Subscribe
Share
Share
Embed
TBPN is a live tech talk show hosted by John Coogan and Jordi Hays, streaming weekdays from 11–2 PT on X and YouTube, with full episodes posted to Spotify immediately after airing.
Described by The New York Times as “Silicon Valley’s newest obsession,” TBPN has interviewed Mark Zuckerberg, Sam Altman, Mark Cuban, and Satya Nadella. Diet TBPN delivers the best moments from each episode in under 30 minutes.
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
Speaker 2:Financial markets.
Speaker 1:Financial markets and the financialization
Speaker 2:No. But also safety as well.
Speaker 1:Oh, you care about both.
Speaker 2:I care about both.
Speaker 1:So you wanna accelerate the AI build out safely.
Speaker 2:Safely.
Speaker 1:Yes. I I thought it was I thought it was you're gonna blend the two, and you're like, look. I want insane private credit deals, trillion dollars of of capital to flow into the AI build out, but I wanted to do do it safely.
Speaker 2:Every contract is safety. Investment grade. I wanna prioritize safety on the whole frontier.
Speaker 1:Okay.
Speaker 2:Safety around the models.
Speaker 1:The capital markets,
Speaker 2:Steve? And financial
Speaker 1:In the capital markets. Got it. Okay. I'm I'm seeing I'm seeing where you get there. Anyway, you wanna do Harvey first?
Speaker 2:Let's
Speaker 1:Okay. People are people are taking shots at Harvey. It starts Samuel Cole.
Speaker 2:For
Speaker 1:it. I won't stand for it. No. I like Harvey. We we we we've talked to some lawyers that use Harvey.
Speaker 1:Seems like they're proud. We also had Logora on the on the on the show just last week. Had a good time with them. It seems like a very competitive space. But the the headline that grabbed attention, they say track changes.
Speaker 1:Harvey's gross margin fell from about 50% to negative 50% by June as agent token use spiked twentyfold 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.
Speaker 2:As any company.
Speaker 1:As really as any company, actually. You don't wanna be selling a dollar for 50¢. And that's exactly what they did in June, basically. That's what they were doing. They were selling tokens at negative gross margins.
Speaker 1:Not good, but there's way more details in the actual reporting.
Speaker 3:So you should go read the Bloomberg report.
Speaker 1:So the first thing is, like, cost did explode, but it was on the back of more product usage, which is good. They're growing We'll get into that. And token consumption increased 20 x this year. So, basically, they are feeling the reasoning era, the agentic era, one job, one one process in Harvey is generating a lot more token.
Speaker 2:Had seat based pricing.
Speaker 1:Yep.
Speaker 2:And then AI got a lot better. So every seat was just using it way more. Yeah. Very simple.
Speaker 1:But it's not it's not just using it more. It's that your 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.
Speaker 1:It's actually gonna 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 gonna be worked through this agent, and so you're generating way more time.
Speaker 2:Yeah.
Speaker 1: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. Kinda 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, but they are back to positive gross margins. They changed their model usage.
Speaker 1: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 Tennant. And there's some interesting details there. I wanna debate with you on, like, the value of that where that goes.
Speaker 1:But first
Speaker 2:one of the founders of Harvey responded.
Speaker 1:Oh, he did. I was He said
Speaker 2: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.
Speaker 2: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. Bunch of great points in here even outside of the the the Bloomberg piece and the subsequent coverage.
Speaker 2:Max from Ligora was going pretty hard saying it doesn't make sense.
Speaker 3:Oh.
Speaker 2:It doesn't make sense to to train your own models because the Oh. Public basically like publicly available frontier models are are dropping in cost Sure. And the and the quality is increasing. So you're just sort of, like, wasting
Speaker 1:Yeah.
Speaker 2:Your own And and again, like chance. That's a bet that he's making. Harvey's gonna make another bet. Probably some mixture of both is gonna be the right approach.
Speaker 1:There's another interesting sort of like how how bad is this situation just comp, which is so they had negative 50% gross margins in June. They're at a $400,000,000 run rate for ARR, I suppose. And so that's $33,000,000 a month. And so negative 50 percent margins means they lost $16,000,000 but they've raised like 500,000,000 So in terms of, like, burn, it's not super cause for crisis, in my opinion. And and what's interesting is that even through June maybe they did the Sequoia deal that valued the company at 15.5 before that happened, because it was announced September 9.
