447 Niching Down on Weird Data with AI Solves Ice Cream, Movies, and Power Bottlenecks | The Pirate Street Journal

0:00Thanks for pressing play and welcome to this very special episode of the Pirate Street Journal. My name is Christopher Lochhead and yes, I am also the host of Follow Your Different. And what you're about to hear is something that I've been dreaming of for quite some time and we got a lot of feedback that people wanted us to do it, so we're doing it. And so what the it is, is a regular look at kind of relevant business news through the category design lens. it's predicated on the crazy idea that, you know, the Wall Street Journal doesn't know how business works. And as outrageous as that sounds, it's true in the context of most business press is focused on companies, maybe products and technologies, but almost never market categories.

0:49And so what the Pirate Street Journal is, is we take three important business topics of the day. And Eddie, Bree, and I look at those topics through the category design lens and share them with you and in and out in 30 minutes. It's the Pirate Street Journal. It's a very different look at business, and I sure hope you enjoy it. And we're going to be releasing PSJs on a regular basis for the foreseeable future and drop our regular follow your difference that are either one-on-one with me or a conversation with a legendary guest. So without any further ado, this is the Pirate Street Journal on Folly or Different.

1:26Hey ho, let's go.

1:42Christopher Lochhead: Topic one, China now controls 90 % of the world's battery cells. America can't fix that soon, so Pioneer rejected the premise and put an entire power plant on a truck. Topic two, what do a Ben Affleck AI startup, a downtown Los Angeles ice cream shop, and a seven-person software company have in common? Niche down AI. And topic three, the most hour obsessed business on the planet, management consulting, just started billing for outcomes instead of hours. Salaries may soon be outdated as well. We are recording on Friday, July 24th. hey ho, let's go. Topic one, pioneers in power. The Wall Street Journal ran a piece this week.

2:21Christopher Lochhead: The key to solar and wind power is battery storage and China is dominating. Five years ago, China had under four gigawatts of energy storage and today it has around 155. It's targeting 300 by 2030. The entire United States has 57. Chinese suppliers now control 90 % of the global market for storage battery cells. More than 90 % of the battery systems installed in America last year ran on Chinese cells. The top 10 storage cell makers on Earth are all Chinese, and China's new data centers are required to pull at least 80 % of their power from renewables. The easy read on this is that China won the power layer.

3:00Christopher Lochhead: Game over. We don't think that's the story. Power is the one square on the AI board that almost nobody owns, and the companies racing to fill it in here are not waiting for Washington to build them a grid. They're building their own. Elon just quietly bought a mobile power company for about a billion dollars, the kind that can stand up a power plant in 30 days and drive it to wherever the demand is. Tesla is selling mega packs to help entire cities with brownouts. Homeowners are playing arbitrage on the batteries in their garage. Mobile, distributed, local, power, you can move. So Christopher, does the AI energy race get won by the country that builds the biggest grid or by the pioneers who make power portable?

3:41Well, I think we're seeing multiple models, right? And I think it is going to tilt more towards portable, more towards network. And so, look, does the monolithic power of the past have a place? It does. Microsoft, via Constellation Research, is firing up Three Mile Island. I never would have thought I would have seen that in my lifetime. That's probably a good solution. What we do know is there are new smaller modular solutions out there that get networked together. And obviously, there are these mobile opportunities. And the interesting thing here is Elon Musk realized that the government was a critical component of his supply chain.

4:23And just like Apple in the past, when they said, you know what, we want to own the majority of our supply chain. And so they brought everything in house. Musk said, I cannot rely on the U.S. government or state governments to be my good partner in providing power that I would pay for. And so I have to have my own power capability. And the interesting thing that's going on here is we see this in multiple different ways. So this is one example of a massive tech entrepreneur duplicating a government service. Another is what MetaFacebook is doing with education. They can't get enough people to build data centers.

5:09So what do they do? They launch their own free community college where they will teach trades and guarantee a job. And so whether it's on the power supply side or the actual human resources side, And what smart entrepreneurs are doing today is say, listen, where government fails, I can't wait. The category can't wait. And so I have to do what the government can't do that they used to do. Well, I think that's completely on the ball because there's really two ways to play the game. And that's what category design is about. Let's change the game, not just play the old game. and that what China is really kicking butt doing is the centralized top-down game, they're dominating.

