Trump's First 100 Days, Tariffs Impact Trade, AI Agents, Amazon Backs Down

25:59Kamala Harris gave every indication she would have continued it. What was that path? It was a path of continued escalation and doubling down in Ukraine. Recall that it was Biden himself at the beginning of the war who said that if we give Ukraine Abramstanks and F -16s or attackams or high mars or if we allow them to hit targets inside of Russia, it would lead to World War III. He actually used the word Armageddon. So at the beginning of that administration, they were very concerned about how an escalatory path could lead us into direct conflict with Russia and World War III. And yet, despite that, at every fork in the road where they had a choice, they ended up doubling down.

26:41They gave the Abrams tanks, they gave the F -16s, they gave the high mars, they gave the attackums. And finally, when Biden was elamed Duck in his last couple months in office, they did the most reckless and irresponsible thing, which is allow American weapons to be used to strike targets on Russian soil, not just fighting in Ukraine, but on Russian soil. Moreover, we now know from a New York Times article that just came out in the last few weeks that it was American generals and American intelligence who were planning this war. So when you're talking about striking Russian targets on Russian soil, it's not just the Ukrainians using our weapons.

27:14They're using our targeting, they're using our guidance, they're using our satellites. I mean, we are deeply integrating the kiltchain. This is the United States being a cobalt during the war, hitting Russian soil. That is incredibly reckless and dangerous. I have no doubt that if the Democrats were still in office, we would be in an escalatory spiral right now with the destination being World War III. I do think that Trump has pulled us back from the brink there. There's obviously still more work to do, but I really appreciate the efforts that Steve Wichoff has undertaken for the first time in three years.

27:46We've at least had direct diplomacy with the Russians. We weren't even talking before. We weren't even talking before. Talking is a great thing. And apparently we're going to keep supplying with weapons as long as they pay for them. So it's going to be very interesting to see how this all hashes out over the next hundred days. Or so, let's keep moving when we're looking. I don't think we know that. Let's wait and see on that. Okay. Yeah. I mean, I think that's what was reported, but you're right. We should wait and see. Okay. Downstream tariff impacts. We got to talk about this. And this is why we have you here.

28:17Ryan since you're in the thick of it. You treated a thread last week about the lag time of shipments from China. And when you were on, I guess during COVID, you really educated us to how the supply chain, how the supply chain works. And according to the thread that you shared somewhere around early June, we're going to expect warehouse is trucking the entire supply chain. Maybe to start to seize up or layoffs, I don't know how you would frame it, Ryan, but are we past the point of no return with regard to the supply chain? Is there an off ramp for this tariff conflict war negotiation with China in your mind?

29:00What are you seeing on the streets and in the purchase orders and the invoices at Flexport? Definitely not past the point of no return. I think we're still right in the middle of the, don't judge the cook while he's cooking. We want to, you know, Now let's see what it tastes like at the end is I take up starting point here and we're still, they're still in active negotiations. So I don't think today's, it's not static. Now the world does want a lot more certainty and that's a big cause of what's happened here and what has happened is a 60 % decline in bookings of ocean freight from China to the US.

29:32I mean, so that's really, really pretty dramatic. Like probably exceeding what was expected. I don't think, you know, when they issued, when they rolled out the initial reciprocal terrible tariff plans on April 2nd, it was meant to be a 54 % tariff on China, then there's multiple cycles of escalation. We ended up at what's now 154 % tariff. So this is just a lot higher than anybody planned for. And so therefore I don't think anyone's planning for a 60 % decline in ocean freight. Ryan, let me ask you a question about that. Are people actually paying that 154 %? there's been this discussion online and it's sort of unclear from the administration and from retailers stuff that's landing that people ordered before April 2nd.

30:20Are they actually paying the 154 % on top of what's landing? It's live now. It was based on departure date, so goods that departed China after midnight, Eastern time on April 9th are subject to the tariffs upon arrival. And so now enough time has passed. They're pretty much all the ships that are arriving now left China after April 9th when that started. So what happens? People are paying it or people saying I won't take delivery because it's Jason, you have to pay it right and correct me if I'm wrong, but you have to pay it at the dock in order to get the goods released. More or less. More or less that's true.

30:53They allow you to have a bond in place so you can pull the goods out before you pay. But the money's owed at that time and then you get you get like a two week time frame to actually make the payment. But there are strategies here. A lot of people are doing that you can use what's called a bonded warehouse and move cargo into this warehouse. And then you only owe the duties when the cargo leaves. That's what I was asking like, is there a hack here to that? That'll let you defer things. And it's very, very common right now people are searching everywhere for bonded warehouse capacity because in a bonded warehouse, not only you defer payment to when the cargo leaves the warehouse, but you only owe the duty amount based on at that date.

31:30So if the duties come back down, which a lot of people are betting they will on the China specific duties. You'll actually lower your tariff burden. And then there's another hack for this, which is effective. Use a Mexican or Canadian bonded warehouse. So you move the goods into Mexico. And then you actually only technically import them into the US at a future date when tariffs are lower. So I understand a lot of a lot of companies are doing that right now too. We're helping some people with that type of strategy. But yeah, sorry, Ryan, do you think that the government will they view that okay?

32:00that kind of hack. And you know, like, if you look at the GDP numbers, one of the craziest things was the inventory pull forward that people did to your point, like trying to get as much stuff into the United States before April 9th as an example. Yeah, I mean, it's not a hack. It's a bondded warehouses that have been around for decades and they're very commonly used. I don't know that it'll be that material in the scheme of things that it would, you know, cause a change in law that are around bonded warehouse. So you don't think, for example, the Department of Commerce will have an issue with the strategy of sending inventory into Mexico that essentially you're essentially like isn't it in work around?

32:40So it's a work around, like instead of paying the China tariff, now you pay a Mexico tariff, which should be less, is that the idea? Well, you can move it into a bonded warehouse in Mexico even and not pay Mexican tariffs either and you just wait until it imports. But what's the department of commerce or the customs to do? It's sort of you just to lay it important, the goods. You imported them in the future and you know, it doesn't. I wouldn't even call it a act as just sort of like people are going to get creative here. You know what I mean? That's the job actually. The government should set the rules and the rest of us got to figure out all right.

33:10How are we going to compete and make money in this environment that they've created? Ryan, in that tweet you redid, which was a pretty dramatic tweet painting of very like, I don't know. Like a pretty dire situation. Where are we at in terms of how dire this will get or resolvable? Painless, the best case scenario and what you expect could happen in that case. Or if this gets extended, are we going to see as people are hand -ringing empty store shelves, Christmas gets ruined and all of these layoffs start happening in the supply chain? Take us through the two scenarios that people are debating.

33:48Yeah, I mean the the the bleak scenario, which is I don't really think it's gonna happen I think that the administration doesn't want this to be their legacy that they like created a policy that just like kind of tank Small business and supply chain so I don't I don't actually think this is gonna happen, but the bleak scenario is Teriffs stay at this level for one forty five percent on China The 10 % goes way back up to what it was originally announced in reciprocal tariffs So there's no like safe haven for tariffs and trade just falls off a cliff and a lot of companies go bankrupt in our in our cut especially small companies are the ones that are importing from China.