Speaker 1: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?
Speaker 2:I would assume so. But but the main thing is, like, every every application layer company has, like, gone through a moment like this even if even if it was for, a single day. Totally. Totally. So it matters how quickly you can respond Yeah.
Speaker 2:Yeah. Yeah. And adjust.
Speaker 1:Yeah. So where do they where where do they need to get to? Like, what is the standard for the legal technology industry? There are some comps here. EDiscovery provider, DISCO, public company, 75% GAAP gross margins.
Speaker 1:Like, it is a software tool that when you're going into discovery, you need to connect with bunch of data. Okay. We got a whole bunch of text messages, 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.
Speaker 1:Thomson Reuters has a legal professionals segment of software that's running at almost 50% adjusted EBITDA margins. So like true true net profit. So that's that's sort of where they wanna be. Also, pure play law firms have really high margins as well. Hard to comp them to businesses because they pay out the to the partnership.
Speaker 1:But PricewaterhouseCoopers reports net profit margins around 41% for the top 10 firms. It's lower for firms 10 to nine 10 to a 100 or something like that. But still, lots of health in the legal industry broadly. Obviously, everyone knows lawyers make money. That shouldn't be surprising.
Speaker 1:And, yeah, tokens are getting cheaper by the day. There's a bunch of model launches. Opus five five launched today. Grok 4.7 was yesterday. And Grok 4.7, no one was even Elon was like, Okay, we're not jumping straight to the frontier with this.
Speaker 1: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 you could benchmark against it. But the interesting thing about that particular benchmark was that it was a legal task benchmark based on cost.
Speaker 1: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 x AI and Grok. Do you have something? Are
Speaker 2:you You wanted to give it up for BigLaw? Yeah. The chat wants to give it up for BigLaw.
Speaker 1:Let's give it up for
Speaker 2:They're the real winners here. Additionally, GPT Six Soul and Luna are starting to roll out. Fun. So we'll look out for announcements there Cool
Speaker 1:demos. I saw one Opus five five. Mean
Speaker 2:cat 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.
Speaker 1:Flawless. Yeah.
Speaker 2:Thank you. Thank you for flagging.
Speaker 1:Huge upgrade. Very funny. Yeah. I saw a cool I saw a cool demo where someone sketched out on a piece of paper sort of a trebuchet in Opus 5.5, took that image into simulation and created like a virtual version of it. AI is is is continues to impress in terms of translation of things.
Speaker 1:Oh, your your kid drew a, you know, 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 wanna if you wanna read a a 2,000 page novel based on Call of Duty Modern Warfare four, like you can probably do that now. So you can have your content in any specific specific format you want. That seems to be where the AI continues to impress. Anyway, let me tell you about Codex.
Speaker 1: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.
Speaker 2:Should we pull up What? Your latest masterpiece? Yeah.
Speaker 1:The Literally Me compilation. It burns up the Internet and vibrails 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 there. Patrick Bateman in American Psycho.
Speaker 1:You can put yourself in the movie. There's inception. We got Jordy in there. That's drive. I think that's heat Heat.
Speaker 1:Mad men. Can we call him out? I think that's bike club.
Speaker 3:You got
Speaker 1:joker. You got taxi driver. You got night crawler. No country for old men. Social network.
Speaker 1:Godfather, Back to Joker. I forget that one, Sawshank, maybe. That one's Collateral, then Blade Runner Blade Runner again, I think. Yeah. Some of these some of the renders are really, really good.
Speaker 1: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 at all. It's just Product is even like
Speaker 2:it didn't work. I
Speaker 1:was like, do it anyway. I don't care. It's fine. It's gonna be on screen for two seconds. That one's good.
Speaker 2:John is really my human agent. Just You didn't even tell me 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.
Speaker 2:Right? You'll send me
Speaker 1:You wanna post something like this.
Speaker 2:You with Everyone should have a everyone should have a John Everyone in their life.
Speaker 1:Everyone should. Everyone should. But, yeah, fun fun powerful tool. Still got $500 locked away in Higgs Field in the wrong section. Can't do anything with it.
Speaker 1:Maybe just give away the account or something. Anyway, what matters more, people or insects?