5:57They don't have nimbyism problems there. The party tells you what's going to happen and it's going to happen. They build a hundred-year plan and they have the local mayors compete to move up into the system and the like. And so we're going to dictate from the top and we're going to build it in the centralized areas. They're going to win the game because the government and nimbyism is just going to be a real problem here in the U.S. And I think, Christopher, as you said, Elon has kind of recognized that about that. Now, there's really, if we kind of double-click, we can unpack the whole power layer of the AI cake into three buckets.

6:31There's powered generation. Solar panels are one, obviously, if you had an oil field or LNG or whatever else it might be, right? Like, where do you create the power from? Where do you store the power? And then where do you distribute it? And I think the way that the signs are shaping up is that the game is changing and the way that the U.S. can actually win is not by playing the centralized top-down game. It's the decentralized all 10 ,000 flowers bloom is really the way to go. Because if you look at the Tesla portfolio in the energy business, there's really three big buckets. There's Megapack, which is tons and tons of batteries stitched together that huge cities in Australia or other places that have brownouts, they're buying and they stabilize the grid because they go right next to where the power plants are and they're sucking up energy when it's cheap because there are different times of the day when energy is more expensive based on when people are using it versus not using it.

7:32Oftentimes, at the end of the day, when people come home when it's hot, usage is really high. And that's kind of one solution at the centralized top-down level. And then there's the power wall on the other extreme. And so that's a consumer solution of like, hey, usually it's you have solar panels like we do in Hawaii, and you have a power wall and that stores it. And so that can be a great solution. But the reality is you don't actually need solar panels for that to work because time of day usage matters. The price of energy during the day matters. And that so much of the magic of the power that we need to solve for is not solving for the grid math overall.

8:14It's solving for peak usage. That's really, it's like when you're trying to call an Uber when it's raining and everyone's trying to get to the airport, right? Surge rates go through. It's really solving for that. And that the reality is, you know, consumers, if they understood the opportunity, would be like, would I like to have a little mini oil field in my backyard? Sure, why not? And I don't actually have to have panels to do that. I can have a couple of Powerwalls, store power when power is cheap and not being used. and distribute it when it's being, it's very expensive and everybody wants to use it.

8:47That would be a meaningful difference in the power. This brings up an interesting opportunity and something that's been lurking in the background for quite a while, which is, do we see a future where individuals, consumers, have solar power on their house? And now because the storage component gets better and better every day, hopefully, certainly every couple quarters, when will we be at a point where you and i can install solar on our homes which we both have store that solar and actually sell it back to the grid well so so that that'll be the interesting part is that that future is long and hard because panels are the logjam because of the raw materials and that China has a monopoly on that.

9:34I think what I am suggesting, and also to that point, selling back to the grid, to the Hawaiian electrics, to the PG, that's tricky. Some of them, that's a really complicated business model. Some of them have incentives to buy power back. Some of them don't. So in Hawaii, we did it too late. We are not grandfathered into the system. So when our solo generates excess power and we give it back to the grid, they're getting a freebie. They're not giving us anything for that. So I actually think it's slightly different in that you don't actually need the panels right now. If you just had the batteries and you just arbitraged time of use, that could make a meaningful difference.

10:15And the reality is in the category design of top-down, centralized, from bottoms up, decentralized, is where we're actually going. is that so some of it is imagine Christopher, not if every homeowner had solar panel, but everyone had batteries, which you can install. You don't have to mess around with the roof and all that kind of stuff. And there's a lot of hee-haw related to that. Batteries are actually much more doable on the bottoms up level. And then the middle category, so if you have kind of these mega packs at Tesla and power walls at the individual level is what's called the mega pot.

10:48Christopher Lochhead: And this is kind of interesting. I don't quite have the, you know. The mega pot? The mega, no. No, that's you. That's a different Christopher category. Now I thought we were talking about something interesting. Okay, keep going. P-O-D. So the Megapod is actually going to be a set of batteries coupled with GPUs. So here's the magic that you can get to, is that I don't actually know, imagine a data center in a box. It's not a humongous, you know, like Colossus like in Memphis that Tesla has. It's some sort of a in-between size with power generation and storage and also compute. The question becomes, who buys this?

