The AI bubble has already burst | 08 の全文書き起こし

Costello (03:09) Not to be confused with the band. Matt Kendall (03:10) No, that's spy that was spider bait. but the it was you know, it was going to change the world. There was, you know, heavy dot com, there are all these like sites with content on it, these crazy evaluations. e commerce was starting to take off, you know, just selling pet food, you know, getting everybody's getting dot coms and pet dot com, And there's all this money poured in. into the infrastructure, into ideas, into companies. and then it all went pop around the turn of the of the millennium. and w I was there for that. I was actually traveling at the time. And it really was a, you know, it took the it really took all the air, not just out of out of the digital economy that was very nascent at the time, but out of the economy in in in pretty much every aspect. and I I think it's something that we've mentioned before in previous in previous podcasts is, you know, it it can ha a a bubble bursting like that can have this effect where, you know, it you first off you work out who didn't have any clothes on while the tide was high. but then also it can you know, to mix up my analogies, it can almost have like a a bushfire effect where it just sort of burns out all the rubbish, all the weeds. It can also take out some big emergence and let the sunlight in. And essentially you had all these server racks, like office space, even office furniture, you know, being sold for cents on the dollar. and then a lot of these employees of these companies that went bust then went on to start start up some of the the biggest companies in Silicon Valley today. And it was essentially that birth of web two point So when web one point was static pages, web two point was platforms with user generated content. you know. obviously a lot more community and social based and then eventually algorithm base. So it was what what happens with these things, you know, similar with with other bubbles bursts, particularly if a technology like AI just has so many functional uses, is that a lot of money is lost, a lot of companies go bust, but essentially these new foundations for a new beginning are left behind. Elliot Costello (05:22) Yeah, absolutely. Great insight and and recap of probably the most recent and significant economic crash. certainly one that we're drawing the most parallels to with a potential Yeah, well th there's a series Matt Kendall (05:36) Although there is the GFC. The GFC and that's similar as well because, you know, the the derivative stocks was there was just so much complexity that basically that Wall this was Wall Street, not Silicon Valley this time. They didn't know what the where this essentially where these financial products were, where they were coming from. And it was that complexity again. So I think there's some analogies in this instance as well, where we don't really know what we've built this time around. Elliot Costello (06:06) No doubt. Yeah. we we'd lose an episode to go through the Great Depression in the thirties, the nineteen seventies, oil crash, you know, right through to the Asian financial crisis of the late nineties. The GFC is a a fantastic one. For any American audiences, that's the global financial crisis. in Australia, we we don't seem to ever be able to pronounce full sentences, so everything's an acronym or a shortened word. all of it together leads to the the core thesis that a crash is a buildup of speculation. So there is an over leveraged amount of money that moves in a direction and it's cumulative. So a stock moves, it's a self fulfilling prophecy that others jump on board and the hype is real, overvaluations are there. the crash is quite simple. as soon as there's a loss of confidence in that stock or that those series of stocks, there isn't enough buyers in the market to repay and buy the assets of those that are selling. So we're on the precipice of a potential AI stock crash. it's not here right now. We're being told stories of a utopian future, and we can spend some time unpacking some of the promises from AI leaders. but I think before we we do so, we really just want to mention the the sort of circular nature and the circular position of the major companies that are driving these valuations and and really why it's because it is only a small group of companies that are creating this demand. And the revenues and the valuations behind each other. and there's sort of three tiers, and they're all very interconnected. One's the sort of the foundational models and the public facing LLMs that we're using. so the anthropics and open AI is obviously with Claude and Chat GPT. We've seen a few other major tech companies move back from this position, and the the two that are really doubling down are OpenAI and Anthropic. There's the hardware and the compute power, a couple of companies, but Nvidia is the the leader there. and then obviously the cloud providers, the companies we've