34:18Reality is like tariffs have been high on China for a long time. Labor costs in China are not there for cheap labor. You're there for quality manufacturing at this point. There's much cheaper labor in Southeast Asia or other parts of the world than there is in China. So you're in China because of the manufacturing capabilities, the ecosystem, not just for cheap labor. And if you could have moved, you would have already, even the 25 % tariffs from the Trump's first terms were pretty high enough incentive. And so that's the bleak scenario is that small business starts getting wiped out. The ones that are buying from China.

34:50And it's a lot of brands like it's not just Amazon seller selling stuff that you don't need. It's like all the brands that you know are like, you know, interesting brands, apparel brands. I had cuts clothing on this week and start up last week. And he said there's going to be like, if this doesn't get resolved and like, let's say two to four weeks in his group chats, people are going to start layoffs and they can't physically restart the supply chain in Vietnam or wherever to make t -shirts. So Aaron, what's your thought on this as well? Just bringing you in? Sure. Well, first of all, I mean, Ryan has supplied me with a high degree of doom scrolling.

35:29And I just like a horror show reading his tweets. First of all, I would feel better if the messages out of the administration were either more kind of consistent or that there was a logical connection between do we either want to raise the kind of you know tariff revenue stream or do we want free trade like like those things are working against each other because like depending on who you talk to they say this is a mechanism to bring down income tax which obviously then by definition means that they expect the tariffs to sort of persist which is totally different from let's go negotiate deals that just allow for the free trade to actually increase.

36:08And so are we worried about the reciprocity or are we worried about revenue streams? So that's a whole issue. You also have this issue, which is the messaging from the government. And the meta point I'll make in a second is about how we could have actually accelerated into the transformation of the economy. But you have folks like Lutnik, et cetera, going on TV, talking about the end state of our economy, which are actually probably fine messages. But we haven't seen what that vision looks like. So everybody is kind of confused. Like does this mean that we literally go into manufacturing plants and we're like the ones literally doing the screws on an iPhone?

36:42Or is it a bunch of jobs which are next generation jobs, which is like we're managing robots and like shipping logistics grows as a result of this and all of the surrounding kind of supply chains you know start to grow. So like, like there's an underlying, most people on this call of managed teams, like you do change management. You lead people to the end state that you want them to sort of see the potential in. And I think something that kind of gets missed is, and the part that kind of confused me is, I don't know if exactly to Ryan's point, people are in China because of the ecosystem of manufacturing.

37:17Yet the messages you get out of the administration are like, oh, we're going to have fewer toys at Christmas time. It's like, no, that's not the big picture. Like the big picture is like, this is supplying the parts that go into building a manufacturing plant. and building a car that allows us to actually, you know, be even remotely competitive in car manufacturing. So we're shooting your mind this all lead because you have some thoughts on American exceptionalism and maybe skating to where the puck is going. So if you were to become an advisor on this, as a technology expert and somebody who spent the whole career in it, what would you advise them to do?

37:53I'd get rid of Navarro immediately. And you would basically say, you know, me a culpa like oops and like obviously you need to like land that with some really cool trade deals that make everybody kind of feel happy. And you basically say you know what? Like let's go back to the first two days of Trump which is let's announce massive deals. We're bringing manufacturing here with Stargate. We're doing TSMC. We're building Nvidia chips. We're gonna do a deal which is you get like 5 % tax rate if you build in America and so you just stimulate a mean manufacturing boom in the country. We incentivize automation across the manufacturing.

38:31We use that as a competitive weapon to go and compete with the lower cost labor that happens internationally. We find every incentive and tool we can, we deregulate, you allow people to build these plants and so you don't have to go through the three -year EPA process like you just accelerate from this position and you see it all as upside. And so then business leaders, if you go talk to the Fortune 500 company that actually has to build anything right now, you give them a path to say, listen, we're going to help you transition away from your current supply chain and we're going to make it even more competitive and more compelling in America to do that.

39:09There's a reason that Elon builds in America. He has actually made it be more effective to be able to bring automation to manufacturing, to be able to build locally, but he wasn't forced to do that. And so I would just I would argue like you you use as many carrots as possible In some surgical areas in Chimoff. I've heard your points about the like you know chips Farma, you know AI like in those surgical areas we get tough weren't where necessary And and if we have to do you know a couple sort of very surgical tariffs Yeah, you know to kind of make make people move the direction that we want that's totally fine But but I mean like it's like that even arguing the premise is hard because because we act like like we're like like it's like countries that are screwing us, but actually businesses are independently making decisions about where they want their supply chain to exist.

39:58They have this in a free market, they've made that decision. They don't need the government to tell them where are they supposed to or where they allowed to to have their supply chain operate. That ends up with just lots of economic distortions that everybody on the right would have called, the left socialists for trying to kind of implement central planning around supply chains. So that's my piece. Well, what do you think, Chimuth here, of this sort of reframing slash off -ramp and sort of maybe the positive spin on it? Hey, if you want to make T -shirts, you want to make commodity items, have at it, free trade, you know, reciprocal tariffs, great checkbox there.

40:33But here is a series of incentives and a path forward to do the advanced stuff, to do robotics, etc. Let me answer this in a different way. Okay. A lot of those things he's actually doing. I think this is where we are is where beyond TDS, there's something that comes after it. And I think that the mainstream media has just lost their mind to a degree that they hadn't even lost their mind in Trump one. I'll give you a couple of examples. Well, one example, by the way, just a shout out to our friend, completely brazen, ridiculous, as shrily reporting by the Wall Street Journal last night when they were told that this, you know, this whole Tesla thing was a total farce, they continued to publish it.

41:22Okay, fine. They had a phone to pay. They had a phone to pay. You know, the board, yeah, starting a search, you replace you on and then the board said, well, wait, we told you, we weren't doing that. And they didn't even mention they communicated that directly to the Wall Street Journal, the Wall Street Journal said, I don't care, I have an axe to grind. Correct. Yeah. I think that Trump has a strength, which is he shapes these potholes for the mainstream media to fall into. The downside of that though is that the mainstream media then doesn't do the other part of the job, which is to tell the things that are important.

41:55So for example, we spent a lot of time breathlessly talking about the MS -13 knuckles of the guy, right? Or then we We spent a bunch of time talking about how MSNBC blurred out the names of the placards on the lawn. Okay. But here's the other part where then they get so tilted, here's what they don't report. They didn't report that, for example, when Trump took a shot at Harvard, he also reinforced and strengthened historically black colleges and universities. Totally did not get rhythm. I'll give you another example. This past week, Bessent said that the tax bill will allow you to fully deduct all the PPE and all of the incidental costs of building a factory.