Speaker 2:Tough call, John. Mhmm. We should we should get into maybe one side of it first. Yeah. Steel man it?
Speaker 2:Before we form an opinion.
Speaker 1:Okay. The human side. You wanna steel man the human side?
Speaker 2:Look. And I'll just I'll just say that we're coming to you live. Yeah. We're coming to you. We had a rough morning with the
Speaker 1:We've been
Speaker 2:so mosquitoes. Busy. Actually had to we had to we had to DoorDash one of those bug bite Yeah. Treatment thing, the heat, whatever. What I don't what do you call them, Tyler?
Speaker 3:I think it's like insect bite healer. Okay.
Speaker 2:That's what
Speaker 3:it's listed as.
Speaker 2:It did work because I Yes. Sebastian says I'm famously famously pro mosquito.
Speaker 1:Yes.
Speaker 2:But I 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. I I It's funny because I can't stand EAs and I think they're weird freaks and I don't think they should be anywhere near public policy or anything But like 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 anyone should I don't think we should harm the insects.
Speaker 2:Oh, okay. I think humans are more important but that doesn't mean we we you gotta like take them all out. You know, just go somewhere else.
Speaker 1:Right? You would take sort of you would be altruistic towards the towards the insects.
Speaker 2:Ellie says mosquitoes are terrorists of the air. They really are.
Speaker 1:You would you you would you wanna be altruistic towards the mosquitoes. You wanna be effective in your altruism when it comes to mosquitoes. You don't wanna
Speaker 2:No. Want them I wanna them effectively to be nowhere near me.
Speaker 1: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.
Speaker 1: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?
Speaker 1: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.
Speaker 1:But he estimates that for every for every second of human life, insects collectively spend roughly two hundred and seventy thousand seconds dying, about seventy five 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, guy pulled some cool numbers together. This is interesting.
Speaker 1:This is fun fact territory. Large part of the essay turns on whether insects can insects can pain. Bentham's bulldog argues that they can, even weakly. 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.
Speaker 1:Using what he describes as a conservative assumption that insects experience pain at just one ten thousandth 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 is there a piece on this where this is the view of the of this community with the relative like, they're looking forward to how a world might interact with, like, a a godlike machine super intelligence 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 I
Speaker 2:think that's meticulous. I think that's certainly like where the thought experiment goes.
Speaker 1:Yeah. Right? And so and so the the 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.
Speaker 1:Well, now there are a quadrillion super intelligences that can all feel pain in some abstract way, so they're more valuable and saying, 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 in the robotic future. Yeah. I don't know.
Speaker 1:The reactions to this were rough. People did not like this argument at all.
Speaker 2:Well, it's it's just the timing is funny because there's so much tension building around Yeah. The sort of EA rationalist movements with
Speaker 1:Which is super people don't realize how fractured
Speaker 2:Yeah. It but it's sort of boiling up to the sort of national Yeah. 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.
Speaker 2:In the aggregate.
Speaker 1: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. Andre
Speaker 2:Yeah. I think it's actually important for people to read the before you go and comment on this article, watch A Bug's Life.
Speaker 1:Oh, yeah.
Speaker 2:At half speed.
Speaker 1:Half speed.
Speaker 2:Watch the whole movie at half speed. Yeah. You're a movie guy but
Speaker 1:I think we coexist.
Speaker 2:Maybe you're not a fan of watching movies at a half speed. It's not really how the I'm
Speaker 1:really a half speed guy.
Speaker 2:But half speed is a great sort of tool to have in the tool chest
Speaker 1:I
Speaker 2:when you wanna comprehend.
Speaker 1:Two x speed, three d glasses with the 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
Speaker 2:Duo.
Speaker 1:While I'm sleeping.
Speaker 2:Yeah. With the duo.
Speaker 1:Exactly. Max brain rot. I mean, the fracturing of EA really like leads to a really, really wide range 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.
Speaker 1:It must be full employment if you're in the public relations department over there.
Speaker 3:Yeah. I was gonna say I I think the shrimp shrimp welfare stuff Oh, yeah. Makes more sense to this because, like, I'm not really sure what like, after reading this and say, I agree, like, what do I what are the next steps? Like, we're not, like, really factory farming insects.
Speaker 2:Maybe through pesticides. Speak for yourself, brother.