11:31Is it, because is it something, so if we have kind of a government and the consumer, it's B2B. That's where I think this gets really, really exciting. If you have a B2B enterprise that has an excess capacity of land that's not being utilized, a megapod would make a lot of sense, not only to offset their electricity costs, but to actually generate some revenue from being a distributed data center. And so I actually think in the long game, the bottoms up, decentralized, distributed, you know, power layer is actually the most sustainable and the highest upside. Now we have to, someone has to do some math and modeling kind of what needs to be true with that.

12:11But this is the part where, you know, Bree, you tell me, what is the amount that an Alaskan citizen gets per year from the oil subsidies and the like? Is it like a thousand bucks a year or something like that?

12:23Christopher Lochhead: That's the average. Yeah. It's the average. I think that's going to be. I don't think most people know this, that the Alaskan government gives the people of the state. I love those bongoose.

12:40A thousand bucks a year just for being in Alaska because we got oil. Right.

12:47Christopher Lochhead: Right. Yep, anywhere from like$800 to$1 ,600 a year, depending on how well oil does. Hey, move to Alaska. Free oil money, people. Move to Alaska. Well, and then partly, I assume that's how Texas is able to not have any income tax and whatnot, right? Is that the mega pod, not the mega pod, the mega pod is actually going to bring the same kind of annuity opportunity that Alaska has and that Texas has that. an enterprise could have. A church could have a megapod. You know, a farm could have a megapod. And that these are the things that I think are the really, really interesting things. Okay, topic two, ice cream, Affleck and AI.

13:32Christopher Lochhead: Three stories that have nothing to do with each other, or do they? Story one, Ben Affleck, the guy the internet decided was the dumb one standing next to Matt Damon, just sold a stealth AI startup called Interpositive to Netflix for$587 million in cash. Here's what it does. It trains a small model on a single film's own footage, its lighting and its style, and uses that to speed up post production. Story two, a downtown Los Angeles ice cream shop called 28 Wishes. Ice cream sales drop about 20 % every time it gets colder than 70 degrees. So the owner bets$20 on a day on cold weather forecast on a prediction market.

14:11Christopher Lochhead: That covers about 43 % of his rent. And story three, AI native companies, teams of seven doing the work of 50, companies hitting$10 million in revenue with fewer than 10 people. A movie star, an ice cream shop, and a seven-person startup, same move every time, niche down, use the weird data nobody else is looking at, and do more with less. The riches have always been in the niches. AI just made the niches enormous. So Eddie is a niche down on weird data, the future of the AI economy? Yeah, I think what this shows is that as much of the upside might actually be not from the huge, you know, Fortune 500.

14:50Certainly the Mag 7, 10, 12 will generate a huge amount of the upside that we're talking about in AI. But I wouldn't underestimate the amount that can come from the SMB side of things, right? And so what of the story that I think, I think, Christopher, you've talked about not large language models, but small language models. And I think the niche down one, what was really, really interesting about the Ben Affleck AI startup was that they're not using it to rip off other artists. As I understand it, they're using it to replicate the current artists. And so what they're not doing is stealing stories, stealing images, stealing any of that stuff.

15:26What they are in fact doing is taking existing footage shot by a director in that director's style, the way they They like to light it in the way that they're trying to do. And using that as the base training model to actually add in effects for the future parts of that person's film. So it's not really applied so that I can steal Christopher's or Breeze stuff. It's that I can write the first draft or shoot the first part of what I want to shoot. and that Ben Affleck's startup, Interpositive, will actually take that style and lighting and whatnot and apply it to subsequent parts so that it reduces the post-production time.

16:04And so, you know, Netflix is not dumb. They are going, you know, full steam ahead on their original content and they want more content. And so I think what will end up happening is that the cost, the timeline of post-production will shrink, therefore more content, therefore more binge-watching. You know, Netflix, you know, created the category of binge-watching and they're just going to keep loading the buffet. And so this is a really good one of, it's not an AI engineer coming into film, but an artist equipped with AI creating a niche down version that makes and creates abundance for more art.

16:40And so you have that one. I think that's a really powerful example of the natives kind of adopting AI for their own benefit. So there's an old saying, if there's going to be a cannibal, you want to keep the cannibal in the family, right? And so have it done yourself. The second one with the ice cream shop is a really tight one around what Amazon has done so beautifully, taking costs and converting it into revenue. Shipping costs became prime, became revenue. Compute costs became AWS. And what I think is really, really powerful is the ability for entrepreneurs, lowercase e entrepreneurs, to take the weird data that is specific to their businesses.