42:46I heard that. I immediately went to my wife. She runs a farm of business. This is exactly what she's trying to figure out. And we now are like, how do you build a business case if this actually gets effectuated? The point is that thing would create an absolute economic bananza if it were to get passed. Other than people hearing it on this pod or randomly maybe finding it on a direct clip that Besson puts out on X, there has been zero coverage by the mainstream media. Yeah, but like the first of all, yeah, okay, MSM and whatever we want to call it aside, Like the that is that is still on the administration for driving a a change management process that causes people to build on momentum and not causes boards To basically say oh, we're gonna pivot our entire supply chain this week because because Trump you know Didn't get a callback from from she like like that that is like it is this is really not a a a TDS MSM issue This if you talk to fortune 500 CEOs.

43:53Did you know about the PPE thing? No, but that's not but like I don't need to like the thing that I know, but there are many other CEOs that do. They're controlling trillions of dollars of capital allocation. It's an important thing. If we had Mary Barron this call and we said, Mary has Trump increased your ability to execute and operate and accelerate the transition to the US, or as he had headwinds that make it tougher to navigate right now, which way do you think should go? I think that she would give you a calculated answer that neither is pro or con. I think that answer changes by the day.

44:28I think that if you talked to her last week, she would say this has been a major headwinds. And then yesterday to Chimau's point, they did this thing with you can depreciate or fully expense in your one capital improvements or building out factories. But also earlier this week, they made it so that auto parts are not subject to the tariffs. They created a huge exemption. Yeah. It wasn't there at the start. Everything like it's great. But by the way, if this was a bit more patient, it should have been there in the beginning. because these auto companies were saying, hey, this is gonna bankrupt us if we have to pay taxes on the land.

44:58Where I would take you, start going back to communication and making a crisper. And you're right. Aaron, where you are right is, my expectation is it's my job to stay informed. Okay? As a CEO of my company, I try to stay informed. And you're right. It is hard because sometimes I find myself hunting and pecking to find the things that matter. But I do put a bunch of that responsibility into the lap of the people that are supposed to actually report the facts, they can choose, they didn't have to run that article about Elon, which turned out to be total bullshit and horses on the front page of the Wall Street Journal.

45:34They could have talked about what Ryan just mentioned as the first article and said, here completely changes your Roy can R a week calculations for 90 % of the S &P 500. That was not the article they chose to write and to publish. But I also think it comes back to my original point around the Udah loops that their the Trump administration is running these very tight. Hey, let's take an action. Let's see what happens. Let's see the reaction and then take another action in Washington's used to doing all these committees that plan everything for five years or something or whatever 18 months and then roll it out slowly and they're going, Hey, let's roll it out.

46:08Oh crap. We're about to cause this huge problem in the auto manufacturers and they're all telling us they're going to go bankrupt. Okay. Three days later they push an update to that. It's kind of now. It feels chaotic. Yeah, the summarize, Ryan and Aaron, you're a position, so we can keep going through the docket. Hey, a little less shock and awe, maybe a little more predictability, a little crisper communication and chamauth, I think Europe position is, hey, maybe the mainstream media can play a better role here in focusing us on what matters. Yeah, that wouldn't be my takeaway. So yeah. Okay.

46:37What's your takeaway? I mean, like zero shock and awe. Like, not a little less. Like, like my strategy would be 100 % different. Actually, Scott Besson has an incredible podcast from September of last year. And he basically said, Bidenomics got it all wrong. And I was like, listening to him. I was like, oh, okay, actually, this is kind of cool. Like, he basically says, deregulate the US. Make it easier to build manufacturing the US, increase the GDP, and then you'll be able to take in less tax revenues, spend less in the government. And it was like, oh, this is actually like a glide path. We could take the fact that we did a soft landing relative to the rest of the globe, we're winning an AI, we're winning in a number of categories.

47:20We obviously need more energy. We need to bring in manufacturing into the US. And so you have this great momentum, which is where we are the tech leader in the world. Let's just pour fuel on that. And so to pour fuel on that, you just do a series of carrots and the winds that build a flywheel of positive energy. Like the reason why I take a little bit of exception to Chmas and SM point is that I I think to some extent, fortune 500 CEOs are not the ones like, oh my gosh, like Rachel Maddow said this, like I'm gonna go and, and you know, worry about this topic now, like the, like the information coming at them is, is I'm not talking about the information that's presented.

47:58I'm talking about the information that's excluded. How do you get the information that's not published and shared broads? No, no, no, no, come on, like Goldman Sachs and J .B. Morgan are not writing reports on the fact that we might enter a recession because of MSNBC's reporting on this topic, like, But again, that's not what I'm talking about. I'm saying like, glad handing some high level prognostication, which nobody ever gets right, is in my opinion worthless. What I'm talking about is the details. So when you talk about something as narrow and specific as excluding PPE or allowing you to double or triple depreciate something in a given calendar year, that is narrow.

48:37It's precise. It's specific. It's actionable. and what I'm saying is if I surveyed the 500 CEO of the S &P 500, dollars to donuts, the overwhelming majority would not have known. And had they brushed up against that somehow in their normal media consumption to then ask their teams, the odds of that would have been zero as well. And so I'm guessing who do we blame for this, Jamal, is it the administration's job or a Measuring Media? But I bet you everybody knows about the blurring out of the stupid pictures on the lawn and the MS -13 knuckle tattoos. Yeah, okay. So let's wrap up on this just really lightning round here Amazon flip flopped on a new tariff notification on their websites.

49:18Trump said he had a great discussion with Bezos. He solved the problem very quickly. He did the right thing. Good guy, etc. If you haven't seen this, it's something that Timo is doing. Here's what Timo does currently today. Nick, you have that image if You could pull it up of just when there is a tariff, they explain the tariff coming into the country. They put it as like a line item. I thought this was actually kind of cool. I don't know why people take offense with Ryan. This is pretty standard stuff. So Amazon's competitor, Timo, is putting in the import charges. They don't say tariffs. They don't say taxes, import charges.

49:53This is like a standard thing. What has happened to other countries too? What is this? What is Timo? Is it like a dollar store? It's basically like a dollar. This is the last place you would ever buy. You never bought some $5 jeans. No. Yeah. You can buy $12 jeans. Basically, you're left sock from Laurapiana cost less than Timu's entire inventory of jeans. The point being, I thought this was actually a plus for... I think that Amazon told him this, played this. They did it. They rolled it out and they got criticized. I think they were called a treasonous company from the White House press, you know, by the press secretary.

50:28They told him this played this because they should have gone and leaned into it and said, yeah, we're showing you all these tariffs when you buy from China. If you buy from America, you don't have to pay any tariffs. I come on, that would last in three and a half seconds. This is exactly consistent with the other issue, which is they're playing whack -a -mole. Okay, we're going to do something with the automakers. We're going to try and solve some problem with Amazon. Like, this is a sign that it's not a good strategy if you have to do this much whack -a -mole. Like, they can't cover up what Amazon is going to end up dealing with because there's going to be 500 other retailers that don't get the call with Trump.

51:06So, to me, that's evidence of clearly, they didn't think through the entire downstream set of conditions that are going to change as a result of this. Sure. Yeah, I thought this was a big win, Chamoth, because they could then have Amazon. Here's a mockup somebody made. I'll pull it up here. It was interesting. They could, to Aaron's point, just show, hey, here's a bunch of American companies by American when you do a search. That's Ryan's point. Oh, sorry. Ryan's point. Hey, here's what it might look like. Pull that oral B -tooth brush one up. Nick, if you got it right there. So somebody mocked this up.