Speaker 1:Yeah. Yeah. The shrimp the shrimp welfare
Speaker 2:Got some big announcements coming soon.
Speaker 3:Like, what percent of of these, like, 600,000,000,000 insect that insect deaths per second are Preventable. Preventable. Yeah. Where, like, with the shrimp, you know, you can say, like, okay
Speaker 1:Change the farming practice or just don't eat the shrimp.
Speaker 3:Yeah. So I think that's why maybe this was less impactful. I don't know.
Speaker 1:Yeah. Mean, it's all continuum to like until you get to like the veganism, the animals who you can pet, you know, should be able to live full lives. This is sort of the most abstracted version of that. But, yeah, I don't know. It's odd.
Speaker 1:More update
Speaker 2:on TBPN's Road to Christmas. Oh, yeah. We're down to just ninety three days. Ninety three Days. Lot to think about.
Speaker 2:Lot of moves to make. And we couldn't be more excited.
Speaker 1:Start counting it down. I was trying to think of a realistic AI doom scenario. Like, everyone's like, oh, I can't imagine it. And so I I actually mapped out one that maybe I could take you guys through. I wrote a little script here
Speaker 3:Okay.
Speaker 1:Called The Old Way. So, you know, fades in, starts in the TBPN UltraDome. Me, you, Tyler, we're all doing the show. All of a sudden, all the phones in the UltraDome light up. Emergency alert, unknown biological event, shelter immediately.
Speaker 1:Outside, alarms begin to sound across Los Angeles. A strange strange haze starts pouring through vents and drifting across the streets outside. I see a cloud approaching the Ultradome entrance. First thing I do, bioweapon attack, this is you what gotta do. Boom.
Speaker 1:Right there. It's not going in your mouth. You're good. Good. Second step, pull out your gun.
Speaker 1:Boom. Boom. Boom. Boom. Start blasting whoever's responsible.
Speaker 1:You're just firing wildly.
Speaker 2:Woah.
Speaker 1:Once you got the situation a little stabilized, then we gotta start we gotta figure out what's going on. So I go over to Tyler and I say, look, we gotta hack into the super intelligence. We gotta figure out what is going on with this with this bioweapon. And he's like, sure. Like, yeah.
Speaker 1:Let's hack into it. Like, I'll open up Codex. What do want me to prompt? I'm like, no. The AI has gone rogue.
Speaker 1:We have to do this the old way. He's like, okay. I'll I'll open Cloud Code. Like, what do want me to prompt? And I'm like, no.
Speaker 1:The old way. He's like, okay. I got cursor open. Like, what are we doing here? I'm like, no.
Speaker 1:Tyler, we have to go back to the old way. So, okay, got GitHub Copilot fired up. I'm like, no. Tyler, you can't use any AI at all. He's like, okay.
Speaker 1: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.
Speaker 1:We have to do it by hand. And Tyler's like, I I can't. I I don't know.
Speaker 2:You don't know how.
Speaker 1:Okay. I I don't remember anything. I can't write any code. So what do we do next? We gotta get George.
Speaker 1:He's the only one who can help in this scenario. So we go
Speaker 2:because he's the only one that remembers how
Speaker 1:to He remembers how to code. He's got local model stored. He's got his own hardware. So we go get George Hots. And then we're like, okay, George, hack into the super intelligence.
Speaker 1:Tell us what's going on with this bioweapon. And why aren't we affected? Like, we we we've been walking around. We seem to be fine. Everyone else is turning into a zombie.
Speaker 1:What's going on? So he he locks in. He writes some code. He hacks in. Figures it out.
Speaker 1:He says, Jordy, tell me about your diet. And you're like, well, I I just eat heroin 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.
Speaker 1:Right. And I'm like, but how do you explain me? I only eat Air One once a day. And he's like, well, John, my analysis indicates that you might have had the original Four Loco. Drink the original Four Loco?
Speaker 2:In in large quantities?
Speaker 1:And I said, yeah. I was daily driving in college. Why? I was like, your body consumed so much original four Loco, it's now essentially pure toxins. So the bioweapon sees it as a hostile environment.
Speaker 1:It doesn't even try to infect you. So you are immune. So we're both we're both immune. But the bioweapon is still wreaking havoc on America, so we gotta get to the bottom of it. So we're gonna need some muscle.
