhone dock? Do you have any advice for Matt there?

B (27:47)

Well, so I have a very particular approach to my phone home.

A (27:55)

I never had any doubt.

B (27:57)

Yeah. So the idea of the dock is those are the apps that are all the time, no matter which screen you're on. I treat the whole home screen basically as a dock. I only have one screen of apps. So you're either on the home screen or you're in the app library, and I invoke you by searching for you. So there's no multiple screens of apps. So it's bas. So if you're at that level, you're at that level. So my doc is X, of course, because I'm an addict. A folder. People don't think about this. You put a folder folder in your doc. A folder of all my messaging apps. But then if you put all your messaging apps in the folder, then the folder has the, like, the unified notification count of. Of what's in there.

A (28:39)

Okay.

B (28:40)

The only apps that are in there are apps where I do check every message. So, like, my email is definitely not in there. I don't want notifications for, like, it's too much that. That somewhere else I have the camera, which is probably not necessary because there's all these dedicated buttons and stuff. Now maybe that's just sort of muscle memory.

A (28:56)

That's.

B (28:56)

I should probably demote that. And I have Google Maps. Because you actually use. I use it all the time and I want it right there, but on the home screen. Like I said, I did demote the Google search. The other thing is just. I don't know. Safari is there. I can search at Safari, but I do have Perplexity Quad and chatgpt all on there. I don't really use Perplexity, to be honest. When I use it, I tend to enjoy using it. It's never quite broken through for me. And I think, like, ChatGPT, adding search actually for me kind of killed any desire I had to use Perplexity. So I'm there as far as them being on the home screen. And the ChatGPT is in the spot right above the dock, on the right where the Google app was. That's sort of my default. And I like it. I use it all the time.

A (29:50)

Well, do you have any recommendations for people looking to incorporate AI capabilities into their daily life? Because, like, if you're not already using LLMs on an hourly basis, it seems like you have to be intentional about the way you incorporate them into your workflow. And I imagine that process takes some time. I mean, how did you approach it?

B (30:14)

Well, I mean for me, for me the, the anchor is I just always want to know everything. I hate it when I get a question or I think of something I don't know the answer. So I look it up like so I like I've been so have just

A (30:28)

wholly replaced search when you go to look up for a certain answer.

B (30:33)

Right. Yeah, I just want to like I. And I mean this is an incredible way to sort of accrue knowledge sort of over time in the long run. If every time you don't know something you look it up immediately,

A (30:46)

do you

B (30:46)

worry accumulate a lot of information?

A (30:47)

So do you worry about the accuracy though?

B (30:51)

Sure. I mean for sure. When I'm. I never go straight from an LLM to say certecary to sort of put something in and I mean, I don't know. I think like I've been doing this for so long I have generally a wide base of knowledge in whatever I'm looking up where I can usually if something doesn't seem quite right, I can catch it pretty quickly. Like there was something I was using it for the CISC risk discussion that I did in the intel article. It just like I was like checking a couple things and actually I think I got a pretty bad error about something about like preserving die space or something. I'm like that's not right. I think Cisk actually uses more die space and so like. But so I knew to look into it. Yeah, sure enough it actually got that. It actually got that wrong. Had I depended on it directly, I would have gotten in trouble. But I knew enough about how it worked that I immediately knew it was wrong. And so it's hard to give advice about that. Everything's a trade off, right? Like I recognize there's a risky way to do this, but my time savings are so astronomical, particularly as just the quality of search again which I think is mostly out of Google's control. Just, you know, that's the problem. You get an abundant amount of crap. Yeah, right. Well, also I hate how Google search is so biased to recency. It's really like you have to. I use the date function all the time. Even then I feel like that's even gotten worse because I'm always trying to find older stuff and you know, and this is actually one of the things I'm iffy about is I feel chatgpt invoke search too often. I actually kind of like that quad doesn't have search built in because it's like it just has to pull on its internal knowledge, which is risky sometimes, but sometimes that's what I want. I want, like, yes, you're only up to date as of, like, six months ago. That's great. I don't want anything last six months. I'm trying to give sort of historical context there. So for me, that's the end. But there's a good example. Like the email that you just read about, like, reading, reading, or invoking, putting all your stuff in your calendar. Like, that's like. So there's another example. Like, you know, I was talking to someone that had a medical sort of issue. They sent me a screenshot of their printout, and this is actually a friend here in Taiwan. So I asked it to, like, translate and explain what was going on. And it did it all in one step. And it was incredible. And the.

A (33:11)

That's awesome.

B (33:12)

Like, for me to go through and translate it and then look it all up would have taken me an hour to sort of, like, do, and it did it in 15 seconds. So it's one of those things that sort of builds on itself. You. You just have to start recognizing opportunities. Like, this could really help me do this thing really quickly. Which is. I mean, there's an analogy here which is like, automating, like, stuff you do on your desktop. Like, this is super nerdy, but if there's something on your computer, you sort of just keep doing the same thing all the time. You should automate that. Unfortunately, automation's generally a little out of reach for people. Actually, I think it was. Jon Gruber had a good point. I don't know. It was on Daring Fireball, if you said on Dithering or something, where this was the idea of AppleScript. Back in the day, Apple had this sort of scripting language for the computer that was like. It was meant to be normal language so you could, like, write little programs to do stuff for you. The problem was it was still deterministic. So it actually was horrible because it was pretending to be normal language, but you had to have the exact right normal language. And there's a bit where having, like, actual programming terms is better because you know exactly what they are and how they work. Right. Like, and. And so AppleScript super powerful. But it was like this art. It was really hard to write a good. There was no, like, real good debugging mode and all these sort. It was. It was a coding language that was trying to not be a coding language. That actually made it worse. Yeah, but there's a bit where the idea is now being realized with these LLMs where you can use natural language to get them to do something for you. And, and part of the. There's different aspects of this whole agent discussion, but one of the more exciting ones is can you trivially get sort of a process in place that just does stuff for you automatically that you did all the time? And part of this is, and this is almost a marketing challenge is how do you make people aware that this is possible? Right. Like our emailer is going to use the calendar ad thing all the time. Then also, what happens if it screws up a calendar ad?

A (35:23)

Yeah, that's why I asked the question. So Rob's email inspired the question as we talk here. And I, in my own life, this is embarrassing to admit, but like the mistake that I made with ChatGPT over the summer with the Demar DeRozan sign and trade. And then again, I checked some salary cap information that I think it was chatgpt. It may have been Gemini, but one of them got it wrong and led to me making a bad argument in a group chat. Those two experiences have me reluctant to trust the ChatGPT or Gemini information on anything. And that's stupid because obviously these are really powerful tools. But as I project forward to 2025, that's one of my goals, is to be more intentional and more proactive about finding ways to incorporate these into my daily workflow and just daily life. Like you said, they're great information resources, but I do still have that sort of trust barrier on some of the searches.

B (36:30)

No, I think that that's a really good point. And like, my core use case of like, looking stuff up all the time is arguably one of the most dangerous areas because it's right so often and then it's wrong and you trust it and then you're screwed. Right. And so there's a bit where I almost don't recommend what I do because I am sort of, again, I'm just. This is just the way I operate. I've accumulated so much knowledge just because that's the way, like, since I'm a little kid, I'm like, I was the kid who would get a dictionary or go to encyclopedia when I would didn't know something because it bugged me to not know something. Like, so, like this is before we even had computers.