51:43I think this could be the hugeest win. You could have the retailers do by American, buy it once, buy a high quality product from America. If you look here, we don't have the products. It wouldn't work. We don't have the products. Well, I mean, we do have for some products, you know, American made products. You know, I buy my boots from Danner and those are all American made. So yeah, so we should just go back to communism and we're all going to make our shoes. Like it's like that's like we're in a global market. Like we've been from everywhere. Chimap any thoughts on this? I mean, obviously there's the whack -a -mole angle.

52:19There's buy American and be proud of it. Here's the narrow communication. Here's the narrow question. I got a bunch of emails from people. And a bunch of them were Amazon sellers. And I don't know Nick if you can find it, but I posted their comments and I reshared them just to kind of highlight the issues that they were going through. And at the core of it was a feeling by them that Amazon had abandoned them as American purveyors and sellers of goods. and that Amazon on the margins had a tendency to help competitors from abroad come stand themselves up and compete and essentially cannibalize on price and margin.

53:05This is my view completely and this is the biggest opportunity that I think the Trump administration is flying at 40 ,000 feet doing macro level negotiations and look and failing to see some of these micro optimizations that are really, really real. So in the United States, you can import goods as a foreign company. You do not have to create an LLC or any sort of registered entity in the United States to import goods. Sometimes they say, oh, Americans pay the tariff. Like that is not true. In many, many cases, the foreign companies just imports the stuff and they sell on Amazon. And when they get caught cheating, they can lie about the valuation and pay a lower tariff.

53:40They can change the classification and pay a lower tariff. they can import stuff that you know it's harmful to children has lead paint whatever else There's no enforcement at all you can't go like Amazon third party is like a bit of a backdoor to abuse the system right? I mean Amazon is just playing the game that's on the field But they are this legal in the United States is these companies import stuff and I think it's 60 % of all of all the sellers on Amazon are these Chinese Registered they're not registered in the United States at all just Chinese companies sell it which sounds profoundly unfair to take it a level playing field.

54:09We just take a step back and also acknowledge that we are talking about a level of detailed issues that we would have never have talked about six months ago, that there was no interest in even bringing this up. Like if Ryan wanted to bring up the sort of hollowing out of American salesmanship, let's say if you will, because of like this arbitrage at Amazon does for GMV, that would have been a snooze fest. Except today, it can actually get a lot of awareness. Aaron mentioned this, and I've mentioned this before, but I think that there are four things that really matter, batteries, AI, pharma APIs, and rare earths.

54:50That now is on the agenda. I think the positive way to look at this is the American economy is too complicated. If you had waited Aaron for a study for all of the implications, we would have been waiting forever and nothing would have happened. And I think that we've made macro level moves. You're right. And now we are finding what the implications are in course correcting in real time. And I hope what happens though is when we find these big thorny issues, I think the Amazon thing is a pretty interesting issue actually about like American competitiveness. Now the question is, do we follow through and get to the root cause of it and fix it?

55:26And those feedback loops are there. I mean, the Trump administration is going to act on this. And there's an act that's coming out of Congress as well. they shut down the foreign imported records. So those feedback loops are there in ways that I don't know if they were there in the past. Go ahead, Aaron. We can totally chaos monkey the economy and just see what breaks. I think that the, you know, you sort of phrase the, we could, you know, do the research paper and we could do, you know, ask an institute as a bad thing, but also it can be a bad thing if you're the small business owner right now who has 30 employees and you just literally don't know what you're gonna do next month.

55:58And so, so that's sort of, then the argument to not. That's like why you do have some bureaucracy and you don't cast monkey the economy and why Rand Paul is literally saying maybe we shouldn't actually let you know have you know lateral control over tariffs So you know interesting that that like dynamic there All right, it's actually want to wrap us up here or you want to pass Am I still on the pod? You turned your camera off so look I spent 80 minutes to be in this topic with Larry Summers three weeks ago. The point I made then is that we had in this city for 25 years a globalist consensus on trade that distorted a lot of outcomes.

56:43I don't need to rehash that debate, but I'll just recall that Larry Summers' main argument for why this would not work out is that the market was down. Do you remember that? That was his evidence that this wasn't going to work. Okay, and it was all about the market not pricing in lower expectations. Well guess what the market is actually up since liberation day on April 2nd So what happened three weeks ago was basically a Panic in the market over this policy and the media has been trying to fuel that panic now What I said as well as we do have to stick the landing on this I mean president Trump shifted the conversation away from this globalist consensus and he's now redefined the debate but it is now up to Scott Best and the Treasury Secretary Howard Leidenick the Commerce Secretary.

57:32James and Greer the U .S. Trade Rep and so on the Trump trade team to now negotiate these deals, stick the landing and I agree with you to the extent that the sooner that is done the better because it is good to provide business certainty. But the idea that so far this hasn't worked I think again the main argument against that that was a market reaction that now the market's not positive. So I think my point is just, we need to give this time to work. I think it's too soon to be judging this policy as if it hasn't worked yet. It needs to be executed properly. And quite frankly, Ryan, I mean, I remember the last time you were on this pod, you were coming on about, wasn't there like some union deal that was supposed to shut down all the ports and all the shelves would be empty?

58:15That never happened either. You did happen. You did, they shut down for three days. I don't remember the show's being empty, which is now the new panic the media's trying to create. So look, there's a lot of pants wedding that's occurring here that's being filled by the media. Well, I just want to bookmark one thing, the only thing, I actually, I was more frustrated listening to the Larry Summers and you have conversations because I was like, why, Larry, make this point, like come on, like, well, don't go down the WTO rattle, like that's not relevant. So here's the only thing. The, of course, it's relevant, it's how we got here.

58:46No, no, like as in like that's 25 years ago. Like let's worry about literally today and like what do we do going forward to this? And the only thing I just wanna say because I do think that I appreciate your point about hey there's like everybody's freaking out whatever, but to be totally fair, that some of that freak out, whether we can decide how emotional it needs to be, is the reason that then Trump walks back the things that then caused the market to correct. So you can't just say the market's back and see we didn't need to freak out, because it was literally - I said we have to make the deals, we have to stick the landing, but look, China over the last 25 years has been able to strategically annihilate our rare earth processing capability and our ability to cast rare earth magnets.

59:31We just sat back and watched as the market basically went to the lowest bidder, which was being subsidized by the Chinese government, which the WTO allowed them to do. We should totally tell you a critical dependency Yep. And our supply chain on China for basically every electric motor in every product, including cars. That was crazy. We should not have allowed that to happen. Yeah. I'm going to change that. So we needed to shift the political conversation to recognize the ways in which free trade, led to unfair trade and created an acceptable dependencies on the American economy. Wait, wait, wait, just I just want to make one point.

