You're watching TBPN. Today is Tuesday, September 22nd, 2026. We are live from the TBPN Ultra, the Templar Technology, the forks of finance.
Let me tell you about ramp.com. Little teaser. That's foreshadowing. You see me wear the yellow suit. I tell you about ramp. Time is money. Save both. Easy. Use corporate cards, bill pay, accounting, and a whole lot more. All in one place. Uh, Jordi, his hat game has evolved. He's now wearing the fedora with the Blackstone cap on top, which I assume can only mean one thing. It means that your top priority is safety in financial markets.
Financial markets and the financialization.
No, but also safety as well.
Oh, you care about both. I care about both.
So, you want to accelerate the AI buildout safely.
Yes. I thought it was, I thought it was you're going to blend the two and you're like, 'Look, I want, I want insane private credit deals, trillion dollars of capital to flow into the AI buildout, but I wanted to do it safely. I want to prioritize investment. I want to prioritize safety on the whole frontier.'
Safety around the models, the capital markets and financial... in the capital markets. Got it. Okay. I'm seeing where you get there. Well, there's no safer place to participate in the capital markets than on the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. Just do it. Stop making excuses.
Ridiculous. Uh, anyway, I got some Harvey takes. Should we run through those? You want to do Harvey first?
Okay. People are taking shots at Harvey. It started with Samuel. I won't stand for it. No. Uh, I like Harvey. We've talked to some lawyers that use Harvey. Seems like a great product. We also had Logora on the show just last week. Had a good time with them. Seems like a very competitive space. But the headline that grabbed attention actually came from Bloomberg reporting. But then this company called the Redline HQ actually did a pull quote on that, got the reporting, repurposed it. So you don't even have to go to Bloomberg. We'll talk about that later, too. No one's going to news websites anymore, but uh, they say track changes. Harvey's gross margin fell from about 50% to negative 50% by June as agent token use spiked 20-fold on rented OpenAI and Anthropic models, Bloomberg reports. And so it seems really bad. Like you definitely don't want negative gross margins as a software company, as a tech company.
As really as any company, actually, you don't want to be selling a dollar for 50 cents, and that's exactly what they did in June, basically. That's what they were doing. They were selling tokens at negative gross margins. Not good. Um, but there's way more details in the actual reporting, so you should go read the Bloomberg report, subscribe, and actually read it because there's way more details. Uh, so the first thing is like cost did explode, uh, but it was on the back of more product usage, which is good. They're growing ARR. We'll get into that. Uh, and token consumption increased 20x this year. So basically they are feeling the reasoning era, the agentic era. Uh, one job, one process in Harvey is generating a lot more tokens.
Had seat-based pricing. Yep.
And then AI got a lot better. So every seat was just using it way more. Very simple.
But it's not just using it more. It's that the lawyer might be saying the exact same thing as they did a year ago. They might say review this. But instead of like one-shotting it with no reasoning, they went to reasoning which generated way more tokens. And then they went to agents. So they're like, 'Okay, it's actually going to fan out, look at everything across your firm, all the other records that you have, all your different policies and MD files, like everything that you have is going to be worked through this agent.' And so you're generating way more tokens. So that created negative gross margins. But the good news is that they're already back to positive gross margins, which is in the Bloomberg article. Kind of criminal not to repurpose that because it really makes it look like they are running at negative gross margins. They had a negative 50% gross margin in June. Uh, but they are back to positive gross margins. Uh, they changed their model usage, of course. There's a whole bunch of models that are cheaper and better for certain things, and they actually fine-tuned an open-weight model. So they did post-training on an open-weight model, and they're calling it Harvey Tenant. And there's some interesting details there. I want to debate with you on like the value of that, where that goes. But first,
First, Gabe, one of the founders of Harvey, responded.
He said, 'The hardest thing about building Harvey is doing what's best for our customers despite immense pressure to do what's easy. The easy thing would have been to force our customers into consumption pricing before they were ready and serve them worse models to protect our margins. We chose to help our customers transition on a timeline that works for them and give them the best models in the meantime, even though it hurt our margins. This meant optimizing our product through routing, harness improvements, and post-training so we could serve frontier intelligence at an affordable price. It also meant building the infrastructure for customers to monitor and manage spend: usage dashboards, per-matter cost attribution, spend caps, and ROI reporting. As a result, we improved our gross margins from negative 50% to positive in a single quarter despite usage doubling month over month and continuing to serve the best models. Our philosophy is simple: do what's best for our customers, even when it's painful, hurts our margin, or draws criticism from competitors, X, and the press. We believe the most important part of building a company is earning and keeping your customers' trust. You do that by doing the hard thing for them, even when it costs you.' Um, bunch of great points in here. Uh, even outside of the Bloomberg piece and the subsequent coverage, uh, Max from Logora was going pretty hard saying it doesn't make sense.
Uh, doesn't make sense to train your own models because the public basically, like publicly available frontier models are dropping in cost and the quality is increasing. So you're just sort of like wasting your own... And again, like that's a bet that he's making. Harvey's going to make another bet. Um, probably some mixture of both is going to be the right approach. There's another interesting sort of like how bad is this situation. Just comp, which is, so they had negative 50% gross margins in June. Uh, they're at a 400 million run rate for ARR, I suppose, and so that's $33 million a month. And so negative 50% margins means they lost $16 million, but they've raised like 500 million. So in terms of like burn, it's not super cause for crisis in my opinion. And uh, and what's interesting is that even through June, maybe they did the Sequoia deal that valued the company at 15.5 billion before that happened, uh, because it was announced September 9th, but I have to imagine that the deal probably got done with at least some insight into those negative gross margins in June. What do you think?