A (37:08)

I was also that kid. Just for the record, I remember the Bill Clinton, Monica Lewinsky sex scandal. I sat down, you learned about the world, a physical dictionary, like a giant dictionary. And I looked up every word that I didn't understand on the drug. It was a very educational experience for me and the dictionary had it all. Everything I needed to know is in the dictionary. So thank you.

B (37:35)

That is amazing. That is incredible.

A (37:38)

So I was. We were. We were like minded elementary schoolers I would say.

B (37:43)

Yeah, there's a. So there is an aspect of this stuff not being ready for prime time. On the other hand, it does. It reminds me of the self driving car issue which is I think there's a good chance that self driving technology is good enough that actually it should be deployed and be allowed to be used everywhere. And even though there will be bad accidents, but because the number of and intensity of the bad accidents is still going to be less than distracted humans or whatever sort of operating that it would be worth it. But the reality is that's not going to happen because the status quo is so powerful, is such a powerful sort of bias and often for sort of very good reasons that self driving cars need to get to a point where they're exceptionally massively. The delta is. Is. Is so huge that they're inexor the drive for them is sort of like the necessity is. Is. Is so massive that then they can sort of become a dominant sort of thing and it's this huge sort of leap forward. And that's might be the case for a lot of folks. Like on balance the amount of time that I think you can save with these tools is probably worth the mistakes you will make along the way. But the mistakes feel so catastrophic and they hurt.

A (39:08)

They feel like betrayals. Yes, right.

B (39:10)

And. And you feel like such a moron when sort of it happens forever trusting, you know.

A (39:15)

Yeah, exactly.

B (39:16)

So I don't know. Proceed at your own risk to a certain extent. But this is, this is the like. Here's another example. It's well known in sort of digital advertising that the best bang for your buck ever is whenever Facebook introduces a new ad unit. Because what happens is they can't be measured very well. But it's Facebook, it works. And so like when Facebook added story advertising, like companies, there were individuals and companies that made astronomical amounts of money by just pouring all their ads into stories, even though they had no idea or no real good measurements whether it was working. But because of that there was much less competition for those ads. So they were super cheap and they were definitely working. Facebook just couldn't measure it very well yet. And so it's like you sort of took the risk. But the problem is if it didn't work and sometimes it didn't work, you weren't getting that super granular feedback on every ad, right. So you were running the risk of putting a bunch of money in and not working and oops, we just lost a whole bunch of money. Unlike most digital advertising where you know exactly your return. So what do you want? Do you want a guarantee? You put this money in and you're going to get this return and this lifetime value of a customer, but it's a relatively small margin or you put a bunch of money in and by the way, if you screwed up because you're having to do it yourself, you just lost all the money because it just went down the black hole and it's gone forever. But if it works, you get this big return. Same thing happened with sort of reals and you have like this one to two year window where you have this huge opportunity but if it goes wrong it's pretty catastrophic because the money's down the black hole. You got sort of no return on it. And like, like that's the state of like. I think using a lot of these alarms is, is there's high risk but there's real high return. And you know, it's going to be interesting to see how long this persists. This gets into my why I think a big bit of these rollouts is going to be top down, big job, wipeout sort of domains. If you're an employee in a company, the risk of using an alarm and screwing up, yes, you'll do it. You want to tell anyone but like if you, if you relied on an LLM to make a business decision and it was wrong, it's on you, right? And so you're going to be risk averse to do it. If you're a CFO and a CIO thinking about this, you're like okay, it's going to be wrong X percent of the time but the cost of them are so much lower. And by the way, we already know humans are wrong sort of Y percent of the time and for now Y is lower than X but the cost of the humans is so much greater. We can just sort of wipe out this whole department use AI. We can run the cost benefit analysis assuming a certain level of screw ups and while the overall expected value is

A (42:00)

still lower, still net positive.

B (42:02)

Yeah, that's right. It's like if you were the czar of the roads in America. You can come in from the top and say we're putting in like we're allowing self driving, we're banning human drivers, we're going all the way Right. It's just going to be a better situation. Are there going to be accidents? Yes, but I'm king of the world. No one's going to like. It doesn't matter.

A (42:20)

Are there going to be accidents? Yes. But will there be fewer accidents overall in that world?

B (42:26)

Right, but in a decentralized sort of like, like way, one self driving car accident is front page news. Right. And that is such a retardant on adoption of these new technologies, even though on an objective basis it might. Now again, I'm not saying that categorically about self driving cars. I suspect that's the case. But let's just assume, assume that that's sort of the case. And I think this is going to apply to this generally, which is, I think is good news if you're sort of an individual that can figure out how to leverage these because you have this meaningful arbitrage opportunity for a while if you can accept the risk.

A (43:02)

There you go. Let's all seize the day over the next couple years until the rest of the world catches up.

B (43:08)

The real shift though, and this is where people really get left behind, is when it becomes table stakes. Right. And this will probably happen in sort of, sort of coding environments or development generally. At what point does using alarms make you more productive? Shift to if you don't use an album, you're just not nearly productive enough. Like, and that is, that's the difficult shift. And it's also the really profitable shift if you're sort of the toolmaker.

A (43:35)

All right, well, one pronunciation watch a couple times on this show you've said inexorably and I don't know whether that's the correct way to pronounce it or not. I've always pronounced that word inexorably and so I guess we'll have to rely on the emailers to let us know which one of us is right. But I'm glad you're right. One of us is wrong. Okay. All right. So inexorably. But it's appropriate way for us. I pronounce the way it should be pronounced. Your pronunciation is fancier. I like inexorably more than inexorably. But in any event, Jonathan says if Ben could teach any business school class, what would he choose? The obvious answer, Some macro class about Internet economics. Since most business classes pretty. Since most business school classes pretty much ignore zero marginal cost goods. However, I personally would love a class where Ben examines different business models using real world examples. Meanwhile, if Andrew could teach any class in sports, media or law school, what would he choose Ben. Do you have an answer for Jonathan?

B (44:46)

Yeah, I mean I'll probably do this at some point. I enjoy, I enjoy teaching, giving back. Yeah, well, and yeah, and it just, you know, certainly I would like to think and hope that business schools are talking more about zero marginal cost markets than they were when I was there. That was, you know, a long time ago at this point and a lot has happened in the space since then. So. And I do get lots of requests to use my articles from professors, you know, like get clearance or whatever. So at least some people are out there. I have some sort of proof in that regard. What I think is more interesting generally though beyond the sort of like technical, like looking at the economics of how things work and Trategory has certainly been about that. But there is a broader point about general alignment between culture and business model and sort of outcomes and the degree of success that comes from that. And I don't know, like I'm sort of just thinking about this off the top of my head but putting together a macro understanding of companies and what impact does that have on their performance. Right. And this is almost a fuzzier area. There's lots of famous business books that like lay out, oh, this is how great companies operate. And you go back like 10 years later, like half the companies are out of business or bankrupt or got acquired or something along those lines. But I think that's also interesting in its own right. We kind of got to it a little bit earlier with discussions about Google or Apple or whatever it might be. Companies form around sort of a moment in time and. Or like an Intel. Right. You have this sort of culture that we've talked about is arrogant and you know, our way or the highway. That was a culture that was written about glowingly when it meant pushing Moore's Law forward and taking big risks and sort of doing new things. And that was aligned with their overall strategy of CISC versus risk. For example, we're going to stick with CISC even though it is less efficient because we already have a software ecosystem advantage and we'll make up for it in manufacturing. You go back to those decisions and discussions in the late 80s and early 90s. Those are like Pat Gelsinger was a driver of that. He was the cis guy. And Pat Gelsinger comes back in 2021 and realizing from a macro perspective intel needs to become a foundry and he is fighting with the ghost of pat Gelsinger in 1990 to do that. That's pretty interesting, right? And that is really integral in Understanding these long run trends of companies. Every bad culture you hear about is downstream from a good culture that made it happen. And what changed? Did the culture suddenly change? Actually, no. The culture is usually the same. The context changed. And I think that's pretty interesting. And that's what you see happen to a lot of these companies again and again.