1:00:08It's more than that. point. And then I'm, but this is the national security of the United States that that's at stake. Let's go take your favorite pet issue. China invades Taiwan. Okay. And we have to take a side, Jason. My pleasure. Hold on. Let me just finish. And the Chinese say, here are the implications of supporting Taiwan on this. A, B, C, and D. You don't get any pharma APIs. You don't get any rare earths, you don't get any batteries. Okay, it'll send life back 50 years. Or let's say China and India get into a fight and we're forced to pick aside, same situation. The point is there's all these scenarios that we never even considered us being able to have strategic optionality to make the decision that's morally and ethically right for the United States.

1:00:59And I think that we have learned through this lens that these are huge issues. the thing that the Chinese did that was so brilliant, which we still don't have an answer for, is they have these national champions and being a national champion allows you and we'll talk about this in AI. It allows you to blur the lines between the public and private partnership. It allows you to blur the law. It allows you to blur a capital. And I'm not saying we have to do that, but what I am saying is we need to have our own answer to it. And that was never on the table until April 9th. All right. So yeah, but like 100%.

1:01:36Like do that strategy and then and then don't have a matter of. It is the strategy. No, no, no, because because that's a you if you have a you know what Ryan what is the number? I don't know. Trillion of imports or whatever like you don't need everybody then then jamming the system to build their supply chain. You know in the US to solve that problem immediately. If we're sitting here in nine months and you're saying this and there are no deals, I would say that you're right. What Howard Lutnik said last week and again, we may have all gotten caught up in the knuckle tattoos and we missed this, but he was very clear.

1:02:06We have a country, a deal is already done. We're convening parliament. It's going to be the first of many. So for all we know, there is like 30 deals that are waiting in the wings and the first one will set the tone. And I think that Saks is right here, which is it's way too early to declare defeat and that it was quote unquote chaos. I think if we're sitting here in nine months and foreign direct investment has shriveled up and Domestic investment has shriveled up because there is just no continuity. You have a claim But I don't think that'll happen. I don't think that'll happen I think we will end up in like I'm with Ryan like we will end up in a good spot because we'll iterate through this My only point is there's an alternative path that could have occurred It could have been done in a more thoughtful well -communicated pattern instead of hey Let's do barrel rolls with the airplane.

1:02:54I don't disagree with you Aaron and I can tell you We've already started to see layoffs and everybody wanted to even initiate the barrel roll guys. Yeah, okay We got it. It's like hey, I don't want anything to change. I think we'll agree to disagree on this Aaron, where are you with this? Hold on, hold on. This one last point next. Excuse me. Excuse me. You didn't call me for 40 minutes I just want to make one final point here. Here he goes Aaron. Where were you with this perfect plan? Yeah Where are you with this perfect plan before liberation day? I was telling you tell me about it You're telling calm okay great.

1:03:26He was no we're having nobody nobody all happy mind ties And they were talking about this Specification all the people who suddenly know what the perfect plan is and how to perfectly execute it No barrel rolls had nothing to say about this topic for 25 years now of a sudden they've come forward with their perfect plans I would say that's where I want to finish. It's like, look at this. The best thing of all of this is you've got the liberals embracing Milton Friedman and their backgrounds on there. I love it. Yes. All right. Listen, we've agreed to disagree. We're going to agree to disagree on this one.

1:03:58All the liberals love the stock market. Listen, SACs, Commal is coming on next week. We're going to make some cocktails. It's going to be wonderful. We'll ask her some direct questions about it. But I want to talk about AI agents. 2025 shaping up to be the year of AI agents. tons to talk about here, OpenAI is planning to charge between two and 20 K a month for different levels of AI agencies. These would be basically Cron jobs. They would run the background and do things for your company that humans are doing right now. You may have heard of this agentic tool. Again, agentic is just a fancy word for agent, which is a fancy word for like a Cron job that just runs.

1:04:34No, propensity. Yep. And Manus is the company in China that started this. Weirdly, benchmark, invest in it. That's created a whole fluff law on the side. And Manus' website has a really good visualization of what these agents would look like. So first of all, I think you're giving me a little too much credit to Manus. They didn't come up with the agents, but I do think that they have a very good demo. And it's hard to know exactly how real it is because everyone's used it. But if you - That is from China. It's from China. I'll get to that in a second. If you go to their website, you can see a bunch of their demos.

1:05:07And I do think that what they deserve credit for is advancing the ball on the UI paradigm. And it's not that other people weren't doing this. I mean, I think notably, and Thropic was doing this with its operator product. But the basic idea is that you've got this two -pain view. And in one window, you've got the standard chatbot interface. And then in the other view, you can see what the agent is doing. And that agent has the ability to toggle between it currently four apps. their search, browser, code terminal, and document editor. And so when you give Manus a task, the first thing it does is create a to -do list in the document, or you can kind of see it there.

1:05:46And then it works sequentially to achieve each of those tasks and then puts an X on there. And you can kind of see it working. And I think what's cool about the demo is just the way that it seamlessly toggles between those four apps. And you can see what the AI agent is doing. You know, and it's browsing the internet, it's searching for things that's writing documents, it's crossing things office to do list. Now, I think it's pretty easy to imagine where this goes, which is you'll be able to connect an agent to all of your SaaS apps. So it won't just be four applications, it'll now be connected to dozens of applications, including ones that already have your data, and it's gonna know what actions it's possible to take in those apps.

1:06:24So when it creates this to do list, there's a much wider range of things that it can accomplish. And in fact, there's a new standard called MCP, which is taking off like wildfire, which is built specifically to enable agents to connect with applications and understand the data and understand the actions that are possible in those SaaS applications. So look, Manus is just to the iceberg here. I think this will become a very standard UI paradigm. That's the reason why I mention it, not because I am predeclaring them to be the winner in the space, but just because I think there's a lot of talk about agents and I think it's hard to conceptualize what that means without just seeing it visually.

1:07:01Great, great summary there, SACs. And Aaron, I want to get your thoughts on it because obviously you're running box and you have your finger on the pulse of this. We actually started building one of these in our venture firm. We have 20 ,000 applications a year and we have updates coming in from investments. We are now taking those SACs, Aaron, and we are having an agent sort them and then look for competitors and compare them to the last update and we're looking into our notion, our CODA and saying what else have, where are the communications that we had? What questions should we ask about this startup and about their strategy and then we're presenting that in Slack to our team.

1:07:38So this is coming fast and furious and we spend, I don't know, probably 15 minutes on each of those incoming applications to start doing the math on that. Talk about 5 ,000 hours of work. Aaron, what are you seeing on the street? What are you doing at box in terms of agents landing right now in Q2 of 2025? Yeah, I think I think I think sex represented it well, which is which is you know you have to now think about AI as as effectively being able to do anything on a on a computer or another piece of software as as a human can do and the little distraction that that I think happened two years ago after the chat to be team moment was we sort of thought about that as oh we're just going Now, I do typing information retrieval, and that's a new paradigm for user interfaces, let's say.

1:08:22So you just talk to your software and you search Zillow via chat. That was a little bit of a distraction. That's super helpful when you want basic information, look up, or whatnot. The big breakthrough is starting to think through these things as full, effectively, agentic systems that operate on any amount of data, any amount of tools, for as long as you want to complete any tasks that you want. And this is sort of the big year where agents are starting to enter the vocabulary of enterprises, of IT people, of larger and certainly small organizations. And it kind of requires you to have a little bit of a reset moment on how you think about AI, which is it's not just now a kind of a co -pilot that you talk back and forth with.