Uh, I would assume so. But the main thing is like every application layer company has like gone through a moment like this, even if it was for like a single day.
So it matters how quickly you can respond and adjust.
Yeah. Uh, so where do they need to get to? Like what is the standard for the legal technology industry? There's some comps here. Uh, e-discovery provider Disco, public company, 75% GAAP gross margins. Like it is a software tool that when you're going into discovery, you need to connect with a bunch of data. Okay, we got a whole bunch of text messages, a whole bunch of emails. Let's put them in a system so we can review them. Sort of, you know, normal SaaS. I'm sure there's a bunch of AI involved now, but that business is running at 75% gross margins. Uh, Thomson Reuters has legal professionals segment of software that's running at almost 50% adjusted EBITDA margins. So like true net profit. Um, so that's sort of the way they want to be. Also, pure-play law firms have really high margins as well. Um, hard to comp them to businesses because they pay out to the partnership. Um, but uh, PricewaterhouseCoopers reports net profit margins around 41% for the top 10 firms. It's lower for firms 10 to 100 or something like that. But still, uh, lots of health in the legal industry broadly. Obviously, everyone knows lawyers make money. Uh, that shouldn't be surprising. Um, and uh, yeah, tokens are getting cheaper by the day. Uh, there's a bunch of model launches. Opus 5.5 launched today. Grok 4.7 was yesterday. Uh, and Grok 4.7, no one was even... Elon was like, 'Okay, we're not jumping straight to the frontier with this. It's going to be a couple more iterations, but we're excited with the progress.' But there was one benchmark that really jumped out, and it was actually the Harvey legal agent benchmark, which is interesting. So I don't know how insanely good that benchmark is. It might be, you know, you could benchmax against it. But the interesting thing about that particular benchmark was that it was a legal task benchmark based on cost. And so Grok 4.7 is particularly cheap for the type of legal work that at the very least Harvey wants to do. So you could see them funneling more token spend towards xAI and Grok. Do you have something? Are you...
You wanted to give it up for Big Law?
Yeah, the chat wants to give it up for Big Law.
They're the real winners here.
Uh, additionally, GPT-5.5 and Luna are starting to roll out, so we'll look out for announcements there.
There were some cool demos. Uh, I saw one, uh, Opus 5.5. I mean, chat says the combination, the hat-jacket combination isn't working. I'll try to adjust a little bit. I'll try to adjust. Hopefully, that's better.
Flawless. Yeah. Thank you. Thank you for flagging.
Huge upgrade. Um, very funny. Um, yeah, I saw a cool demo where someone sketched out on a piece of paper sort of a trebuchet and Opus 5.5 took that image into simulation and created like a virtual version of it. Uh, the AI continues to impress in terms of translation of things. Oh, your kid drew some crazy mythical creature. Turn it into a story. Turn it into a movie. Turn it into a video game. Whatever format you want. If you want to read a 2,000-page novel based on Call of Duty: Modern Warfare 4, like you can probably do that now. So you can have your content in any specific format you want. That seems to be where the AI continues to impress. Uh, anyway, let me tell you about Codex. Codex is a powerful workspace for getting work done with AI agents. Whether you're writing code, analyzing data, creating content, or automating business workflows, Codex helps you move projects forward from start to finish. I've been having a lot of fun generating AI videos with Codex.
Should we pull up your latest masterpiece?
Yeah, the 'Literally Me' compilation. It burns up the internet and vibe reels all of the time. But now if you see those 'literally me' characters and you're like, 'I am Jake Gyllenhaal in Nightcrawler. I am... I forget who else is in there. Uh, Patrick Bateman in American Psycho.' Uh, you can put yourself in the movie. There's Inception. We got Jordi in there. That's Drive. I think that's Heat. Mad Men.
Can we call them out? I think that's Fight Club. You got Joker. You got Taxi Driver. You got Nightcrawler. No Country for Old Men, Social Network, Godfather, back to Joker, I forget that one. Shawshank, maybe that one's Collateral. Then Blade Runner. Blade Runner again, I think. Yeah, some of these renders are really, really good. That one, like your face doesn't quite fit on Leo's head, but that one's really good. That one works pretty well. This one's a little awkward. The Joker, it's really just Joaquin Phoenix's face. Didn't actually swap you in.
It didn't even work. I was like, 'Do it anyway. I don't care. It's fine. It's going to be on screen for two seconds.'
John is really my human agent.
He's sort of like an ambient, always-on agent. That's what I'm saying. I don't even have to tell you. You just predict my needs, right? You'll send me.
Everyone should have a John in their life.
Everyone should... But yeah, fun, fun, powerful tool. Still got 500 bucks locked away in Higginsfield in the wrong section. Can't do anything with it. So just give away the account or something. I don't know. Um, anyway, what matters more, people or insects?
We should get into maybe one side of it first before we form an opinion.
Okay. The human side. You want to steelman the human side. Look, and I'll just say that we're coming to you live. We're coming to you. We had a rough morning with the mosquitoes. We actually had to DoorDash one of those bug bite treatment things, the heat... whatever. What do you call them, Tyler?
Uh, I think it's like insect bite healer. That's what it's listed as.
It did work because I... Yes. Sebastian says, 'I'm famously pro-mosquito.' Yes, but a mosquito did bite me on the face today and a bunch more times. So we ended up... I ended up capitulating and having to get one of these. But I don't like using these. I don't like using these. It's funny because I can't stand EAs and I think they're weird freaks. Uh, and I don't think they should be anywhere near public policy or anything like that. Um, but the one thing that I can agree on with Ben, this guy writing this post, that insects are more important than humans, is I don't think we should harm the insects, right? I think humans are more important, but that doesn't mean you got to take them all out. You know, just go somewhere else.