A (47:50)

I would say I would take that course, but I'm sort of already signed up for that course hosting this podcast with you.

B (47:56)

Yeah, you're a good point. I just want to keep. I just want to keep having the same.

A (47:59)

Let's keep podcasting for the next 25 years. Yeah. Yeah. Well, in law school, the two classes that I would like to teach, and to be clear, I would need to do a lot more studying to feel comfortable teaching either of these courses. But criminal procedure and constitutional law are just very interesting studies of American history generally and the way courts have adapted to solve problems over the years. And you end up with a bunch of good philosophical questions that don't have clear cut answers. So I think for me, selfishly, it would be intellectually stimulating to teach in those two areas. If I ever went to teach at law school in sports media, I feel like I would be very good as a professor of podcasting. My true passion project though would be to work with the next generation to restore the lost art of good sports column writing. Because I feel like people just don't really do that anymore. And there's a lot of analytics writing out there that is frankly pretty boring. And the youth of America has forgotten how to just have a take and write it in 800 words, say something interesting, say something that challenges the audience and make it fun. I would enjoy trying to implore a bunch of 22 year old grad students to do that with their writing. I unfortunately don't think there's much of a market for that book anymore.

B (49:34)

People should probably not be going to grad school for column writing given the current environment. I do think it's a huge problem. I feel like there's an entire generation of sports writers that I just can't stand that that is basically single handedly driven me to make clear to everyone, no, I am still Gen X. It might be the last year of Gen X, but I'm not one of these like, and it's this sort of sense that you have to sort of ironically always be aligned with Zeitgeist.

A (50:05)

Yeah.

B (50:05)

And like you, you, there's a real fear to your point of takes. And I think the, the analytics movement is part of this where People just like throw a number at you. And it's funny because like, this is an issue in business, right? Like there, there is a, there is a whole wave of, you know, what's the phrase? You know, God gets opinions, everyone else brings data or something along those lines. Like, I'm butchering, I'm butchering the concept.

A (50:28)

They throw numbers at you because they're afraid to be wrong. And if they have numbers to support a given opinion, it feels more solidified, right? You get to play the number for retroactive criticism. And it's like, all right, well no, but this is the same thing.

B (50:44)

This is the Intel Boeing thing. It's this sort of retreat to numbers. And what you see again and again, you see this in sports, in sports analysis, you see this in companies is everyone gives. Not everything can be measured. You say that and everyone's like, that's right. Not everything can be measured. We agree. Then you say like, well, so we need to make sure to not depend on just the numbers and making decisions. There's like, that's right. We need to make sure to not depend on the numbers when making decisions. But when it actually, the rubber hits the road, if numbers are available, people retreat to them again and again and they just win arguments. And this leads to, this is the recipe for short term optimization and long term screw ups. Because the long term is the bit that's uncertain. The long term is about uncertainty. It's the unknown. And you're sort of hedging against that or investing in something that is not sure. And if you're going to put numbers against it, the numbers are made up. And this is just a real danger I think you see all over the place. And it's one of the things that's, that you could apply this to the whole globalization discussion, right? Like if you pencil it out, if you, if you do everything based on numbers, the choice is always obvious. What you can't pencil out is what does it mean to, to hollow out your sort of industrial base over the next 50 to 100 years, right? Like, like, sure, there's no wars on the horizon, but if there ever were to be in the grand scheme of history, which by the way, if you look at the grand scheme of history, wars are a regular occurrence. It might be something to sort of like just assume it might happen. Is there anything you would do different? Like, well, the numbers say xyz. It's like, then numbers go out the

A (52:26)

window in certain scenarios. Yeah.

B (52:29)

And this is arguably a tech derived issue. Like what was the big killer app for the PC, the spreadsheet. Like, like maybe we could blame tech. If you want to blame tech for everything, start there. Like once you had the means for everyone to apply numbers to a problem, you got massive amounts of optimizations in all sorts of areas. There was all this low hanging fruit that was addressed. The problem is the lure of having a number is irresistible for anyone but the strongest leaders. This is part of the reason why founders are feeded in tech and given so much leeway. Because they have the cultural authority to do stuff that's not in the numbers. Everyone else has to bring data like that, that, like that's maybe the, the how the statement should be changed. It's not God and data, it's founders and data. And the reason why founders are treasured is there's just this sense they're the ones that don't have to have the numbers and people will still listen to them and do it. This is why Mark Zuckerberg is like king of the hill. He's still there, he's still engaged. Of course Facebook is one of the most data driven companies on earth, but there's still someone who can say, yeah, yeah, there's feel. Exactly. And once you see this in one area, you realize how it's pervasive everywhere. And you like one of the reasons NBA Twitter is so irritating. Like there was like people just, they'll literally make up numbers and then it's like, am I going to spend the time to go in and investigate this number and see what's going into it? Right. Like, you know that defensive plus minus incorporates assists on the offensive end. Who knows?

A (54:10)

Really?

B (54:11)

Yes. That's why Joker's defense DPM numbers are so high and it's used as evidence. Oh, like his defense is really good. It's not that like Giannis is like, it's pretty comparable on that end. It's like if you actually dig in and understand how these numbers are calculated and these weird assumptions that were made because you're sort of fitting the data to a curve. So you get someone, a center that passes, has a ton of assists, who breaks the model and then you use that to say, okay, we're going to decide who's the best player in the world, XYZ and who's going to win Gazillion MVPs. I love Joker. This isn't an anti Joker thing. It's a.

A (54:45)

The discourse is the eye test matters.

B (54:48)

The eye test matters. It matters in sports, it matters in business, it matters. Like one way to think about intel and Boeing is, they stopped relying on the eye test, they started relying on spreadsheets. And again, this isn't an anti spreadsheet bit. It's not even an anti analytics bit. It's an observation about human nature that people will give lip service to. Yes, there's things that can't be measured when you think about more than numbers, but when you're in the boardroom, when you're in the meeting room, the numbers tend to win again and again.

A (55:23)

Yes, it's absolutely human nature to seek the comfort of numbers that could support whatever argument you might be making.

B (55:31)

That's right. It's an abdication of accountability because you can just point to the, point to the spreadsheet.