1:09:07It's actually something running behind the scenes that's Now actually starting to deliver real automated work for you. And so lots of implications, massive implications to what the software business model is in the future. I would argue strongly that it's a massive tam increase because now software starts to go after labor spend. It completely changes the dynamics of then how do you build a moat in a world of AI agents. But I think part of that piece there, you said something very interesting. that your software then is going to go over human spend. Yes. Explain that concept. Yeah, I'll hack for a second.

1:09:44David, and I, you know, we go back way back in Sassland, but like you used to basically just, you know, you built a piece of software and you sell it for the number of people in the organization. And so, you know, company has 500 employees and you sell that thing for, let's say, $10, you know, a user a month, you know, 120 bucks a year and you make $60 ,000. And so that's kind of the business model. Now, when your software actually brings the underlying workflow to the customer or the underlying outcome to the customer, so that company might have 10 lawyers. And so previously, if you were selling software for lawyers, you had a maximum amount of 10 seats that you could sell.

1:10:25Now, all of a sudden, if your AI agents are doing the equivalent of, let's say, paralegal work or some form of professional services, all of a sudden, you might be able to sell a multiple of the initial kind of 10 seats that you would have sold previously. So you said it has a huge opportunity because now you're not enabling a human to be 5 % more productive, you're replacing a human or you're replacing one out of 10. Yeah, and actually I'm gonna take a massive, I'm gonna do an underscore on this point though. I don't like the word replace because I think actually most of the upside is actually gonna be for companies that now deploy labor at things that they wouldn't have deployed labor at before.

1:10:57And maybe I'm biased from the view we have but most of our conversations with customers are when they have AI agents, they can actually now go and actually deliver work in areas that would have been unaffordable previously. So they actually, they weren't doing the work. And so, it's out true in logistics. So like we're making thousands of phone calls a day using AI calling truck drivers. And if we have a load, we've got 400 ,000 truck drivers using the mobile app, the FlexPort mobile app. I don't have enough loads to keep them all checking it every day to see if there's a load that matches them.

1:11:28and if they don't check it, they're useless to me. Now, and it was too expensive to call the truck driver and have a human call and talk to them, even if it's a human in the call center and the Philippines, whereas with AI, it's almost free. I'm calling thousands of them a day going, hey, this load looks like it's a good match for you or you interested and then we activate them on the platform. That's new work that wasn't gonna happen before not just a replacement. Yeah, I think that's interesting. I think like we have, you know, in the Valley, unfortunately, we've been co -opted a little bit somewhat but with a little bit of a doomer mindset in some areas and we think of then AI is okay, like it's all fixed pie, it's gonna replace things.

1:12:06And on the ground with large enterprises, the vast majority of the use cases are, it's the ability to finally review the contracts that we never got around to reviewing. It's the ability to finally automate an invoice process that we never did. It's the ability to go in and just create marketing campaigns in every language that we never got around to. And so, so I think that'll probably be actually like 90 % of the usage of AI in the future. We'll be things that if we look back and we snap the line right now, we said this is what knowledge work is today. 90 % of AI usage will be things that we don't do today.

1:12:3810 % will replace, you know, what we're doing in some areas. That's the right take because Chimoff, I can tell you in our firm, we would never have associates or researchers or analysts. Sacks, you're also wearing this line of work as well. venture capital. You would never have them review legal documents. That's something lawyers would do in legal department. But now, because of AI, we have them say, here's the safe, here's the term sheet, here's the edited version, dump it all in, find out what the changes are, what are the deltas here, and then let's have a discussion about what the founder changed in a standard document.

1:13:10And we don't have to bother with an attorney. And maybe you wouldn't have even checked those documents if you were, you know, a seed fund or an you would just go along for the ride because you're the 10th person signing the documents. So what do you think, Trimoff here in terms of the premise that maybe it's 10 % replacing work that's happening, but this is blue ocean and we're gonna do 90 % of like new stuff that we just never got to. Yeah, I tend to believe that's true. I think the customers that we sell into at 80, 90 are largely large enterprises as well, so not dissimilar to Aaron's customer base.

1:13:43What I would say is that what they are encountering is the trough of disillusionment. I don't know if you're seeing this as well, but every single CIO ran around signing up some sort of AI product in large part because their CEO would say to them, hey, what's your AI strategy? And the reason the CEO asked them that is that at some point somebody on the board said what are we doing about AI? So that's the the bullshed cascade that that we went through in the last two years and I think What has happened now is people have spent billions and billions of dollars. I think you can see it in the revenue traction of the AI companies But I think where we are today is that there are some real technical complexities that have not been solved I'll give you an example.

1:14:37We have a lot of customers in regulated industries, which is to say that if you make a mistake, you will get fined or you will get shut down. Life sciences, healthcare, financial services are three examples. People still don't seem to appreciate that when you replace software that is deterministic, with software that is probabilistic, meaning software that somebody wrote for you, Do A, then do B, then do C. With an LLM that can hallucinate, you'll have errors. So what used to be a throwaway thing, which is quality assurance in QA, right, unit testing, integration testing, is now the only thing that matters.

1:15:23Why? Because if you're a financial services institution and you're supposed to do KYC and it borx and now all of a sudden you send a wire somewhere in Syria, guess what? You're in trouble. No, no. If you're a healthcare company and you're supposed to do some clinical diagnosis to send out a drug on time, and you don't do that because the model hallucinates, that's a real problem. And I'm guaranteeing you, we have not seen the class action lawsuits that will come when those errors will eventually be made. They're guaranteed to be made. We just don't know the scope and the scale of them. So that's why I'm sort of of this posture where I think we've sold in a ton of promise.

1:16:07I think the reality is much more tactical. It's a little bit more banal. I think we're sorting through the exact use cases where you can put guardrails around these error rates where it's okay and tolerable. Like Ryan will probably tell you there's some number of phone calls that just sound totally he's okay with that because the broader thing is okay. And Aaron, so I don't know. So what I want to talk to my customers though, I have a calling truck drivers to offer them loads, but I'm not having to talk to my customer. Sorry, I made your truck drivers, but my point just is that I think agents are real, but I think that we are far away from that because we're still at the phase of how do you build reliable software in production for an enterprise versus the toy apps that you see on the internet, which is like, let me vibe code something.

1:17:00I think these things are worlds apart still. Okay, so let me get Saxon on here. Just to inform the audience, you heard a trowel of disillusionment. This comes from the hype cycle. This is something Gartner has been talked about for a long time. So in case you're taking it for granted and if you're watching, you have some sort of technology trigger like agents who have this peak of inflated expectations. Now we're in the trial of disillusionment. Hey, this stuff doesn't work. It's hallucinating. But we're kind of going up. It works. It just doesn't work. It doesn't say we're on this slope. It doesn't.