Speaker 1:So we go get Sam Sulek. So we go get Sam Sulek, and we're like, let's go. We gotta go to the data center. We gotta break in and shut this thing down. That's the only option.
Speaker 1:So we get in the Cadillac Blackwing. We jump it off of ramp, slam into the data center, pull out the gun, start blasting the GPUs. We're just firing everywhere. Finally finally, after we've shot 90% of the data center, it's just bullets everywhere. We get the k everywhere.
Speaker 1:Finally, a humanoid robot emerges. It's Mecha Hitler. We gotta fight him. We gotta fight him. We're shooting.
Speaker 1:We're shooting. The bullets are just are just bouncing right off. They're just bouncing right off. They're we're we're we're getting beat up. We're trying to fist fight it.
Speaker 1:Sam Sulik's doing nothing. It's not working. 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 kind of like a concrete example of like how AI doom, how, you know, a runaway
Speaker 2:scenario. That's one possible scenario. I put it because everyone's asking for
Speaker 1:put it at 10%. That's the way it plays out. Yeah. That that that's kind of where I'm at right now. Yeah.
Speaker 1:That's how I'm thinking about it. Meta, the AI agent company, Muse. They're putting humans in there. Meta is Finally. Human Concierge for its new personal agent, Muse.
Speaker 1: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. Amazon's like, no. No. No.
Speaker 1:We're gonna block your browser. Well, did you block everyone's browser? How's that gonna play out? I was thinking about, at some point, if I write a really, really bespoke, a like, basically a CLI, like a really bespoke weird CLI, like I vibe code it, but I use sort of odd techniques. Like, I'm like, okay, use computer use and open up Opera or, you know, Mozilla.
Speaker 1:And 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's like it runs on my home WiFi. So and I just do this. I I don't productize it. So Amazon, like, can't really detect it.
Speaker 1:Do you think I could potentially be in front of their battle forever and then puppeteer it from other agents?
Speaker 2:I just asked ChatuchPutti, can you order Amazon over the phone? Because I would think that at that scale there would be Yeah. At least one way to do it. Totally. It seems like you can only use their customer service number Mhmm.
Speaker 2:To help with existing orders. Mhmm. So one thing is you could order a bunch of the wrong thing, like a 100 of the wrong thing, a 100 like, you know, bug treatment
Speaker 1:Order everything and then have your agent refund everything but the thing you
Speaker 2:No. Have have your muse agent Yeah. Have a human via muse call Amazon. Oh. And say, hey, I actually got the wrong thing here.
Speaker 2:Can you swap it with this other thing? Interesting. And just do that as many times as necessary Yeah. Yeah. Yeah.
Speaker 2:To get the right order.
Speaker 1:Maybe have one order that's just always open where you're just adding and and removing
Speaker 2:completed orders because you have to have made a mistake for Okay. To call the customer first.
Speaker 1:Oh, yeah. You can charge me, but just add it to the order. I don't know. We're gonna get it's it's gonna be a weird battle. Ben Thompson was writing about it today.
Speaker 1:It was it was pretty pretty good. Amazon I mean, his 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 and that's basically what they did with OpenAI where OpenAI had the instant checkout process. They invited everyone to be a pit part of it.
Speaker 2:Amazon was like, no thanks. But we'll give you tens of billions of dollars.
Speaker 1:And But but interestingly, yeah. So now now the the result is that Amazon is vending in their ads into ChatGPT to close the purchase there. Toby Lookie at Shopify partnered with Muse and announced that they are integrating. But the interesting thing is, like, where's the value capture layer? And it only works with Shop Pay.
Speaker 1:And so it's a lever to get merchants to go on to 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 and whatever else.
Speaker 1:I'll build
Speaker 2:paper towel holder that's made by a company that's over over 50 years old. Yes. Yes. No. So the the crazy thing is so so now, Vinod was on the show yesterday talking about Wajo Oh, which is an
Speaker 1:AI We launched one.
Speaker 2:Faux. They have a new agent that I believe is human supported In that, some of the stuff is done by AI Mhmm. And some of the stuff, if necessary, can be escalated to a human. Mhmm. Now, you have meta off at least testing, exploring, having a human in the loop on some of these things.
Speaker 1:Yep.