A (55:36)

Well, and as far as human nature, and you mentioned this is all sort of a tech story. The other aspect of the story is that, that in the past 30 years ago, when every newspaper paid columnists to write interesting things that challenged consensus and made readers think, you didn't instantly get feedback from anybody who disagreed with you. And I think particularly for young people, that feedback, the negative feedback that they encounter also incentivizes just sort of going with the crowd on any given issue. And that's unfortunate, but I think it's also sort of a byproduct of human nature where we internalize negative criticism far more than we internalize any sort of positive feedback.

B (56:22)

Yeah, this whole generation likes to think about, oh, we're the first ones that are native to the Internet, we're the ones that know how computers work, blah, blah, blah, blah. Actually, no, the problem is you're native to the, you're the first ones on the Internet. So you had zero antibodies against the massive force for consensus and sort of like, and so you, you have the, like, you just see people who, their opinion, you zoom out and their opinion is so clearly motivated by trying to fit in with the zeitgeist. Right. And so you, you have, you know,

A (56:51)

I was, it's just self preservation and psychological preservation and it's.

B (56:55)

So it's entrenched at this point that, that people don't even realize that they're doing it. Right. They're, they're. And so you end up with this absence of independent thought. And yeah, this was a real value to having the random newspaper columnists in a geographic area who. Yeah, maybe get a letter to the editor.

A (57:15)

Exactly. You get a letter or so. There's a lot more friction in that process than there is today.

B (57:20)

I, yeah, the, the, the, I insulate myself. This is one of the reasons I got super active on Twitter as sort of Ben Thompson. Because it's like I'm one of the most disagreeable people in sort of like the personality sense ever. Like, all those tests I get like 0% or whatever. Like, I just don't really like, like, you're going to have to convince me I'm not going to be cowed. But even with that being said, it's hard, like, and it does have an effect on your mood and sort of the way that you feel about things. And you can understand why. There's a, there's a, there is a self preservation instinct that kicks in and you get this irritating, ironic, detached tone that is perfectly aligned with everyone else thinks. It's like, what is the, what's the value here, right?

A (58:09)

Like the, this is, this is what I'm going to fix when I go to journalism school. I'll go teach at Northwestern. I can, you know, commute to.

B (58:16)

Well, maybe, maybe we've reached the point where going to journalism school is going to be sort of a big enough zag. It was like, don't do that. What a waste of money that you'll actually get the disagreeable source.

A (58:28)

You know, that would be my number one bit of advice to anybody out there. Don't go to journalism school. I don't know what you're paying for at this point.

B (58:34)

I mean, you see this in the independent publishing sort of area, right? It's the folks that succeed are ones that have pretty distinct views and opinions. And you know, like there's a bit where substack sort of bore the brunt of that in a certain extent, where it's like the mainstream critiques them because look who's succeeding on your site. And is that a function of substack, like promoting a certain point of view, or is that a function that the only way to succeed with sort of takes on the Internet is to be different than the consensus. Like, what value is there? You're not going to succeed because you're style of prose is slightly more attractive, but your opinions the exact same as everybody else. And, and so yeah, the, again, it sucks. Like, this is one of the, I think bad things about the Internet, like, and social media in particular, it's this massive consensus driver that pushes people to sort of extremes because people just don't want to be criticized. But hopefully the next generation, there are more sort of antibodies against this. And you know, but the problem is

A (59:47)

Millennials though, we were the first off the boat and it didn't go well. But such is life. I'm Doing fine. I remain disagreeable, although agreeable. On podcasts, I'm a weird mix between agreeable and disagreeable. But I like to think I can maintain original thoughts and am operating independent of the group think machine.

B (60:07)

No, you're, you're, you're. By feedback to you is you're an inherent. I mean we. The tech word is contrarian. Your sort of podcast word is a zagger. You ought to zag what everyone is zigging. The problem for me in terms of basketball coverage is I feel like, you know, you talk about being a member of Giannis Inc. I just feel like after you became friends with me, you became much more critical and anti buck.

A (60:36)

It's true. It's true.

B (60:37)

Because you weren't, you were just constantly responding, being contrarian to me. And so it actually ruined my listening experience. It was.

A (60:46)

So that's also a reflection of the current state of basketball media because I don't consume very much other basketball media. I watch the games, I form my own opinions and then I'm not exposed to a lot of the national conversation. And so a lot of the conversation that I'm getting is text messages from you. The most relentless propaganda campaign. I can't help but resist it.

B (61:10)

No, I take responsibility. I'm not blaming you. This is, this is, this is a me problem, not a you problem. That is an interesting point though I don't read much tech analysis, I would say by and large and like Twitter generally, it's not super tech heavy and maybe, maybe that is sort of a surprisingly consistent angle. I wonder to what extent that applies to some of the more interesting folks in any arena. Like, like is there. I read a lot. Like I said, I, I dive very deeply into information about things, but I usually am less interested or I don't go out of my way to read other analysts and I don't know, maybe that, that, that, that helps.

A (61:51)

Yeah, well, speaking of other analysts, Mark says in a recent Dwarkesh episode with Dylan Patel, Patel described TSMC chip production as incredibly complicated, where a bunch of hyper specialized individuals on an assembly line know their narrow specialties, but very few people, if anyone, understand the big picture. Hence the real value of TSMC is not in the physical fabs, but in the human capital or tacit knowledge that this network of highly specialized workers possess. Patel then suggested that AI can actually help the chip design process by being a super intelligent systems designer sitting on top of the process that understands each specialty quite well and can thus optimize chip design with overlapping knowledge. Is this an accurate Description of chip production. Do I understand it right? Do you have takes on what Patel said? Ben, do you have takes?

B (62:49)

I do have this episode queued up. I haven't, I haven't listened to it. But those are, you know, it also, also Asianometry. Those are like three of my favorite people that were on that. So I will, I will listen to it just based on this summary. Yes, this is, and this is a point that comes up a lot where people are like, oh, you think TSMC is so great, why don't you just buy asml or whatever? And it's like, no, actually there, there is. This is the tacit knowledge point is you have. There's hundreds of steps or maybe thousand or whatever in making a chip. And all those have to work well. And the way to get them to work well is there's just a massive amount of understanding that comes from actually doing it. And it's why it's so hard to sort of move down the chain. Like what, Intel? Yes. You can do five nodes in four years in the lab. What do you actually learn when you're doing it at high scale? And can you actually just jump down when there's all these lessons you needed to learn at each step along the way that you're trying to do all at one time and they all interact with each other. So one variable changing here changes other ones elsewhere. Right. And if you're trying the more variables you're introducing into equation, the massively, like it's an exponential increase in complexity or quadratic increase or whatever it might be, and the overwhelming possibilities just overwhelm you and you fail. That's why you kind of have to just move down the line. And why TSMC is in such a powerful position, because they're the ones that have stepped down this line to where we are today. Everyone else is trying to skip steps. That's really hard. You don't just go buy some ASL machines and make a chip. It's not how it works. Now, this bit, could AI actually have the capacity and understanding to internalize not just broadly speaking, how chips work, chip manufacturing works, but actually all the individual details along the way and how those might interact? Yeah, I think that's possible. It's actually a pretty good articulation of. Of what? Like how do you actually solve a sort of multivariate sort of problem. Right. Where you change one thing and you do the other thing. It kind of goes almost back to our quantum mechanics. Like there's a bit where AI is like the quantum mechanics sort of here, it's not at all, to be clear, I'm just making an analogy. But deep learning going through and basically just inhaling all the data and discovering and picking out the patterns that actually matter and make sense. It's this capability that humans don't have because of the vastness of the data. And it's sort of like it's exploring all the possibilities all at the same time and picking out like understanding the right ones. And it's this degree of parallelism and exploration. There's a bit where the mistake with AI is trying to fit it into human patterns. And what you need to understand is patterns that no human would do because it's not possible, but AI actually can because it approaches things differently. And actually there's some human things it doesn't do well because it, like humans have this learning functionality. Like humans are very slow to an extent, but every time we're doing something, we're learning. So the second time we do it, we literally just incorporated what just happened. A big problem with LM today is they don't do that. They had this massive learning stage and then they're just kind of dumb going forward. And so you think about this tacit knowledge piece. It's a bit where every single iteration and every single run, you're adding to your knowledge base. And it's that slow accumulation, but the fact that it accumulates immediately that makes the human aspect of TSMC very hard to replace. But at the same time, can you,