You would be altruistic towards the insects.
Ellie says mosquitoes are terrorists of the air. Uh, they really are. You want to be altruistic towards the mosquitoes. You want to be effective in your altruism when it comes to the mosquitoes. You don't want to arrest...
No, I want them effectively to be nowhere near me.
But I don't want to harm them.
Okay. Okay. Well, but let's get into this.
Let's get into the Bentham's Bulldog post. First, let me tell you about Figma. Agents meet the canvas. Your AI agents can output and modify your Figma files with design system context. An effective altruist who writes under the name Bentham's Bulldog went viral over the weekend for an essay titled 'Insects matter more than people in the aggregate,' arguing that the combined welfare of insects outweighs that of humanity. His basic argument is one of scale. Insects are so numerous that even if their capacity for suffering is only a tiny fraction of ours, their total suffering could still dwarf human suffering. It's almost just a thought experiment. There's so many other directions you could make. Why not go pound-for-pound? Maybe cows are the most important because there's a lot of cows and they weigh a lot. There's probably more aggregate pounds of cow than humans. Maybe they matter more than humans in the aggregate, right? Or fur, like the amount of fur on all the monkeys in the world is more than all the fur on all the humans in the world. And so maybe monkeys matter more. I don't know. It's completely arbitrary where people draw these lines, but uh, he estimates that for every second of human life, insects collectively spend roughly 270,000 seconds dying, about 75 hours, and notes that more insects die in a single second than the total number of humans who have ever lived. Wow, that's a crazy stat. I mean, the guy pulled some cool numbers together. This is interesting. This is fun fact territory. Uh, a large part of the essay turns on whether insects can experience pain. Um, Bentham's Bulldog argues that they can, even weakly. Uh, if they can even weakly, their suffering should count morally, just as ours does. What species you are is just totally irrelevant to whether it's bad for you to feel terrible pain. Using what he describes as a conservative assumption that insects experience pain at just 1/10,000th the intensity that humans do, he concludes the sheer number of insects would still produce vastly more suffering in the aggregate. He ultimately argues that insects may experience more suffering in a single day than humans have throughout all history. Now, is there a piece on this where this is the view of this community, with the relative... like are they looking forward to how a world might interact with a godlike machine superintelligence and humans? So the idea is like we got to stand up for the insects now because we are the insects of the future. So we want them to make the same calculus.
I think that's certainly like where the thought experiment goes. Yeah. Right. And so this feels almost like bait or like a trap where it's like if you are the type of person that jumps on this and says, 'No, I disagree, like insects don't matter more than people in aggregate,' then you're setting yourself up for like the dunk that's, 'Okay, well now there are a quadrillion super intelligences that can all feel pain in some abstract way and so they're more valuable and now you're the insect.' And so your previous argument can be used against you to advocate for your own eradication. And so it's like stand up for the insects now lest you be discarded in the robotic future.
I don't know. Um, the reactions to this were rough. Uh, people did not like this argument at all. Well, it's just the timing is funny because there's so much tension building around the sort of EA rationalist movements with... which is super fractured. People don't realize how fractured... but it's sort of boiling up to the sort of national level in a way that it hasn't before. Yeah. And so during the midst of that broader conversation to come in and being like, 'Yeah, insects actually matter more than people.' Yeah. In the aggregate.
Uh, what were some of the reactions? Doomer said, 'If you truly believe that a bug's life is worth more than a human life, you will immediately feed yourself to bugs. Do it right now. If you don't do this, you are a fraud, a liar, a poser.' Um, Andre...
I think it's actually important for people to read the... before you go and comment on this article, watch A Bug's Life.
Yeah. Watch the whole movie at half speed.
You're a movie guy, but...
I think we coexist. Maybe you're not a fan of watching movies at half speed. It's not really how the...
You're really a half-speed guy.
But half speed is a great tool to have in the tool chest when you want to deeply comprehend.
2x speed, 3D glasses with two different movies interlaced so I can get one movie in one eye, another movie in the other eye, one audiobook here, one podcast here.
While I'm sleeping. Yeah.
Exactly. Max brain rot. Um, no. I think this is an odd... I mean, the fracturing of EA really leads to a really, really wide range of outcomes that get sort of lumped together. Everything from the 'Pace the Frontier' essay to this. These are clearly very, very different pieces of communication and yet they get sort of lumped together. Uh, and uh, it must be full employment if you're in the public relations department over there.
Yeah, I was going to say I think the shrimp welfare stuff makes more sense than this because like I'm not really sure what... like after reading this and saying I agree, like what are the next steps? Like we're not really factory farming insects, maybe through pesticides or something.
Speak for yourself, brother.
Yeah. Yeah. The shrimp welfare... got some big announcements coming soon. Like what percent of these like 600 billion insect deaths per second are preventable, where like with the shrimp, you know, you can say like, okay, change the farming practice or just don't eat the shrimp. Yeah. So I think that's why maybe this was less impactful. I don't know.
Yeah. I mean, it's all a continuum to like until you get to like the veganism, like the animals who you can pet, you know, should be able to live full lives. Uh, this is sort of the most abstracted version of that, but um, yeah, I don't know. It's odd. Uh, yeah, very, very... but I mean, I guess it's a good opportunity to go viral if you're looking for some... more importantly, update on TBPN's road to Christmas. Oh yeah, we're down to just 93 days. 93 days. A lot to think about, a lot of moves to make, and uh, we couldn't be more excited.