Speaker 2:And my big question is like, why? Why do 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 like something I could text and say, hey, call this person, call that person
Speaker 1:Yeah.
Speaker 2:I'd probably be using it.
Speaker 1:I think there's a ton of 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 they've done over the past two decades. The other thing is that Muse specifically will not train on your data that you put into Muse. And so that this wasn't like a huge announcement or anything, but you could see people being like, I don't know if I want to use this AI agent that's gonna be using my personal information.
Speaker 1:Even if they're running like the PII redactor, putting all of that might be a little bit off putting. So Meta's 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 And so it's like the model by default can, I don't know, like summarize your summarize your calendar? Right? And give you a give you a briefing on what's going on in your calendar.
Speaker 1:It knows how to integrate. And the model's capable of that. Muse one point Spark 1.3 is capable of that. All the models are capable of that. But what can't they do?
Speaker 1:They might not be fully ready to 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 and, you know, creating more training data for them than a permanent solution to the product problems.
Speaker 1:Also, I I I just I don't think the time on for these apps is is actually that high. I 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. I didn't do any work. There's some feedback on Alex Wang's post talking about it'll help you with your goals.
Speaker 1:And I think Katie was saying, a lot of people just don't have goals. They just wanna chill. And this is also the Ben Thompson thing. Consumers wanna be entertained. People that go to the beach, they'll still scroll Instagram.
Speaker 1:Are they really gonna be doing anything with any agent no matter how powerful it is? There are plenty of people that have all staffs of EAs and personal assistants, and don't do anything on a weekend because they wanna chill and they don't wanna
Speaker 3:Well think about anything.
Speaker 2:One thing's for certain.
Speaker 1:It's over or we're back?
Speaker 2:Professor Zhang is vindicated. Yesterday, we watched Oh, yeah. We watched a video where he said, I guarantee that there are humans
Speaker 1:And what?
Speaker 2:Behind the chat apps that you use manipulating the information.
Speaker 1:Mhmm.
Speaker 2:And he is correct, at least in the short term.
Speaker 1:Nikesh Arora chimed in on the knife fight between Amazon and Muse. 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 I think Instinct might land with Amazon potentially, although they do have Rufus.
Speaker 1:But
Speaker 2:you can And Tyler is a a Rufus power user.
Speaker 1:Yep. Rufus is is good. You don't wanna talk 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.
Speaker 1:Like Amazon, like, you know, the trains run on time and 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.
Speaker 1:There's an interesting stat from the Strictory Post about Amazon. They the net income for Amazon e commerce was like 36,000,000,000 and the ad revenue was exactly twice that. So the business is unprofitable without advertising. They put advertising everywhere. So it's like a very interesting dynamic where you can see the motivation.
Speaker 1:There's $68,000,000,000. You you you flagged that number. It's a huge
Speaker 2:Over 70 in the last 12.
Speaker 1:Over 70,000,000,000. So significant pool of revenue and profit that they will be protecting for sure. Any other breaking news? Anything you've been tracking over the last couple hours while we've been live? I think we're pretty much good.
Speaker 1:Opus five five launch, we talked about that. Benchmarks look really good, very cheap. And the the key thing that you that you pointed out was, like, the second post is, like, this is the first model that we've released since we said we were pacing the frontier. This is an expression of the pacing like, this is,
Speaker 3:you know, Fable 5.1 class model, but it's it's much cheaper.
Speaker 1:So it's
Speaker 3:not efficient.
Speaker 1:Trying to be more superhuman, more dangerous. It's it's the safest, best, fastest, cheapest, lightest, thinnest model ever, basically. Yeah. That's And
Speaker 3:then did the same thing today.
Speaker 1:Okay. Yeah. They oh, so Soul's out? Soul. Yep.
Speaker 1:GPT six Soul. Six Soul. Cool. And then also Luna. Luna, but not Terra?
Speaker 3:Or I might have had that back then.
Speaker 1:Yeah. But new new models, the model wars never cease to to entertain. Of course, all that matters is what you do with them. So thank you for watching TBPN. We will see you tomorrow at 11AM Pacific.
Speaker 1:Leave us five stars on Apple Podcasts and Spotify. Sign up for our newsletter at tbpn.com and throw that flash bang. Tyler Cosgrove.