A (66:46)

the top, catch up?

B (66:48)

Right, right by. By somehow getting all that sort of in place. So I don't know, I guess I'm almost talking myself now out of the analogy because of this, this incremental learning sort of bit. There is a bit where I think for alums generally, iterating on them quickly is important. Like iteration speed is critical. And I do think iteration speed, though, iteration learning is the area where humans, like we do something that we haven't figured out how to do on computers yet.

A (67:20)

Well, we'll stay tuned on that front. But while we're talking tsmc. EW says. I saw Ben's tweet about his latest mug acquisition from the TSMC Starbucks. So this is a lower stakes TSMC question.

B (67:34)

Now that you've seen this mug, it's incredible.

A (67:37)

It's a really cool mug. So EW says, as an avid mug collector myself, I'd love to hear about some of Ben's favorites from his tech company Coffee Cup Collection. So for anybody who missed it, on December 1st, Ben tweeted, acquired the best Starbucks mug of all time and the pinnacle of my tech company coffee cup collection. This was only available at the TSMC Starbucks and features the Starbucks logo as a chip on a silicon wafer. So I wanted to ask you about this like 25 days ago and I'm grateful for EW emailing in with this question. Can you explain the tech company coffee cup collection and this particular TSMC Starbucks mug?

B (68:22)

Well, just back in the day, every time I'd visit a tech company, I would, I would just try to get one of their coffee cups. So shop. Sometimes, you know, someone would let me lift one from the cafeteria, whatever it might be. So by and large they mostly stink. I mean, the TSMC one looks awesome. It is however, from my estimation, low functionality because so the perfect coffee cup that has never been beaten from a functionality perspective, which also looks bad and this really just fits perfectly is the Microsoft coffee mug that I was given on my first day of work, the clunker 2011. I still have it. It has a little chip in it. I will be so sad when it breaks. What it is is first out as the old Microsoft logo, like the original one, like the little thing in it, it has some sort of phrase on the side. I can't remember what it is, but it's like it's all faded out at this point. But what it is, it has a very fat base, super wide, but then it curves up into a very narrow top that sort of comes out a little bit. So it, it's super stable, it doesn't spill. But the narrow top makes it really easy to drink from. You don't get like the splash of like coffee over you. I've actually bought, I actually have one here. This is just a one I got off Amazon where I bought. I know once this coffee cup breaks, the functionality is so high I need to have replacements in place. That's right. So that is the best coffee mug ever. Again, we're coming up on 15 years. I still use it. It's still my most used coffee cup. Kudos to Microsoft. Looks terrible, but whatever. The second best tech coffee mug I had from a functionality perspective was actually an Apple one I got from the gift shop. It was just, it was a normal shaped one, but the handle you could get all your, you all your fingers in. It was big, but it wasn't too wide. It was just like. So it was a straight up and down one. It looked good. It was black with like a small white apple logo on it. Again, Apple, like a very good marriage of form and Function couldn't give up the form enough to make it perfect functionality. Like Microsoft. But it didn't sort of like lean all the. You know, this was pre Johnny, I've Apple. It didn't lean all the way into the sort of form. Warm too. Yeah. So the TSMC mug is too wide. It's a big coffee I hate.

A (70:44)

So is it difficult to drink out of? Is that the issue?

B (70:48)

Yeah, your coffee gets colder faster and it's just like in the. The handle. You can't like fit your whole hand into it. It's. It looks amazing, but it's not. Like, honestly, I did checkering mugs a few years ago. I. I didn't love them because they're too big. All way too big also. And the handles too small. So the, the marriage. The marriage is. Is tough to get right. Nothing will beat my Microsoft cup. Sort of from back in the day.

A (71:09)

But does anything look cooler than the TSMC mug?

B (71:12)

No.

A (71:13)

And we'll link it. We'll link Ben's tweet and photo in the show notes for people who want to check it out, because it does. It looks like a microchip.

B (71:19)

Yeah. You could buy them on ebay, I think. So. Like, I. I had a friend who's a. Works for a TSMC supplier was from tsmc. I saw like. Actually, I think I. I saw about it. Yeah. And I'm like. I'm like, oh, like, this is amazing. He's like, oh, yeah, I have one. He's like. He's like, like, do you.

A (71:34)

So you can I pay $500 for it?

B (71:36)

No, I'm like, no, I'll go buy it. He's like, no, no, I actually have two of them, so I, I use one myself. So I'll give you the other one. So thank you. Thank you too. Yeah, but it is, it looks very cool, but it is not as functional.

A (71:49)

Well, while we're talking unbelievably nerdy collector's items, I do feel obligated to make our audience aware that there's a LEGO version of the Twinscan EX C5000, which is ASML's top of the line EUV lithography machine. We'll link that in the show notes as well. I was reading about that EUV machine. The real machine weighs 165 tons and costs around $380 million, which is just staggering. The Lego version is only $225, but tragically, they only sell it to ASML employees. I tried to buy one and I thought I had beaten the system because My purchase initially went through, but then 10 minutes later it was canceled by the system. So I won't have a lithography machine Lego, but I'm glad that it exists. For the nerds out there at asml, I'm glad they have a LEGO version of their lithography machine.

B (72:51)

I'm trying to make some sort of joke about coal in your stocking. Given that we're dealing with sand or whatever. I'm completely failing. So let's just.

A (72:57)

No, but a good effort, a good thought. So I hope the audience gives you credit for that. Dan says, I'm sick of the neutral TV announcers. RSNs are the only outlets with Homer announcers and they are dying. So what are your thoughts on national telecasts offering a choice of audio streams? I'm sure plenty of Twitch streamers would call games for free if given the platform. Ben, do you have thoughts?

B (73:26)

Well, I. We know that Dan isn't a Celtics fan because you, if you were, he'd be happy to know that the national

A (73:31)

media takes care of you. Don't worry, Dan.

B (73:33)

That's right. So. No. Well, it is interesting. There is a. Actually a really interesting startup called Playback where you can sign now because all the rights issues you have to like sign in with your subscription or with your cable subscription or whatever it might be. But you can watch games with anyone can be an announcer. So you and I could sort of announce a game. Maybe we'll do this at some point and I would love that.