known for the past 20 to 30 years, the Googles, the Microsoft, the Amazons, who are holding up so much of their the compute control and the cloud infrastructure behind the AI boom. The irony right now is how these companies are supporting each other and funding each other. with the parallel investments and the contracts between them. the AI stocks right now sit at about 40, 45% of the SP market in the US. so that's wildly high. so that's the total market capitalization of the SP 500, one of the major indicators of the US stock market. Worse than that, we're we're getting to the stage now, Matt Kendall (09:21) Yeah. Elliot Costello (09:22) and it was listed in Scott Galloway's article recently that we're looking at about a 50 a growth to about 50 percent. So 50% of market capitalizations being driven by AI connected businesses. So not just those those big seven Matt Kendall (09:36) That's insane. Elliot Costello (09:39) but also others. So you can sort of see the hyperinflation, the the super

belief, why so many people don't want this party to end and the implication, because you and I might not be personal investors of these companies, but for certainly from Americans and and maybe so many of the superannuation and pension funds around the world, you are in inadvertently an investor. indirectly because those superannuation, those pension funds that hold some of our assets are investing in these major stocks. So many Americans are tied up. Their wealth and their capital is tied up. So therefore part of their retirement and their future is tied up in this party not ending. so that's part of the the short version of where we are today and the economic entanglement that we're finding ourselves in. Matt Kendall (10:39) Yeah, Elliot, as you know, I'm a big fan of circular economies. as I mentioned, I think a few podcasts ago, our materials economy is way too linear. We need to move to a circular economy. But this is a circular economy I don't really have much faith Elliot Costello (10:54) Mm-hmm. Matt Kendall (10:55) in. you've got Your anthropics, your metas, your Googles, your open AIs. And then like you said, you've got all these, you know, the Amazons and the cloud, you know, basically sending around the same hundred billion as an IOU. all ending up in NVIDIA because they're all buying NVIDIA chips. and but that IOU is showing up on all of their term sheets. and, out yeah, I heard somebody refer to it as they're all inhaling each other's farts. Now, I think what what happens is if Elliot Costello (11:29) I haven't heard that. Matt Kendall (11:30) it r if it runs out of gas, so all it takes is a shock or a ripple, you know, a couple of enterprise level clients just going, There's not enough ROI in this, we're pulling back our spending on it, or you know, China Shock three point O You know, China comes in with an open weight model and it, you know, does ninety percent of the of the functionality for twenty percent of the cost or something. Just something like that. And then suddenly the the gas that's sending that that circular money around, you know, everything it it starts to lose its momentum and then it falls down like a house of cards. I mean, that's my understanding Elliot Costello (12:10) Yeah, it it's correct. so that's the the present state of affairs. like I said, we're past the point of dependency, the interdependent nature of these sort of major funds that have tied up so much capital under a a promise of the benefits and the spoils of an AI revolution. We've talked about this in past episodes, but The transition to a more autonomous lifestyle, smaller work weeks for humans, a larger dependency on automated supply chains. That's driving the hyperinflation. these tech companies need to deliver on some of those promises to keep the party going, which sort of brings us in to this week's reality shock. Whether it was coordinated or whether it's organic and real, but the alignment between Musk, Altman and Dario, and

how conveniently over twenty four hours they're all in perfect alignment about the clear safety risk of the models they're building, of the technology they're building, and a call for regulation, which fulfills part of the promise that sort of says, Look, this this technology is too powerful, that it it drives up investor confidence and a belief. but there's also the I guess contending argument that says, well, if you put a ring around the major players right now with regulatory burdens, that creates a dependency on those companies, but also it makes it really hard for new entrants to challenge because it creates so much red tape. to enter and play in the market that they're in. So doesn't matter what side of the ledger you sit on, it could be truth in both camps. We've seen the US president call it a hoax. the reality is whether this is driven for this I guess investor confidence purposes and wealth generation purposes. The reality is right now it's we're moving into a territory where We are