17:19and not just the weird data, Brie. This is, it's weakness data, right? Where is my business weak and turning it into a strength, which is what I think Amazon has done uniquely with this one. And that, you know, there's a lot of categories like this, ice cream, beer. They fluctuate up and down with the temperature and the ability to take that data when their business is going to be slow and convert it into a hedge that generates revenue so that they generate, you know, nearly half their rent is a genius way to run the business. And you just kind of wonder, what are the other things that you are sitting on from a weird data perspective and a weakness data perspective that could actually be monetized in another business model?

17:59And so what AI is actually doing is not only, it's the other side of the magic triangle, right? It's not just your product and your offer. It is the company and what your business model you have. You have many more different business models at your avail than you had beforehand, thanks to AI. and what better use case of it than to take the weird data that I have that I don't really recognize as weird and the weaknesses of my business and turn them into strengths. And I think that's what you really are seeing with that third story is that three different really specific use cases. And Brie, what I love about this one is the journal article talks about 15 % fewer managers and 15 % lower costs as a result of this.

18:35And so the reason why college has gotten so expensive is that it's not that there are more professors, it's there are more administrators. The managerial class has driven the cost of college through the roof. And if I were running a university, there's a lot of administrators probably aren't necessary that AI can do the work of. And if you could cut that by 15%, you could cut college costs by 15%. And that you would actually see really meaningful difference in outcomes when people can graduate without all the debt that they have there. So what I love about this is that it's not all the surplus economics from AI are not just accruing to Anthropic and to OpenAI and to the Mag7+.

19:18It's actually going to be more prevalent, I think, because those businesses are optimized for AI. Most of these kind of medium SMB businesses and lowercase e-entrepreneurs are not. It's probably where some of the greatest opportunities are, which is where, for me, someone who grew up in consulting, serving the Fortune 500, it might actually be the opposite that consultants serving SMB using AI as your co-founder might do even better. All right. Well, first of all, we have to discuss, Bree, the fact that on one hand, insurance company Affleck made a giant marketing genius move when they hired Ben Affleck to be in their Affleck ads.

20:02that was legendary what was not legendary is they stopped the campaign because if i was the cmo of afflack the company i would have had an ongoing set of campaign ads where ben affleck walks into somewhere and people say afflack and he says no affleck and you know etc but i digress okay the interesting thing about this i was just on the ben affleck side and i'm not a big fan of celebrities and stuff. So, you know, I don't love that. But here's what, there's an interesting thing to know. Ben Affleck, as far back as the late 90s was hanging out in Silicon Valley and investing in various angel circles and so forth, he was smart enough to realize that he could convert his celebrity into being a VC.

20:50And several celebrities have done that. We had Andre Iguodala, the of the warriors on my podcast, he said one of the main reasons he wanted to come to San Francisco and play here was exactly this so that he could get in the startup ecosystem. So that's very smart. Now, that said, what's the other common thread? So Eddie showed us a few common threads. What's the other one? All of these people saw what you might call a missing, saw what you might call a problem or an opportunity. And they looked at it from, and I'm going to use these words on purpose, a very different lens. So remember when we talked about the Invest America accounts that Brad Gerstner is the category designer of for all the kids, right?

21:37If you look at the wealth inequality problem and you say, well, there's this horrible wealth inequality problem and the problem is the rich people have too little and the poor people have too much, the rich people have too much and the poor people have too little. And so we need to take from the rich to give to the poor because that's more, quote, equitable, right? That's what they say. However, when you flip the lens and you say, well, wait a minute, how could we make everybody more wealthy than different solutions show up? And so what's happening here across these is how can we use AI to solve a problem?

22:12And invariably, of course, a problem that the category designer is experiencing themselves. And so they want to solve it for themselves and then give it to the world. This is an unlock for everybody. And the thing that's different today, everything I just said was true before AI, except for the give it to the world part. So anybody who sees a problem, anybody who sees a strategic missing, anybody who looks at something that the world has looked at one way, you look at it a different way, and there's a big unlock there. There's a breakthrough that happens there. Well, now with the power of AI, AI is the greatest creation technology category of all time.