A (73:59)

That'd be great.

B (74:00)

And you got your.

A (74:01)

Watch basketball.

B (74:02)

Yeah. Or you could just be on with your buddies and you could just have an ongoing conversation and the streams, all the streams all synced up because actually like we'll go through Playback servers so you know, you're all watching the stream at the exact same time. That's probably realistically in the wrong. Well, in the future, I mean Playback, I think, I don't know if they'll as a standalone entity or if they should be acquired by one of the leagues or whatever it might be. But the concept of you're going to have the barbell effect, you're going to have sort of the national big platform ones and you're going to have user generated content on the other. Like that kind of comes for everything and it's already probably the one going to come for this. Like, like the RSN model is just fundamentally economically broken. It still works in some markets, which is why it's going to be hard to sort of move on. But that fits very much in the middle with your newspaper columnist. That used to have good takes.

A (74:53)

Yeah. Well. And I mean my thoughts having been a member of sports media, I guess I'm still technically a member of sports media with the goat, but announcing a game is harder than it looks. And so I don't know that most

B (75:06)

of these announcers are terrible.

A (75:07)

Yeah. You realize not that appealing to me in practice, but the theory is appealing and maybe that is the direction that things will go. I've never been that invested in announcers. One way or another, an announcer has aggressively bad for me to care.

B (75:23)

Yeah. The thing with user generated content is the. The vast majority is horrible and terrible.

A (75:29)

Yeah.

B (75:30)

And so that's the challenge of building a meaningful user content platform is you need so much volume and then you need sort of the algorithms to pick out what's good such that the absolute amount of your good content is very high. But the absolute amount being high is a function of the total content being astronomical because your hit rate is super low. So this is only something that. And this would be a challenge like a playback, for example. Right. Is how do you get. Particularly if you have this constraint of signing up for other. You have to sign in with your other actual service and tie it in like the. How do you get the volume high enough such that your absolute amount is good? Because the relative quality is always going to be terrible.

A (76:18)

Right. Yeah. You need a bunch of people to buy in and try it and then hope that a fraction of them are actually decent and appealing to the 90% of people who don't post well to keep it moving. A couple more questions toward the end here. Two quick notes on Twitter. Karthik says I moved to using Twitter exclusively on my Chromebook a couple months ago and have been very happy with the change. There are differences in my own behavior. For example, I'm more inclined to open a new tab and follow up on a lead myself. And thus I find myself reading more long form articles. But also I wonder if the algorithm presents laptop or web users with less slope than app users. I wonder that too. I don't imagine they can. Well, I don't know. What do you think on that second question?

B (77:06)

I don't know. I mean I. I think it's hard to divorce that from just. It is striking how different your own behavior will be with like it's amazing. A different input method. Right. Like the doom scroll is a real thing. It's just so easy to swipe and move back and forth, whereas moving a mouse and clicking is a very sort of deliberate action. And it's a really powerful example of what a difference friction makes. Right? Like this is why you talk about all the time. This seems so innocuous, the sort of difference. But yeah, I think about my behavior when I'm looking at Twitter on the web versus on the phone and it's totally different. It's much more delivered on the web. I'm more likely to stop and read something or go off another tangent or flip over to ChatGPT to look something up and on the phone it's scroll, scroll, scroll, scroll, scroll.

A (77:56)

Yeah, and it's just a lot more compulsive. So thank you Karthik, because he sent this about a week ago, maybe 10 days ago, and I actually made this change in my own behavior and it's just much better. So great tip from Karthik. And another note from Kendra. She says, andrew, here is a trick to make Twitter great again. Just mute the thread emoji. It filters out the most annoying of all threads. I've also found the pointing down emoji is worth muting for the same reason. I haven't tried this yet, but I will try that. And Kendra says P.S. please bring back the TikTok segment. I don't have kids yet, but I love hearing you and Ben give your tips. So on that note, we have Haken here who says, I have a question for your next TikTok segment. Is tutoring a good idea? We have three kids and our oldest is heading for elementary school next autumn. I'm not worried about grades or getting into the right university, but my wish for them is to gain confidence from learning and being allowed to deep dive into what interests them. Parentheses play to their strengths as per Ben's life advice. Short of opening my own school like Elon Musk has Elon opened his own school? I was not familiar with that in my Elon lore. How can I give my kids the best chance at being pilots in the future? Which is of course a reference to Ben's Wally metaphor in the MKH MK hbds for Everything article mkbhd.

B (79:32)

But yeah, okay, look, this is. I'm hesitant. I don't know why. Well, maybe I do have reasons why I don't want to give. Here's the way I think about it. So I'm presenting what I think.

A (79:46)

This is not prescriptive. This is just your own experience.

B (79:49)

Okay, well, it is interesting because it's a very intense sort of topic and subject. Being here in Taiwan, where tutoring or Bushiban, which is like cram school, like, like after school School that you go to and there's extra homework and then you stay up until 2am Doing it is the expectation. It's the norm.

A (80:07)

It's like voluntary workouts in the NFL.

B (80:10)

Yeah, exactly. And so it's, it's a tough. So that's the context that I am thinking about this question. And so I'll just say what worked for me and what may or may not sort of work for you. So I'm generally opposed to this viewpoint, but again, that's the context. Like I do think by and large one to one instruction and meeting a student where they are is a good thing. However, in the context where I'm at here, that's not the way it manifests. It manifests as a sort of a baseline expectation. And yes, of course to get good grades is a good thing, but what actually matters in life? Grades don't matter, right? They're a proxy at some point to sort of help you achieve something. The question is, did you actually learn? And the problem with a, a lot of sort of intense academic environments is actually what people want is good grades. They don't want to learn. It's like the measurement bit. Like people optimize for what is measured and you end up sort of focusing on grades and no one cares if you sort of learned anything. And there's other things to learn.

A (81:19)

Professors these days, apparently people are very, very annoying about grades and that's all any student cares about. And parents. And it's just like that's a lot of professors spend their entire time responding to angry students and angry parents who are hectoring them about, you know, an A minus instead of an A.

B (81:39)