Start counting it down. Yeah.
I was trying to think of a realistic AI doom scenario. Like everyone's like, 'Oh, I can't imagine it.' And so I actually mapped out one that maybe I could take you guys through. Uh, I wrote a little script here, okay, called 'The Old Way.' Um, so, you know, fades in, starts in the TBPN Ultra Dome. Me, you, Tyler, we're all doing the show. All of a sudden, all the phones in the Ultra Dome light up. Emergency alert. Unknown biological event. Shelter immediately. Outside, alarms begin to sound across Los Angeles. A strange haze starts pouring through vents and drifting across the streets. Outside, I see a cloud approaching the Ultra Dome entrance. First thing I do, bioweapon attack. This is what you got to do. Boom. Right there. It's not going in your mouth. You're good. Second step, pull out your gun. Boom, boom, boom, boom. Start blasting whoever's responsible. You're just firing wildly. Once you got the situation a little stabilized, then we got to figure out what's going on. So I go over to Tyler and I say, 'Look, we got to hack into the superintelligence. We got to figure out what is going on with this bioweapon.' And he's like, 'Sure.' Like, 'Yeah, let's hack into it.' Like, 'What? I'll open up Codex. What do you want me to prompt?' I'm like, 'No, the AI has gone rogue. We have to do this the old way.' He's like, 'Okay, I'll open Claude Code. Like, what do you want me to prompt?' I'm like, 'No, the old way.' He's like, 'Okay, I got Cursor open. Like, what are we doing here?' I'm like, 'No, Tyler. We have to go back to the old way.' He's like, 'Okay, I got GitHub Copilot fired up.' I'm like, 'No, Tyler, you can't use any AI at all.' He's like, 'Okay, just let me know what code you want to write and I'll look it up on Stack Overflow.' And I'm like, 'No, the internet has been contaminated, Tyler. We can't use anything. We have to code the old way. We have to do it by hand.' And Tyler's like, 'I can't... I don't know.'
I don't remember anything. I can't write any code. So what do we do next?
We got to get George Hotz. He's the only one who can help in this scenario. So we go...
Because he's the only one that remembers how... he remembers how to code. He's got local models stored. He's got his own hardware. So we go get George Hotz. And then we're like, 'Okay, George, hack into the superintelligence. Tell us what's going on with this bioweapon. And why aren't we affected? Like we've been walking around. We seem to be fine. Everyone else is turning into a zombie. What's going on?' So he locks in. He writes some code. He hacks in, figures it out. He says, 'Jordi, tell me about your diet.' And you're like, 'Well, I just eat at Arowan exclusively.' And he's like, 'Yeah, my analysis shows that your body has never had a toxin in it, and so you are immune to the bioweapon. It doesn't affect you at all.' I'm like, 'But what? How do you explain me? I only eat Arowan once a day.' And he's like, 'Well, John, my analysis indicates that you might have had the original Four Loko. Did you drink the original Four Loko in large quantities?' And I said, 'Yeah, I was daily driving it in college. Why?' Like your body consumes so much original Four Loko, it's now essentially pure toxins. So the bioweapon sees it as a hostile environment. It doesn't even try to infect you. So you are immune. So we're both immune. Uh, but the bioweapon is still wreaking havoc on America. So we got to get to the bottom of it. So we're going to need some muscle. So we go get Sam Sulk.
So we go get Sam Sulle and we're like, 'Let's go. We got to go to the data center. We got to break in, shut this thing down. That's the only option.' So we get in the Cadillac Blackwing. We jump it off a ramp, slam into the data center, pull out the gun, start blasting the GPUs. We're just firing everywhere. Finally, after we've shot 90% of the data center, it's just bullets everywhere. Finally, a humanoid robot emerges. It's Mecha Hitler. We got to fight him. We got to fight him. We're shooting. We're shooting. The bullets are just bouncing right off. They're just bouncing right off. Uh, we're getting beat up. We're trying to fist fight it. Sam Sulle's doing nothing. It's not working. Uh, and then at the last second, John Cena breaks through the roof, slams down the humanoid robot, rips his head off, and humanity wins. So that's a kind of like a concrete example of like how AI doom, how, you know, a runaway rogue... just one scenario.
That's one possible scenario. I put it at 10%. That's the way it plays out.
Yeah, that's kind of where I'm at right now. Yeah, that's how I'm thinking about it. Yeah. 10% if we don't do anything. Yeah. Around safety.
Well, that's what we'd have to do. We'd have to like this if the bioweapon comes.
Cover up and eat a lot of Arowan and drink a lot of original Four Loko.
Round of applause. I didn't know that was coming and I didn't know where it was going.
I'm just trying to think more concretely about how these things will play out. And be prepared. You got to be prepared for what happens. You also have to get on console.com. Console builds AI agents that automates 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets. Um, let's see what's next. Uh, Meta, the AI agent company Muse, they're putting humans in there. Meta is testing human concierge for its new personal agent, Muse. Uh, Meta has been testing human concierge for its new personal AI assistant. This is on Bloomberg, which entails having human contractors quietly handle some of the phone calls placed by the digital agent. You got to use this to order Amazon. Like Amazon's like, 'No, no, no. We're going to block your browser.' Well, did you block everyone's browser? How's that going to play out? I was thinking about like at some point if I write a really, really bespoke... like basically a CLI, like a really bespoke weird CLI, like I vibe code it but I use like sort of odd techniques, like I'm like, 'Okay, use computer use and open up Opera or, you know, Mozilla and click around and add some randomness and do this.' And it's like, and I basically create a CLI that wraps an agent where it runs on my home Wi-Fi, so Amazon... and I'm like, and I just do this. I don't productize it, so Amazon can't really detect it. Do you think I could potentially stay in front of their battle forever and then puppeteer it from other agents?