22:53And so when those of us who see opportunities and problems want to get after it, the level of resource it takes today in a post-AI world to get after it, aka prosecute the magic triangle, build a legendary company product and category, the barriers of entry are lowering by the moment. I still love your Affleck and Affleck. I didn't hear that story with it. And to me, the ability for companies to use AI to find non-obvious opportunities, that to me is an example of kind of innovation because so many of these insurance ads, And, you know, the more I watch them, the more I say, that makes me less inclined to buy their insurance, even if I was in the market, which I'm not.

23:42And so I do think that. Well, and the funniest thing, too, about slogans is most people can't remember or taglines. They can't remember which company it is. So everybody remembers the tagline. Can you hear me now? Almost nobody can tell you which carrier it was. The thing about Ben Affleck is his last name is the name of the company. So we give you these prompts. These are the weirdest parts of my business. AI helped me turn this into a revenue opportunity. These are the weakest parts of my business. AI helped me turn this into a revenue opportunity. Because what I would say is in the insurance space, what Lemonade has done, everybody hates insurance, is a weakness of the insurance category, right?

24:22It's like you buy insurance, you file a claim, and they don't pay your claim. So for them to lean into the fact that this is a weakness of the business and we're trying to address it with our POV is a legendary use case of it.

24:38Christopher Lochhead: Topic three, the death of the billable hour. The Wall Street Journal calls it consulting's messy shift away from the billable hour. McKinsey now ties 25 % of its global fees to outcomes instead of ours. Its UK managing partner is on the record wanting more performance-based deals. Bain says 30 % of its business is now AI and tech enabled on the way to 50%. BCG expects AI work to jump from a fifth of revenue to 40 % by next year, and that's at a$12 billion firm. Meanwhile, the model these firms defended, like scripture,$300 to$500 a consultant hour is cracking. Some of them now have 150 former McKinsey, Bain, and BCG people under contract training AI to do the entry-level work.

25:24Christopher Lochhead: Consultants are the ultimate fast followers, the most conservative hour-worshiping business on earth, the last people you'd ever expect to touch the billable hour. So when they finally change how they get paid, pay very close attention. That is about as close to signal flair as the professional services world ever gives you. So Eddie, what does this mean for consulting and what does it mean for the rest of us? So I think for consulting, it represents a longstanding shift away from time-based revenue and pricing to deliverables. And then eventually the outcomes. And so let's be clear about what is likely going on because I believe the numbers that McKinsey and BCG and Bain are talking about, that the percentage of consulting projects that are outcome-based are rising.

26:11Having done some of these, what you do know is that the devil's in the details, is that when Christopher does it in Silicon Valley, the outcome is very simple because it's usually the incentives are aligned with stock. The stock go up, did it go down? everyone has the same scorecard. Within consulting, that's a little tricky. I think there are examples of consulting firms who have taken on equity in exchange for projects, but that's not the norm. And that most likely what's happening within consulting is that there's really kind of a handful of types of categories of consulting. Things that are kind of revenue, growth-oriented, market cap, have so many X factors and variables that I would be surprised if a lot of the outcome-based consulting is on that end of the spectrum.

26:54And it's most likely on the cost, organizational, KPI type of thing, right? So we published in our latest mini book, The Agentix CMO, the difference between the set of metrics from KPIs to market cap. It's probably lower on the list. And before it hits revenue would be my guess. So one is that that's a lot more in the client's control, easier to measure the consultant against it. And therefore that's kind of what's going on. Now, why I think this is important writ large for everybody is that I think this shift of pricing, which is how you get paid, how consultants get paid from time to deliverable to outcomes is something that we talk and preach about in the Category Design Academy all the time.

27:34It's happening within consulting. If it's happening within professional services firms, you better believe it's going to happen at some point with a trickle down to the rest of how everyone gets paid. and that in the broad kind of spectrum of time, I'm making some large generalizations, right? Is that the way that everybody has been paid historically has been time-based. You were an hourly worker and then eventually you became a salaried worker. So it was not purely time-based, but a version of time-based and you had overtime and this and that. And then eventually you got to bonus and incentive compensation occurred and then you got to stock-based compensation, options and equity.