Well, there's other things to learn in life beyond subjects, right? And I think like sort of the whole AP approach or whatever, which is more and more I think out of favor in like the most, most sort of cutting edge schools for good reason. I have chatbots right here. Why do I need to look up the, like the king of XYZ country and whatever, right? Like, like the, the. So there's a bit about sort of how like curiosity and desire to learn is actually increasingly rewarded given the capabilities we have to learn on the spot. If I could impart anything to my kids, I would want to impart my desire to look stuff up all the time. Recognizing now in my 40s, the degree to which that has accumulated. I am super powered by AI, not just because I'm willing to use it, but because I have so much contextual knowledge. I know when the AI is screwing with me, right? And I'm not going to like, like get X, Y, Z wrong. But there's also things like resilience and working through something. When you're struggling in a class, are you going to buckle down? Are you going to have the courage to go talk to your teacher to work at something and to get better at it, or are you going to have your parents go complain, sign you up for a class? And yeah, you got a good grade. But number one, did you actually learn something or were you just sort of taught the formulas to sort of like reproduce an answer? And number two, did you actually dig down within yourself to work through it and figure it out and maybe fail? Like, the degree to which we avoid failure is a big problem. Yeah, not just to handle failure, but failure is how you like. It's cliche, but it's a forgotten cliche. It's one of those cliches like, yeah, we know everything can't be measured. We know you need failure to succeed, except no one actually operates that way. No one actually sets their kids up to fail. And I'm not. That sounded wrong. You're not trying to set your kids up to fail. But failure is useful and to sort of work through something and figure it out. And you know, the. This is something that I actually credit. You know, my wife doesn't listen to podcasts, so I can tell her, but like I was, I was more. Even though she is Taiwanese, I was much more inclined to be the sort of tiger parent and wanting to help my kids and make sure they have all, every advantage and sort of getting ahead. And you know, xyz, and she was much more of the, like, let them figure it out even from a very young age. Like, they're smart. They can figure it out though, even just dumb stuff or whatever. Like that has nothing to do with academics. And it's something that I emulated and try to learn from. And also turned out I picked out a career where I'm pretty busy and I can't be deeply invested and the kids just sort of have been, have been raised that way. And I am feeling pretty salvatory because, you know, like, my daughter's like going, you know, she's going through college applications. These on those lines wouldn't even let me look at them. She's like, no, she's. She's doing it all herself. We didn't hire a consultant. We do didn't do xyz. She hasn't gone to bushy band. She has good grades, not the best in her school because she doesn't have an A plus in every sort of advanced class probably in park. She's not going to tutors and sort of xyz. Maybe she's not smart enough. I don't know.

A (84:52)

Whatever.

B (84:53)

I think, I think she's pretty smart but whatever and, and you know applications coming in and she got into great school. So congratulations to me.

A (85:01)

Absolutely. And her.

B (85:04)

That was meant as a joke. I think it probably came across the wrong way.

A (85:08)

I wasn't looking. I clicked out of the camera for that part.

B (85:11)

Yeah, no, no. I'm glad you called it out. So I realized that came across the wrong way. It does. I don't want to over index on me in my experience but what I do feel very pleased about and confident about is there were times that were hard and there were times that she had to overcome difficult things and work through them on her own. I didn't go fix them for her. She is empowered and independent and maybe overly so. I mentioned I'm very disagreeable. Like there's genetic components to this. She's going to do it her way for better or worse. And it's funny because you have like we're in the middle of this and there's lots of kids and that got all the good grades and they didn't get into their schools. And I'm not trying to. This isn't a taunt or anything otherwise other than to say like what? The grades aren't everything. What like what what in defense of universities and snowing these days wants to defend them. But I do think there is validity and this is a critique I have about a lot of the complaints about college admissions generally the obsession with grades and everyone who got a high SAT score is the only thing that should matter. And by the way, I'm pro sat. Like I think standardized tests are good. I grew up in a pretty lower blue collar like went to a non prestigious school. The reason why like or even you fast forward as an English teacher why did I get into a top business school? Because I had good test scores. Like test scores I think level the playing field even if there are advantages that come from going to classes or sort of xyz. So this isn't an anti test score rant but there's a lot more to success in life than good test scores or good grades. Like the people that succeed think for themselves. Our whole complaint about the current generation of sports writers is the lack of independent thinking and there's a general sort of lack of resilience. My big critique of parenting generally is exactly what you talked about this sort of helicopter parenting approach that by the way is miserable for the parents and it creates kids who live up to the paper marks. Yeah, they got good grades, they did all the extracurriculars xyz and are deeply, deeply uninteresting people who are not, who are not going to do anything. Right.

A (87:31)

Yeah.

B (87:32)

Congratulations. You can go to a good school and then be a consultant and then go work for a company and I guess that's good from a material perspective. You'll make a good living and you will not have any impact on the world.

A (87:44)

Yeah, yeah.

B (87:44)

I mean so.

A (87:46)

Well, good takes. My son is 21 months old so we're not at the tutoring stage yet. My wife informed me on Sunday that Charles is at stage three on Mildred Parton Newhall's six stages of play. So Charles is at onlooker play where he watches other kids plays and learns social cues without participating himself. He went to his first birthday party on Sunday and a two year old girl just came up out of nowhere and shoved him to the ground. And he was resilient, he was unperturbed, did not cry, just looked up at the two year old girl like what the hell is your issue? And went about his business. So his tutoring era is still like 10 years away.

B (88:34)

I think there's two areas for tutoring just to defend it. I think number one, if you're just really struggling like, like one to one is like the best sort of educational experience is like perfectly customized to the child. Right. So like if you could do like all tutoring and like and then move the child, that's probably ideal. But if you're in a industrial level assembly line schooling situation, like if you are behind, definitely useful. And by baptine I don't mean you're getting B pluses or A minuses, like you're getting Ds, right?

A (89:04)

Yeah.

B (89:04)

You just don't know what's going on. And especially things like math, if you fall behind in your foundation it's going to be a struggle for life because it's to the degree that it sort of builds on itself. So that's definitely valid. The other one is I think there's kids school's too easy and they're actually deeply interested in a subject they want to push further. And that's another example where I think a tutor is totally appropriate, where you're leaning into your kids strengths and you're enabling them. I mean it's horrible these school districts that are like cutting off like the advanced classes or Trying to keep kids sort of locked in together. Like kids are different, they're good at different things, they have different strengths. And I think one thing as a parent you can lean into your kid. Where I have a big problem is the broad based. You have to get perfect grades. It doesn't matter if you're interested or not. You're going to do sort of X, Y, Z.

A (89:52)

See, that's my instinct as I imagine how I'll handle all this. If you're doing it because of social pressure from other parents or because it's something you think your child doesn't like and just needs to be good at to succeed in life and have the right profile for colleges, it's going to feel like work for the kid. And I think you probably see diminishing returns from any sort of tutor in that scenario. But I also, when I was about 16, I had the sort of experience you were describing with parents leaning into their strengths where my mom recognized, I mean, I was starting to really like English and my mom took note of that and I really liked writing. And there was one teacher I had who was like 24 years old, he had played college sports and he just made it cooler to care about English. He was an English teacher. Mr. Piazza, shout out to you. So the summer after my sophomore year I had a job and then for one hour a week I would meet with this teacher. My mom hired this teacher to just meet with me for an hour every week and I would write a 2,000 word essay, we'd talk it through and then he'd assign me something else for the following week. And I did that the entire summer. I'm not going to say that whole experience directly led me to where I am today, but it helped solidify my confidence in my writing and my interest in pursuing writing further as I continued through high school and college and beyond. And that's an example of where I think tutoring can be really great if you're leaning into something that your child already likes. And so I think if you're hawken, I would lean toward that as opposed to the like socially enforced tutoring to get into the right university. But it sounds like he's already not interested in going that direction. But those are my tutoring instincts.

B (91:41)

It's so, it's so. You see this in lots of places. It's so hard. And this again, I think there is a big tie in back to this, you know, for kudos to your mom. But sort of in present day the spreadsheet mentality I'm Talking about where if I feel like I can control it, I have to control it. Right. Like, yes, I theoretically, I know colleges don't just care about grades. Yes, theoretically, I know that, like resilience and things actually really matter for success in life. But grades matter too. Would it be better if they got all A's and then totally. Like, it's just, it's really hard to push back on it and it's hard for the kid. Like, again, I might be over indexing on being in a nation country and sort of going through this.