Well, I just asked ChatGPT, 'Can you order Amazon over the phone?' Because I would think that at that scale there would be at least one way to do it.
It seems like you can only use their customer service number to help with existing orders. Mhm.
So one thing is you could order a bunch of the wrong thing. Like a hundred of the wrong thing, 100 like, you know, bug treatment.
Order everything and then have your agent refund everything but the thing you want.
No. Have your Muse agent...
Have a human via Muse call Amazon.
And say, 'Hey, I actually got the wrong thing here. Can you swap it with this other thing?' Interesting. And just do that as many times as necessary.
To get the right order. Maybe have one order that's just always open where you're just adding and removing.
They have to be completed orders because you have to have made a mistake to call the customer.
Oh, yeah. You can charge me, but just add it to the order. I don't know. We're going to get... it's going to be a weird battle. Ben Thompson was writing about it today. It was pretty good. Amazon... I mean, his whole framing is like Amazon is like the strongest in the AI era of the hyperscalers because they own logistics, infrastructure, the final step, the real world. And so he anticipates that they will hold their ground for a very long time and not give in. Where... and that's basically what they did with OpenAI, where OpenAI had the instant checkout process. They invited everyone to be a part of it.
Amazon was like, 'Thanks, no thanks.'
But we'll give you tens of billions of dollars. And but interestingly, yeah, so now the result is that Amazon is vending in their ads into ChatGPT to close the purchase there. Uh, Tobi Lütke at Shopify partnered with Muse and announced that they are integrating, but the interesting thing is like where is the value capture layer and it only works with Shop Pay? And so it's a lever to get merchants to go onto Shop Pay, which Shopify will still make money off of. But for merchants, a lot of people are worried about the implementation that happened with Walmart, where ChatGPT integrated with Walmart and the end result was that conversion rate was one-third of what it was in the app and on the core website and that carts were smaller because people were just going and saying like, 'I just need a roll of paper towels. Just send me that.' Whereas when people are like, 'Oh, I need to do my shopping. I need soap and paper towels and whatever else, I'll build... paper towel holder that's made by a company that's over 50 years old.' Yes. Um, so uh... no. So the crazy thing is, so now, uh, Venode was on the show yesterday talking about, uh, Waymo, which is an FO... they have a new agent that I believe is human-supported.
Uh, in that some of the stuff is done by AI and some of the stuff, if necessary, can be escalated to a human. Now you have Meta at least testing, exploring, having a human in the loop on some of these things. Yeah. And my big question is like, why, why, why? Like voice models are getting pretty good. It feels completely unsustainable for Meta to roll out a product to billions of people where users might just be like... like if I had something I could text and say, 'Hey, call this person. Call that person.' Yeah. I'd probably be using it.
I think there's a ton of value. First, the Muse install base is small right now, so it's not that crazy. They have the ability to stand up huge offices whether it's through Scale AI and that team that came over or through their, you know, all the things that they've done over the past two decades. Um, the other thing is that Muse specifically will not train on your data that you put into Muse. And so this wasn't like a huge announcement or anything, but uh, you could see people being like, 'Ah, I don't know if I want to use this AI agent that's going to be using my personal information.' Even if they're running like the PII redactor, uh, putting all of that might be a little bit off-putting. So Meta is not getting that data. But if they have a human in the loop, the human who does the task that's just beyond what the model's capable of, they're also generating training data because they can probably train on that. And so it's like the model by default can, um, I don't know, like summarize your calendar, right? And give you a briefing on what's going on in your calendar. It knows how to integrate and the model's capable of that. Muse 1.3 is capable of that. All the models are capable of that. But what can't they do? They might not be fully ready to go and have a complicated conversation with someone if they're ordering flowers over the phone or something. And so the current models might fall down. So you need more data to solve that. And so you put the humans in the loop and then that's your extra training data for the next leg up. And then you just keep repeating that. So I would view this human-in-the-loop thing more as them doing data collection and creating more training data for them than a permanent solution to the product problems. Also, I just don't think the time on site for these apps is actually that high. I think there is like an excitement when you jump on, you do a bunch of stuff, but then with all these AI tools, like there's a reason why OpenAI comps to like weekly active users because there's plenty of days where people are like, 'I went to the beach and I didn't touch AI at all. Like I didn't do any work.' Uh, there's some feedback on Alex Wang's post talking about like it'll help you with your goals. And I think Katie Nottopolis was saying like a lot of people just don't have goals. Like they just want to chill. And this is also the Ben Thompson thing of like consumers want to be entertained. Like people that go to the beach, they'll still scroll Instagram. Are they really going to be doing anything with any agent no matter how powerful it is? Like there are plenty of people that have staffs of EAs and personal assistants and they don't do anything on the weekend because they want to chill and they don't want to think about agents. One thing's for certain. Uh, it's over or we're back.
Professor Jang is vindicated. Yesterday we watched...
We watched a video where he said, 'I guarantee that there are humans...