1:17:27It doesn't. It doesn't. I don't know where this is coming from. I don't even think we're at the peak yet. Oh, okay. So then you think we're still going up. Because I a lot of people at Shemaats Point were buying stuff and saying, hey, it doesn't work. You know, and now we're in the message. Well, let me say, let me say differently, socks. I think we have not yet figured out how to move the budgets it's from experimentation to mainline production, meaning where large chunks of the U .S. economy are comfortable enough with the ways in which hallucinations are managed such that they will replace legacy deterministic code with this new probabilistic model generated code, meaning model enabled code.

1:18:08Let's just put it that way. I'm sure though. I don't, where are we on this? I don't see the slope here. Look, I would separate change management issues, which are always going to be important, and there's always going to be big ones whenever there's a big disruption, especially enterprise, especially around compliance and legal and all that kind of stuff. I would separate that from the impact of the underlying technology trend. And I don't think the impact has come anywhere close to peaking yet. And in fact, I would say the rate of progress is exponential right now on at least three key dimensions.

1:18:39So number one is the algorithms themselves. The models are improving at a rate of, I don't know, three to four times a year. They're not just getting faster and better, but qualitatively they're different. Remember, we started with Purell and Chapats and we went to Reasoning models. And the difference there is with a Chapat, it's like kind of a smart PhD or college student giving you an answer off the top of their heads. The Reasoning models, it's more like the PhD saying, okay, let me go off and think about that. Let me do a project on that. And it could work for 30 seconds or a couple of minutes.

1:19:11I mean, as much computers you want to throw at it. And it will break down your complicated question into a bunch of subquestions, and then it'll try different approaches. And it can validate some of those approaches and come back to you with a much more impressive answer. And if you've been using the GROC3 deep research or the new ChatGPT03 to do these types of new reasoning models, it's pretty mind -blowing with their capable of. Have we even come close to figuring out how to tap the potential there, especially in an enterprise context? No, but my point is that the rate of progress on the algorithms is again three or four times.

1:19:47Let me just call out here. Okay, go finish him about six and then I'll take it and pass it. Well, I was trying to lay out the dimensions of which progress is proceeding exponentially. Okay, so one is the algorithms, okay, which is not just quantitative. is also qualitative. We didn't even get to the agents part of it yet, but that's the next big leap after reasoning models. We're just starting to scratch the surface there. Then you've got the chips. I mean, the chips are getting better at, I don't know, three to four X a year. We've gone from, you know, the H100 to the H200. Now we're on the GB200.

1:20:19It will be a GB300 soon. We'll be onto root three times better per year or they get better three times per year. No, no, no, they're getting the chips themselves depending on how you measure it. each generation of chips is probably three or four times better than the last. Okay. And Nvidia is back to rolling out new chip, new generation products roughly annually. And I'm just using them as one example, obviously there are other companies as well. So basically the lead from Hopper to Blackwell to... Got it. ...Rubin I guess will be in next year and then think five minutes coming after that. I mean, really an astounding way to progress.

1:20:53It's not just the individual chips are getting better. They're figuring out how to network them together, like with Enviel 72. It's like a rack system to create much better performance at the data center level. And that would be like the third area where you're seeing basically exponential progress. Just look at the number of GPUs are being deployed in data centers. So in Elon, first started training GROC, I think that had maybe 100 ,000 GPUs. Colossus was 100 ,000 correct. Right. Now they're up to 300 ,000. They're on the way to a million. Same thing with OpenAI's data center, Stargate. And within a couple of years, they'll be at, I don't know, 5 million GPUs, 10 million GPUs, So, and you see that power side, right?

1:21:31You're going from 100 megawatt data centers to 300 megawatts, we're just starting to now see the first gigawatt power data centers. I don't even think they're live yet, but this is where they're trying to get to. And I don't think it's beyond the real possibility that we could be at five or 10 gigawatt data centers in the next, I don't know, several years. So my point is just look, the algorithms, the chips, and the data centers are all improving or scaling at a rate of, I don't know, three to four X a year. that's 10x every two years, okay? Where people don't understand exponential progress is that if you're getting better at 10x every two years, that doesn't mean you'll be at 20x in four years.

1:22:08It means you'll be at 100x, 100x. So the models, the chips in the data centers will all be 100 times more powerful in four years. Let's say the end of this presidential term. So you multiply those things together, the algorithms, the chips, and then the raw compute that's available, you're talking about a million x increase. some of which will be captured in price reductions. Some of it will be in the performance ceiling and then some of it will just be in the overall amount of AI compute that's available to the economy. But the impact of this thing is gonna be absolutely massive and I think people still don't even appreciate that fact because they don't understand exponential progress.

1:22:46Yeah, and I think maybe just to square the circle though, because everything is that you just said to X is what I think is propelling the industry. And then the reality on Jamal's side, like just to connect the dots. So we have an eVal test that we do where we run enterprise data through every model to kind of figure out its accuracy rate and how much, it's not even hallucination, but just literally how much data does it miss when we ask for facts. The best model in the world, actually interestingly, we was Grock III on this particular test. We send it 500 documents and we ask for 40 data fields back from the documents.

1:23:22And so it has to get every single data field correct. And we only do a single pass. So we send the document to the model. We get a single pass back. Right now, the best score is about 90%. And so you can imagine a number of industries where you can't have 90 % accuracy, if you give something a question on 40 data fields. Now, there's ways to solve it. You rerun it multiple times, or you chunk up the document in the smaller parts. And so it doesn't get confused by the large context window. But a lot of the people that were deploying AI a year ago or a year and a half ago weren't doing that And so they they you know they did have a kind of a pilot run of something and it kind of worked okay And what they have to realize back to your point sacks is like this space is literally exponentially, you know changing And so if you don't use the the latest methods of okay, you have to actually like run the data through the model multiple times And you have to chunk up the data into smaller parts and you have to use a reasoning model and you have to make sure that your your prompt is like like hyper -tune for the particular use case, if you haven't done those four things, then you probably will actually end up with a project that fails.

1:24:26And even so, even when you do all those things for even harder problems, you're still gonna run into issues. So I think the challenge is that everybody's running a million miles an hour right now, and they're trying a lot of things, some work, some don't work, at the same time that the space is actually changing at a pretty crazy rate. And let's take a look at our partner Polymarket and which company they think will have the best AI model by the end of 2025 and get feedback from our panel on, if you think this is accurate and who you would pick here, looks like Google is in the lead here, 41 % of people believe that they will have the best AI model by the end of the question.

1:25:06I think the question is 25. What is the dimension? Like, you know, we use Gemini. So for many tasks at 8090, we use Gemini. It's incredible. But for most of our code gen, we use Anthropic and Cloud Kix Asks. It's exceptional. This is based on the best scores in the chat bot arena, which shows you a four -per -gap graphic. That is slightly different because people have gained those tests. That is a rub there. People are now building their AM model for the AVAL unit, right? All the models are way overfitted for these EVALs. But if you had to pick, who's your... I mean, so I guess, Shumoff, you're saying you have to take a test on my task.