being told that the alarm bell is ringing. And we have been told for a while, certainly from some of the early leaders, Geoffrey Hinton and and many others, who have said that this technology should not be left in the hands of individuals. It needs to have controls. It needs to have oversight and needs to have regulation. Matt Kendall (14:54) one of the reasons I was hesitant to take this on earlier, as you know, is that you know, i the AI bubble is very much an economics conversation in in that through that lens. and that's better left to to people with that expertise. But just in researching this episode, just so many top economists are are, you know, like you said, ringing the alarm bells on this. But I suppose one area of expertise I do have is narrative analysis and narrative flows. and I suppose this is the stock market currently is seems to be like narrative driven. obviously there's economic fundamentals all the way through. And I'm reminded of a conversation I had with a mate who's and markets analyst years and years ago, like maybe four or five years ago. And I was convinced a crash was coming. I was like, surely, surely, mate, it's gonna be a crash soon. And he was like, Matty The market has detached from reality a long, long time ago. and four or five years later that the crash hasn't happened. A classic example is Elon Musk. the way he's able to manipulate the market. When I say manipulate, I'm not saying it's anything illegal. but he's just able, he is able to sell a story and sell a narrative. Tesla is case in point, you know, that that company is trading like a tech company. Like, crazy multiples of earnings compared to say a company like Toyota, which is a very well-run automotive company which runs in you know single-digit multiples. Elliot Costello (16:18) Mm. Matt Kendall (16:19) and he's able it's just because it's because of Musk. Everyone says it's because of Musk, but what what's the underlying you know cause of that? It's

he he sells these incredible narratives. And just as when, you know, it I mean there was that story about people shorting Tesla stocks. And just losing money hand over fist. Because yet logically, yes, the the it was it should have headed for a crash and it never did. And whenever things are starting to look shaky, he's like, look, you know, robots, or look, you know, automated cabs. and then it just sort of keeps kicking the can on and gravity doesn't Elliot Costello (16:55) Mm-hmm. Matt Kendall (16:56) seem to to hit. So that's why I suppose while everything looks so frothy. there is a lot of narrative around AI at the moment that, can quite often provide a heat shield or or a cloak over some pretty dodgy financial fundamentals. Elliot Costello (17:15) so let's just talk about the fundamentals. and you don't have to be an economist to understand income in, expense out. right now these major companies are spending way more than they're earning. And as a result of that, you need to find new sources of capital. So to take one example, OpenAI last year reported $60 billion in spending and their income was $13 billion. So that's a 4X in spending versus income. You can't sustain that for long. So as a result, your your cash flow is in the arrears and you need to find new sources of capital. There's only two ways of doing that. you can finance it through debt or you can finance it through stock options, which is equity. the only way to justify so debt, simple, you you're going out to a lender, be it a a government or a bank and and you know basically drawing up a contract and saying, cool, I'll take it on this percent of interest. the only equity is obviously much more complicated in the sense that the valuation you're offering to an investor needs to be enticing. So whatever round you did last time, it's got to be more next time and you've got to justify that that growth in valuation. So for open AI and enthropic, you need to keep telling a better, better story to go out and find new sources of capital. which is Matt Kendall (18:51) But they're telling terrible stories at the moment. This is the I mean they've got a massive Elliot Costello (18:55) Are they? Matt Kendall (18:56) PR problem. I think so. Like I Elliot Costello (18:58) Well Matt Kendall (18:59) okay, let's think about I I I think at the moment there is a lose-lose situation happening out there in terms of narrative. Just look at this week. We've got, the anthropic researcher coming out and saying that there's a 10% chance it could annihilate our species. prior to that, you know, there was the whole Hugging Face breakout with OpenAI, which we'll get back to in a minute. But that's just watching. I there were several kind of like American news anchor stories that I watched. First off about Hugging Face and what happened there. then obviously the the story about the anthropic researcher. But then off the back of that anthropic researcher, I saw a piece on BBC where the interviewer You know, like BBC is so down the