Behind the chat apps that you use, manipulating the information. And he is correct, at least in the short term, for today. Trini is thinking about the knock-on effects of all these personal agents. He says there's no way any of us can predict the impact AI will have on companies, sectors of consumer behavior, which has been for all of history until now, human behavior. But to claim that it won't have an impact and simply brush it off lacks both humility and creativity. Tactics that worked when consumer behavior was dictated by human psychology will fall by the wayside, and those that failed to adapt will be left behind. Every day from now on, we will see new examples of areas where friction was profitable becoming cost centers for companies that don't react in a timely enough manner. Today, you might realize just how many flight credits, loyalty points, or promotional offers you've never used that will end up utilized by agents. What impact will that have? Tomorrow, you might see something that makes you ask how much money health insurers make simply because people won't sit on the phone for five hours trying to get coverage approved unless it's a massive expense or life-or-death matter. What about your savings account? The money could be in a money market fund. You haven't bothered to move it. How much in net interest income is made because of that friction? Yeah, people have money locked away in Venmo, like a couple bucks here. And if it's frictionless and actually free to move stuff around, you get a lot different. What do you tell?
This was like the David Deutsch anecdote about how he fixed his dishwasher. He was about to buy a new one and then it just told him that he could.
Yeah, it's like ghost GDP. We'll need a new metric.
So it doesn't show up anywhere, but there is a new metric.
No. But I mean, so I don't see very many of these adjusted GDP. That's how we get to 10% GDP plus AI. E equals MC squared plus AI. Yeah, this was all kicked off because a flight was delayed seven hours. Someone asked Muse to file a compensation. Five minutes later they had a $250 credit in their Delta account. That's something that you'd normally have to hop on the phone for. It was a lot of cumbersomeness. So people on that team were clearly modeling like, well, 1% of people will ask for it and 50% of people will be able to be like, nah, we can't really do it, and so we think this will only be like a dollar per flight. Well, now it might be $10 per flight, but of course the end result is that they will put AI on the other side and the two AI agents will just negotiate and find out a resolution, I imagine. Anyway, oh, Catrini has a little prediction here. Look at the follow-up to this post. They put a, what is that? An RSA hash or something. That's cool. So they made some prediction there that they will see. Anyway, that's one of the greatest hacks when you want to take a victory lap in the future, but you're not willing to actually post both. You post like, I think the stock market will go up by 100% in the next 10 years, and then I think the stock market will go down in the next 10 years. And then you just delete the one that wasn't accurate and then you're good. Nikesh Arora chimed in on the knife fight between Amazon and Muse. There will be a bigger battle. This will be a bigger battle than anyone anticipates. It's only a matter of time before there is an Apple and Google version of Muse and possibly TikTok in addition to the frontier LLM agents. Maybe a commerce agent from Amazon. That's sort of my prediction. I think Instinct might land with Amazon potentially, although they do have Rufus. But you can...
And Tyler is a Rufus power user.
Yep. Rufus is good. You don't want to talk down on Rufus, but Instinct is clearly tapped into something really special with the energy and the support that they have and the community they've built, but there are a lot of little rough edges that need to get sanded off. And that's the domain of Amazon in my opinion. Like Amazon is the, you know, the trains run on time and it's a pretty efficient company. So every app that's a service marketplace or commerce app will need to existentially decide to open APIs for consumer agents to interact. Smaller players have no choice. Ad revenues are more than transaction fees. Either the consumer benefits or distribution aggregators will demand a higher fee transaction. There's an interesting stat from the strategy post about Amazon. The net income for Amazon e-commerce was like $36 billion and the ad revenue was exactly twice that. So the business is unprofitable without advertising. They put advertising everywhere. So it's a very interesting dynamic where you can see the motivation. They have $68 billion. You flagged that number. It's a huge...
Over 70 in the last 12 months.
Over 70 billion. So a significant pool of revenue and profit that they will be protecting for sure. I don't want an agent for each app, which does sound very annoying. Super agent, go and use all the other agents. Maybe that's the future. I don't know. Yeah. This goes back to like roll your own agent that then puppeteers your Muse, which has good integration to Facebook, and then your Amazon Rufus. Yeah, I wonder if you could order via Rufus because you could probably reverse that API, but you'd still get banned. You're just chatting with... It's just Muse chatting with Rufus to order you paper towels. No, I said over 50 years old. Basically, you had an idea yesterday, a potential monetization channel or strategy for agents, which is like sponsored agentic flows, which I thought was really good. Like imagine you and your friends, you all have agents and then...
The example was Spider-Man.
Yeah. It's basically like it drops in and it says like, 'Hey, I got five tickets to Spider-Man.' It says, 'I know from your Instagram history that you're a Tom Holland fan or you like the Marvel Cinematic Universe. Spider-Man: Brand New Day is coming out in two weeks. I can get you tickets Friday night. There's nothing on your calendar. I can get a Saturday matinee. I know you've been sending reels to these five friends. Do you want to plan a trip to the movies?' And I will book it and message everyone, set up a calendar invite, text everyone and say thumbs up if you want in, thumbs down if you're busy, and I will put together a group outing. And that does feel like something where Spider-Man, the Brand New Day marketing team would put dollars behind advertising that to get people.
The question is how much, how will consumers tolerate that at all? Remember how pissed off people were about the BMW Spider-Man ad blow?