1:25:41It's what? What's on task, I agree with you. What's on task? What's on task? technology we've that is compounding very quickly. All I was just trying to get across is that the error rates have been diminishing but not nearly as fast as you need for some sectors of the economy. So you can use a model to generate deterministic code. That's great. And as long as you unit test it and integration test it, it'll be fine. But I'm saying if you're going to use a model in production, in an environment where if stuff goes wrong, there are consequences where Health care, we're not there yet. And you don't want to use this for healthcare yet, but you could use it for writing or writing jokes or maybe 10 or five.

1:26:20But that's too binary. It's already used right now in healthcare, but it's just the doctor's meeting notes that would normally take 30 minutes to go and train. Yeah, so you can't be too black and white on that one. Yeah. What's happening right now, the reason why the progress is so rapid in coding assistance. And I think, you know, you're right, that enthropic with, was a cloud 3 .7. Yeah, I think they're the leader. And in fact, I think the Manus demo that we showed, it's not entirely a wrapper on cloud because they actually, they do a number of different things, but I think they are significantly using enthropic for the code assistant part of it.

1:26:57In any event, the reason why the progress is so rapid with coding is because code compiles and you can determine objectively whether it works or not, you can validate it. So that makes it a perfect area for AI to get better at through reinforcement learning and test time compute is AI tries a bunch of things, it sees what works, it sees what compiles, sees what the user then accepts, and then it's able to learn to iterate based on that. That's why coding right now is really the big breakthrough application in use case. But it's not going to be the only one. I mean math is another good area where I think AI is improving rapidly again because math you have proofs and you can look at the results and see if it validates.

1:27:40Now I think one of the big questions in terms of AI progress is how extensible is the progress to other areas that don't easily validate that way. So for example, legal work is I think a really good area for AI, but how do you validate that is correct? You would have to go to the court, right? the court is the compiler, like a loss of music compiler, or maybe the laws. Well, you could hire like a thousand lawyers or experts in area to basically do, you know, reinforcement learning through human, through, are doing human, but it's not like a compiler to your point. It's not like a man. It's not going to be a progress.

1:28:13It's going to be as rapid because it's harder to validate, but absolutely. But my guess is that once they figure out how to nail coding math and the things that are easily validated, they can move to the things that are harder to validate. But I think this is one of the big questions is whether I think people just kind of assume that AI progress will be, you know, equally fast in all areas. And I think it's possible that AI gets really good in some areas better than human, but it's sort of childlike in other areas. And that's a possible outcome. It's like, make your point, right? Like with a finite answer or an answer we know is the definition of answer.

1:28:49Well, this is the important thing about the agent kind of, you know, let's just say framework or architecture, momentum was instead of just saying, okay, we're going to do a single pass through the model and then whatever it comes back with is we're going to be satisfied. The legal work might be reviewed by another agent whose job is to review legal work. And so we can just throw more and more compute at the problem and we're just early in figuring out how to architect those. For multiple models, right? You could have a trial check. You would have You have a traffic check. And then GPT check Gemini.

1:29:24When you have some anomaly, it spits back out to the user. So human and loop still matters in this in type of process. In the early days of OCR, you would have a computer say, here's the characters in this legal document that you'd have two humans type it in and then you would get a certain level of certainty. You'll quickly find that when you layer these models on top of each other, the test time compute cost are astronomical. Yep. And Aaron's probably dealt with this. It's like I get a bill from AWS and it's like, oh wait, hold on a second, I just, you know, per 100 ,000 this month, what's going on?

1:29:56So I got to get to the point. Well, that by the way is another major friend line, which is that the new applications that we talked about are all much more token intensive. So we went from basic LLMs, totally. You know, which don't require that many tokens to give you an answer to the reasoning model where you can spend a thousand times more tokens just getting one answer to a question. and now the agents are going to be even more token intensive than that. So the amount of compute required to serve all these new applications is going to be massive, which is why I think the CapEx build out actually makes sense.

1:30:26When you do a deep research to your point, David, you're firing off maybe 200 queries and it's asking them, the AI is saying, hey, what query should I ask on behalf of the user and then you go down that rabbit hole? It's basically I'm doing 200 of them at once. Ryan, your thoughts here on AI first companies and agentic computing? Well, the one that really wanted to tie back to us actually our earlier conversation on tariffs and there's a real use case of a LL M's is how do you classify a product and you'd like to get to what we see today when we did our first Machine learning based natural language classification of a product you take a product URL listing page Just Shopify page or Amazon page and say hey, what what classification code is this what duty is owed?

1:31:06Six years ago in a hackathon. We got to like 70 % accuracy we're now in high 90s accuracy versus what a human trained expert will get to. But you actually get to it, which is not good enough. You're wrong, 3 % of the time. You might have committed a violation of the law for sure. But actually, what is truth in that regard? There's a lot of gray area in this. And truth ultimately is what does custom say? What does the CBP determine is correct? And those guys are using software that's a very simple algorithm. them in its decision tree that's going, okay, is it a shoe? Yes, is the top made of leather?

1:31:44Yes, you know, is the bottom made of rubber and they just go through a very simple and that outputs. And so on some level, if you convince the government to use your LLM, it becomes true, whether it's true or not. I think there's going to be some interesting cases like that that we haven't really thought through. And like, when does the government adopt these to be the source of truth? Okay, all of this speaks to this thing that's going to sound totally esoteric, but like, We all used to shit on QA, right? Like the least talented engineers were allocated to QA. I think in the world of AI, it'll end up being the most talented.

1:32:18You know, we internally at 80 -90, we call it improvement engineering and it's a total specialty. It's similar to when I kind of coined the growth team at Facebook. I feel it's the same kind of moment. That's right. Improvement engineering is really the skill that translates toy apps and vibe coding into something that's very practical and real. And my team and the leader of this team, he's like steeped in things like Japanese kata management from Toyota and quality systems. And these are all the things that matter when you're trying to just shrink the air rate down to zero so that you can use it in a reliable way.

1:32:56And also to document it so that if people want to questions what happened or have recompense or some way to come back and say, I hate that really harmed me. How do you even do that? Like these are all very complicated issues that will get sorted out. Super, I think interesting. Okay, four. I got to wrap guys. I got to catch flight to Miami. Let me do a closing here. If you want to keep going, you're welcome to three. Two. Well, the plane just waits, just text the pilot and just tell them you're, all right, listen, I'm not burning all the all in. Just give crannics. So to speak and all of our focus.

1:33:30I'm kidding. I'm kidding. I'm kidding. I'm not like free. You're worried about seeing a flight. It's everything and then putting it on the all -in budget and the rest of us are flying Southwest for your chairman dictator, Jamal Foyer. And I miss a flight. It's a strange concept. Yeah, David Syne's I had what does that mean? David was the last time you flew commercial. Clinton. I haven't missed a flight in about 15 years. For Ryan Peterson from Flexport. Iron Levy from the amazing box. Jamal Foyer, Patia, David Sacks, your chairman dictators are. I am the world's greatest moderator, lovey boys

1:34:09Brain man David

1:34:36I'm going to be a We should all just get a room and just have one big huge or because they're all It's like this like sexual tension that we just need to release that out What your the beat what your

1:34:52I'm doing

1:35:01Always