28:12And so if you look at the kind of history and the mark of time, even slowly, is that the large consequences that the way that people have gotten paid has shifted and that you might think this is only a Silicon Valley or CEO thing. Well, the whole kind of tipping craze that's going on about like, oh, I'm forced to automatically tip is a function of an hourly worker that has been, you know, paid a much lower hourly rate and given an outcome based thing that is not really outcome based because some of it's in the waiter or waitress's control, but some of it, the kitchen might be late. Some of it, you might be a high maintenance, you know, patron or whatever.

28:48And so suffice to say, I think what people kind of have to get ready for, not just consultants, is that the advent of, hey, I can trade time for money easily is probably going away and changing meaningfully. And so what that means is that there are two types of people. If you are somebody who is naturally a missionary and you drive towards outcomes, you're going to be great. You're going to be paid more, in fact. and that the ability for you to get paid in purely time is going to be a real benefit because you out-deliver, you out-perform, and you care about the outcome, and you're going to do great.

29:23The people for whom they are coat riders, free riders, and who are like, I don't know actually how to drive an outcome. I'm not even sure how to drive a deliverable. Some scary times are afoot for you, right? Is that you are going to have less certainty in your salary. You're going to have less certainty in an hourly wage, and more of it is going to be, quote, at risk with it. And so I think the natural consequence for, I would say, everybody to think about is how do, which bucket am I in? Am I in somebody, the category of somebody who, regardless of what I get paid and how I get paid, I drive out?

29:58If so, you're great. You should be looking for ways in careers and categories where you can get paid more in outcomes. E.g., you know, you can be, you know, a cashier anywhere. And we just kind of published in our recent PSJA, better off to be a cashier at Costco where you get stock options and the category is, they're a category king and the stock is going to go up versus being a cashier at somewhere else where you don't get options and the category is going nowhere with it. So if you're an outcome-based person, good for you. Congratulations. We think you're awesome and legendary. Go get paid by looking for a category king and choosing a career in a category that it actually works for you.

30:37If you're going into consulting where we started, look for those outcomes. Look for those firms that will take more at risk because you will make more money and you will build a more legendary reputation and reputation capital because of that. Now, if you are somebody who is going to be a little bit afraid of this, then you have to prepare yourself for the fact that you will not be compensated for your time as safely as you have been historically and that you need to figure out how do I actually deliver deliverables and outcomes because the hiding places are going to go away. No one cares how much time you spent on it.

31:12No one cares. No one cares. This is an example of being so smart, you're stupid. Management consultants are supposed to be the smartest guys, gals in the room, right? When you have a business model that is predicated on something bad happens for the client and something good happens for you, that's called a perverse incentive. So what am I talking about? When you charge for time, you have an incentive to sell more time. So if the project takes too long, that's good for you. When we built our home, there were penalties for not doing it on time, which is exactly the opposite of what is true for the incentive system for most consultants.

31:58And so look, this is just bad business model. It's dumb. This is smart people being really, really dumb. The reality is outcome-based pricing. That's the world that everyone else lives in, okay? If we don't produce outcomes, nobody buys from us. And that's true for the vast majority of industries. Produce value, produce outcomes, and you'll make zillions, charge by the hour, and you're a donkey. I remember, Christopher, one of, at Cambridge, our longtime super consumers, this guy named Jim Kiltz, who was the CEO of Kraft often than the CEO of Gillette. And the founder of Cambridge, Rick Cash, had met him early on in their careers when Jim, I think, was the brand manager of Countrytime Lemonade.

32:44And so that relationship goes all the way back. And then when everybody, you know, fast forward a couple of decades, you know, massive outcomes for everybody in the career on both sides as Jim went up through the ranks and then continued to hire Cambridge. Jim was great. He would always kind of, you know, sit there in meetings and, you know, smart people get up there and present and stuff. And his version of you're being so smart, you're stupid, Christopher, he would just sit there and say, I'm a simple man. And when you heard that, you're like, oh, crap, I really flamed out on this presentation, right?

33:16But, you know, when we asked him for a testimonial, you know, Jim, why did you keep hiring Cambridge? He goes, every time I worked with you, I got promoted. Period. End of story. It's not complicated. Deliver outcomes. Deliver legendary value that others cannot deliver. Period. Yeah, but how many hours does it take you to charge you an hour? Nobody cares about a nose-picking discussion about how many hours. Literally nobody.

33:46Christopher Lochhead: We'll see you next week, Pirates.