I think they will if it's good. And I think the long-tail small business advertising will work, especially if it's like... it's a little bit tricky to figure out how, like when you're in these agentic workflows, there's often multiple beneficiaries. So Spider-Man benefits, but also Fandango might benefit, but also like a particular theater might benefit if you go to, you know, Lowe's versus AMC, or I don't know, they might be the same company now. But so it's interesting to see like where the sponsorship would break down. But you could totally imagine like you have a wedding coming up, you've been seeing ads for this tie company, you've also been seeing ads for this sock company. Do you want me to put together like a whole new outfit for the welcome dinner that might look really good? And those companies pay... unclear like where all that goes. But the scale of the platform I think is what unlocks it because you know Sean Frank at Ridge Wallet is going to be like, I'm going to test that, and there's going to be millions of other small businesses that are going to test something or other. Anyway, let's see. Products like Muse and Instinct are going to saturate customer service teams. Again, we talked about that. Not a huge problem because you flip it around and the customer service team is using AI. So I think the equilibrium is just a lot more AI on both sides. Hopefully smoother transactions. Signal is talking about Walmart here. If only Walmart was actually competent enough to realize the agentic commerce opportunity ahead of them. And it's interesting with the backdrop of the Vibe acquisition, Walmart seems to be going after the same pool of revenue that Amazon went after with the ad market scale-up and the ad business, when in fact if they didn't have that, they would be sort of counter-positioned to lean a little bit more into agent commerce because they'd say, 'Look, we don't have an ads business to lose, so let's be the default option for agents that buy anything.' There's also an interesting knock-on effect which is like with Shopify going so deep into Muse and agentic commerce and agent checkout. I wonder if we will wind up seeing a different type of Shopify merchant emerge because Shopify merchants, it's been a lot of clothing. It's been a lot of like DTC single product, like a single shoe company. But I wonder if you will see someone who's like, yeah, I sell a lot of stuff on Amazon. I have a factory that makes paper towel holders, but I also got paper towels. I also got socks, T-shirts. I got sort of basics. And I wonder if someone will go and try and win on Shopify and set up the actual economic structure to be able to provide the experience because consumers will clearly want to be able to go to agents and say, 'Order me a new pair of socks.' And just have it done, right? And so if they say, 'Sorry, can't do it on Amazon,' they're going to figure out another place to do it. The water will flow into the cracks. Yeah. And so...
No, it's a big, big opportunity for Shopify to be able to lean in here.
But the issue is still the real-world physical fulfillment, having a bunch of different products in...
But the individual Shopify merchants solve that on a diffuse basis. So in theory, if it's like, okay, I want next-day shipping. There's five Amazon merchants that sell... or there's five Shopify merchants that sell socks. You want socks tomorrow. There's only one of them that has a distribution center in your town and can actually get them to you tomorrow. We'll just go with that one. So you could maybe piecemeal together an Amazon-like experience, but that's the bet. That's the knife fight. Like that's what we're going to figure out. Either Amazon wins and their moat remains a moat, or the war for the open web continues to battle it out. That's what Alex Wang is hoping for. He says, 'Once more unto the breach, dear friends, the new open web, Meta, Stripe, and Shopify.' The hilarious thing here, of course, is like Meta is solely responsible for sort of killing the open web in the former era. Like it's why all the news sites are not, you know, initially... why everything went closed and like you can't just Google search for stuff on Instagram. Everything's closed off. It's not truly open. But it's like the Meta flags in the front because we'll be the aggregate...
Open web for thee, but not for me.
Everything below the fold will be the open web, but at the same time it is just fun to watch the battleships go off to war and see how it pencils out. Because there is a scenario where the equilibrium solution is actually a more open web. We will see. We will see. What is going on with McLaren? We got to talk about McLaren before we get any further because they rebranded and there's a new logo. They went from an angled M that looked very fast, very bold logo, and now it looks like Miss Len.
Yeah. It's interesting. They seem to be trying to reposition as a luxury brand versus a motorsports brand. Like when I think of McLaren, I think of like fast, sporty cars, period. I think of racing, F1. I love the old logo.
Lineage too. I love the old... They've never made like a mass-market car really. I love the old mark. It's one of those things, you hate to be the guy that when a brand, when a company rebrands, you're like, ah, it's brutal, because oftentimes you revisit it in six months and you're like, okay, it wasn't that bad.
But this feels like bad. And but the business has not been performing right. The business has been in a rough spot. They don't have the... they don't have like the Cayenne that they can sell. They're doing an SUV. That's coming. But I'm saying like an Urus competitor. To date, they don't have like a mid-market car.
I'm sure their Urus competitor will do well, but still...
Part of what enables the Urus is like the Q5 and the Q3, right? The core selling economy. Yeah, can't forget about that. It's still growing. So, so anyways, pretty brutal rebrand. But again, it's a tough one just because it was always like I always just thought the brand... their brand was off.
The steelman here is that they are taking a number of bets that seem to speak to interesting, potentially valuable pieces of the market. So you have the SUV. The Urus is incredibly well-selling. So that could just be a cash cow business for them. Then you have the MCL, their manual supercar, and that could speak more to the enthusiast, right? Because enthusiasts want manuals. Porsche's been delivering manuals. Ferrari is starting to put them back in. And other brands are starting to get that memo that it's not enough to compete on speed. EVs have you beat there. They haven't gone full EV. I haven't seen a lot of McLaren signaling that they're going to remove the combustion engines. So that seems like they are positioned for the future. That's certainly a trend is don't go full EV, don't kill all the gas cars. And then they also have the W1. So they're staying in the hypercar world and the W1 lineage. I think it clarifies what McLaren's been doing because when they were doing the Senna and the Speedtail, there were just too many special cars. And the problem was that you were like, okay, Senna, Speedtail, where do we sit? But F1, P1, W1, very clear lineage. You know, it's the best. You know, if it's that, it's going to be the craziest thing they have. And so if they can create some differentiation between the true hypercars that they create, the supercars that now come in manual, the SUV for the more everyday daily driver, you have the trappings of at least satisfying the demand that exists in the market. You do have to go create more demand. You got to win in F1. But...