A (92:26)

No, it's like that in northwest dc. I mean, you could talk to Bill Bishop about the environment in Northwest dc. It's like that here. And there's a ton of social pressure and it's insane.

B (92:36)

Yeah, it's so crazy because we were kids, like the, you know, the all good grades, they're the nerds. Right? Like, it's, it's amazing how that is the sort of prerequisite now, but, you know, once you've submitted some applications or whatever. And my daughter came in, she's like, she's like, I thought I had good grades, but I, I heard about some of my other friends, blah, blah, blah, and like, you know, xyz. And again, she's like a minus, like, sort of, sort of level, by and large. I usually give to personal evasion, but, like, they're good. It's not, I know she's not like valedictorian or anything on those lines. And it's like you get like very. I remember, like, I was at this, I was in business school, right? Everyone. I was always a big proponent of, look, I want to work in tech. I'm not going to apply to a consulting job, not going to do xyz. And then everyone, the consultants all hire their first. Like, they're super early in the process. And so you have all these people that are getting jobs or whatever and you start feeling like I've done that. Like, what if I don't get a job? Like, what am I going to do? Right? Like, and it's like that sort of environment interview.

A (93:30)

Yep.

B (93:31)

Right. And I told her, I'm like. And I. This is actually a bit of parenting too. There's a bit, I think everyone, as they grow up and get older, there's, at some point you're always going to be mad at your parents for something. And I was actually always a little reserved for my parents for not pushing me into my strengths more. And I'm like, like, no one's told me to go to A good school. I just went to University of Wisconsin because that was like, that's where smart kids went. Like, in Wisconsin. Like, I was, you know, town, a small little school. No one even told me to apply, like, somewhere better. And I was always, like, sort of like bitter and resentful about that, even as I'm sitting here. Extremely successful and widely regarded as being smart. It's worked out things, right? Yeah, it's fine. And honestly, I think I benefited greatly from not going through the quote unquote, elite pipeline and having you're also wearing

A (94:15)

a Wisconsin sweatshirt as we record this podcast, which is really hot.

B (94:20)

Right.

A (94:20)

Sorry.

B (94:21)

But the, the. Yeah, I always, like, I've always credited, like, I've gotten so much mileage out of being right on Facebook. And I'm like, that's just a function of me coming from normal people where they actually care most about being connected to their friends and family instead of all you weirdos that are running away from them. Right. Like, and so, so the. I've been. I remember I was going through this. I'm like, I'm like, this is that point. I'm like, in my 30s. I'm like, why am I hold. This is ridiculous. Like, it's my life. I'm responsible for what has happened. Like, they did what they did ever since then. It's up to me. Holding on to any aspect of this is dumb. And there's a bit to there where that actually made me a better parent in that, know what? Someday my kids are going to be in their 20s or 30s, they might be mad at me for choices I made. And you know what? That's going to be their problem. They're going to have to deal with it. It's going to be their life, and I'm just going to be the best parent that I can make. The choices that I think is right, some of which are going to be wrong, but it is going to be what it is. And you know, I told that to my daughter when she was there. I'm like, look, I've not pushed you on grades at all. I didn't send you to tutors because I don't think that's the right way to approach it. I think high school is 5% of your life and ideally should be the most. Some of the most enjoyable aspects of your life. I'm like, look at some of your friends who are miserable and they hate their parents because they're nose to the grindstone sort of constantly. I'm like, I just don't agree with that. And I'm like, maybe it's the wrong choice. Maybe all that, like, you did need better grades, you're not going to get into. She had her heart set on one school she really wanted to go to. And I'm like, maybe it's not going to happen. And if that happens, someday, you're going to have to forgive me for that. But that's the choice that I made, and I think it's the right one, but that's just sort of what it is. And like I said, I'm flying high because it worked out. She got in.

A (96:15)

She got in.

B (96:16)

But even if she hadn't, you know what? Like, I did what I thought was right, and it's going to be her life. And I think that's a broader point to think about your kids. At some point, you do your best, and the best thing you can actually give them is letting it be their life. And the way for it to be their life in the future, if you want them to forgive you when they're adults, is let them start thinking for themselves today. And to be grow up feeling and thinking that I'm in control of my life, it's my choices, I'm accountable. And when you fast forward, that'll actually be to your benefit in the long run because they already learned. They already learned. It's not your fault, it's their fault.

A (96:59)

There you go. Well, TikTok always gets a little profound, but a nice note to end on and also a huge congratulations to you and your daughter and thank you for the question. Hawkin. Thank you to everybody who wrote it.

B (97:14)

I've always like, I'm like, like, I do want to do the victual. I see. I was totally right. My approach worked. But of course, it's an anecdote, but

A (97:21)

it's not about results, it's about the process. You know, there is a little bit

B (97:25)

where, like, some are like, her, like, more like crazy, intense friends. Like, I feel really bad for them because they're devastated because the whole college process is weird. It's much worse than, like, like when we were younger. But this whole, like, early decision thing versus early action versus regular decision that existed in like a couple schools when I started out. Now it's like sort of everything. And so the early decisions, things are coming out. And like, some of her friends, it's like, it's like believably stressful all this time, all A pluses. And I then I didn't even get what I wanted to. And then it's like, you know, wait,

A (97:59)

she's just sailing through.

B (98:01)

Yeah, you're like you. It feels what. Hey.

A (98:05)

But hey, it's a great way to enter the holiday break for her as well.

B (98:09)

So the vacation is going to be so much less stressful. I was already, I was already thinking, like, this vacation is going to be very binary. Either my daughter is just going to be stressed out to all get out the entire time, or she's going to be in sort of like the best mood ever. So unfortunately, I think it's.

A (98:25)

Honestly, it's one of the best feelings in life. When you get into college, you don't have to care about school for the remainder of your senior year. That's not prescriptive for your daughter, but that's certainly how I handled it at 18 years old. And. And you just. That weight is lifted. It feels like the most important decision and sequence of your entire life at 18 years old. In reality, it's really not, but it's hap. I'm happy for her that she has the weight off her shoulder.

B (98:52)

It's funny, this is such an area where my background is just so much different and maybe, like, this worked. I got lucky that it worked out because I don't understand, like, the elite world and elite parenting. Like, for me, of course I'm gonna go to Wisconsin. Of course I'm gonna get in, like, like, even though I didn't actually get accepted to, like, March, it was like, it, it was, it wasn't very sort of dramatic sort of at all.

A (99:14)

Right.

B (99:15)

But so it's funny, like, yeah, it's so, yeah, maybe don't take advice from me. I have no idea. I just looked into everything so.

A (99:21)

Well, I come from a very over programmed, uptight, elite ecosystem here in D.C. and once you get past that first acceptance, man, oh man, life gets so much easier. So enjoy the holiday break with your daughter. Ben, we didn't get to hit everything on this rundown. We'll roll some of these over. We're going to do a mailbag in early January coming out of the break. Thank you to everybody who's written in all year long. God, I love doing this podcast. Always fun. And I look forward to lots more in 2025. So I will talk to you on the other side, my friend.

B (99:58)

Talk to you in two weeks.

A (99:59)

There you go.