And the new mark's not coming out of nowhere. On the website they say, before McLaren became a mark recognized around the world, it was a family name painted above a service station in Ramura, New Zealand. Those original letter forms inspired our new wordmark reimagined for a new era. So they're going back in time. McLaren was started by a guy from New Zealand, but it was started in the UK, but the founder had been born in New Zealand. So...
Well, on a related note, there's some data that might be bullish for the supercar, hypercar market here. Toy sales among adult-only households have outpaced households with children for toy sales, growing 16% through June. So when I think of adults buying toys, I think about Myers-Mans, I think about dune buggies. I think about dirt bikes. I think about supercars, boats, yachts, wakeboarding. These are the toys that adults buy. And it's not surprising to me that there is a trend among adults buying toys. I mean, these are just fun toys. And so it's the fact that it grew 16% in June. A sign that, you know, adults are having fun. They're buying toys. And bullish for the supercar community, right?
Don't you think? Judging by the picture here, you might be incorrect.
Wait, they're talking about like children's toys. Interesting. Yeah, I think that's a little bit more nuanced. I would go for, you know, gaming rigs, racing simulators. These are the toys that adults should be buying. I think that's...
I did... I was a teenager when I realized that the Lego Ferrari that I had built years, you know, probably when I was like 10, had a market value. Like I was shocked.
Yeah. Because I had some Lego, like really complicated Lego Ferrari that I probably built when I was around 10. And then as a teenager, I'm, you know, cash-strapped teenager, I was like, wait, someone will give me $100 for this. That's crazy. Sold it. Of course, I'm not super sentimental about that kind of thing. But I think that trend has just been building and now there's people that just buy every new Lego that's released and just sit on it and it's...
I saw a... I was looking with my son at an old Lego set that I played with as a kid and I was like, I can actually go and buy the exact same Lego set and share that with him because I thought that one was really cool, but it has gone way up in value. I was like, I should have just held on to it because it's like a hundred bucks now.
So I built a Ferrari Lego as well when I was younger. Probably not as complicated, but I remember being like, oh, this is like a sick Ferrari. Like maybe one day I'll get a real one. And then I finally got older, learned more about it. I was like, it's an F40. So I was like, I'm probably never getting one of these.
But maybe, maybe. I don't know what F40s are doing. Thompson says, 'Today's adults increasingly behave like the children they never have.' Feels like a rude thing to say about people. And yet toy sales among adult-only households have now outpaced households with children for toy sales. That is so crazy. And black-pilling and sad and it's over. It is so over. I don't, you know...
You're literally me. You're doing the literally me thing.
I think one factor here is like children, in my experience, don't care about the cost of a toy. And so a $5 toy will bring just as much joy to a child as the $100 toy. And so the adults are like, I have to get the $500. So I think it's like for every toy that a parent buys their child, it could be like a dollar, $5, $10. So I went toy shopping with my son and he finds this like Hot Wheels thing and it's like a book that has like some Hot Wheels inside and stuff and it has like the craziest, most electric...
Look at this. Look at this comment from Winom. 'I never had a quote-unquote Lego set. I just had a sorted Lego brick. I used to fill the bath with water and make oil rigs which used to annoy my mom.'
This guy was in the oil rig.
Hit the oil rig. Fantastic. No, on that point, I went to the store. My son finds this Hot Wheels thing that's like kind of structured like a book, and it has some Hot Wheels cars in there, but on the cover of this box is like the craziest, most elaborate loop-de-loop thing. And so he gets super excited about it. And I'm like, okay, I got to check this out. And it turns out that what you're actually buying is like a book with some stories and then just like a big sheet of paper, like a poster that has that illustrated on it. And you can imagine the loop-de-loop and the actual track. And I was like, dude, let me clue you in. This isn't the real thing. Like we can go get the real thing with the actual loop-de-loops and stuff. Like let's find the real toy. And he's like, no, this is the one I want. This is what the one I want. And I'm like, it's the power. Okay, fine. But like, I'm like, this is going to end poorly because he's going to be dissatisfied when he actually opens it up. Opens it up. He's like, this is the greatest thing ever. He's like, my imagination is running wild. I would never want the alternative. And so your point is 100% accurate.
Toys. The best toy company in the world, period.
Is this company Play Forever. Okay. Have you seen any of these things? No. They're the most well-built kids' toys in the world. So well-built.
Yeah, we should. Let's get the founder on.
They are Play Forever. They are absolutely incredible.
Okay. Okay. Yeah, let's bring in our guest. As much as I want to keep talking about toys as an adult, I would rather have a surprise guest. Shopify is the commerce platform that grows with your business, lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with AI agents. We have a surprise guest.
Our favorite guest, Ben Thompson.
Oh, wow. I didn't know that. Maybe.
We actually were able to fully...
We were able to fully surprise...
It's a very, very short list.
We were able to fully surprise John. It was me coordinating with Ben. So this is like a surprise.
This is a wonderful surprise.
Well, I mean, you were going on and on about toys and I was sitting here trying to decide who has the most ridiculous jacket on today. I think it's down to a tie. Hey.
I do it on the ramp side. I got to wear the ramp suit. This is my next suit. I've been wearing this as we're trying to get to the bottom of this whole AI safety debate. I'm trying to embody the movie...
In the mind. In the mind.
How cold is it there? I don't know. I thought you guys are in LA.
Yeah, we're in LA. It's warm. Well, we don't have any windows.
Great insight into the booking process. Nick's been doing great work trying to get me on. I kind of was committed. Oh, you could do it then, do it then. He's like, oh, about Friday. And I'm like, I don't know what's going on on Friday. So I'm like, can I just go on now and then I will stop feeling guilty about it? And somehow I'm here I am. So those...