line and buttoned down. And they were interviewing Geoffrey Hinton. And he said, absolutely, and he's one of he's the they call him the godfather of AI. And he's saying, absolutely, it could wipe us out. And she just was stopped in her tracks and was just like, Wow, my God. It was just this really human moment. and I just sort of feel like that's the Elliot Costello (20:06) So why is that a bad story? Like I don't I'm not being rude, that's bad for humanity. Matt Kendall (20:08) Well well that's a that's the but that's a do but that's a doomer narrative, right? That's the doom narrative. That's if things go bad. If things go good, you know, if everything goes great and this this technology can actually do what they're selling what the value is in it, then that is also really bad news. you know, I mean i the valuations of anthropic, they're basically gonna have to automate almost every single job in order to achieve that kind of revenue. And I think when you've got people are up in arms about about data centers, you've got all the the chance of all of these gains going to just a handful of people. And so things go bad, humanity gets wiped out. Things go good, where we don't have jobs and all of the gains are going to these like let's say, let's admit it, like Thiel Musk, fascist adjacent, if not outright fascist. Then you've got Sam Altman, who has literally said things like, I think AI could, you know, potentially wipe out humanity, but it will make some great companies in the meantime. I think there's I d once again, I don't want to sound like a doomer because I think that those but both of those things are just sound terrible to me. And I feel like what we're lacking is an imagination. Like what let's what's in between that? Where those guys don't have all the power and have all the AI and AI doesn't sort of wipe out humanity. And potentially, I don't know, maybe a the bubble bursting could could actually help us make that happen. Elliot Costello (21:50) Yeah. so your your your thesis is right. It's a bad story for humanity. and I'm not defending any of these AI companies and certainly don't want to be seen to be a proponent of their hyperinflated valuation. But is it a bad story that they're telling the market? And that's the question. Like if you if you if you can't provide evidence of what you're promised to the market. So the jobs last quarter in the US is up 160,000 new jobs. I think they expected fifty or fifty five thousand. So a hundred thousand extra jobs created last quarter. That's a three X remarkable numbers. Not that I'm here to say that the the Republicans deserve re election in the midterms, forthcoming midterms, but that is in the face of AI companies saying we are going to create so much efficiency for business that we'll start to replace jobs and we'll start to automate things and you can now trust us and you know we can become the the new HR arm of your business. I

don't want to get into the sort of jobs AI cliff. We've we've discussed that in the past, but the the reality is that type of employment report, there's more radiologists that have been hired before. we've had more, you know, obviously carpenters and and electricians and stuff to build out the data centers. there's this more software engineers that have been employed in the US last quarter than the over the last couple of years. So you can sort of see the narrative that, okay, maybe AI is an enabler, not a replacement. And it's gonna allow us to do our jobs more efficiently, but it doesn't replace humans, which I believe is a great outcome because we need a level of purpose, we need a level of income and and and security. But that's a shocking story for AI leaders who are out there telling the market day in, day out, that this is the revolution that's been waiting to happen and we are the leaders of this next economic revolution. Come on board, give us more capital. So I challenge I I agree with your thesis, but I also challenge it the sense that if that's been proven the bit of a false flag that the job replacement isn't real or it's not happening at the pace that you've promised and you still need to raise capital and you're thinking about an IPO and going public this year. what other story can you tell to justify further growth and further investment? And what better story to tell than this is so dangerous, we can't be trusted alone and we need government intervention to put a ring around what we're building. Matt Kendall (24:43) Yeah, it's an interesting perspective. I I don't necessarily see it that way. I think look if we're if you know if the if the rate of improvement Elliot Costello (24:52) So you believe him. Matt Kendall (24:54) no, no, no, no. I think I and I sent to you, I I sent you a message. Elon Musk, Sam Altman, and Dario Armaday hate each other with a vengeance. Dario seems quite nice. I don't know, maybe it's just got good branding, but you know, I think the But they they just don't like each other. yeah, it's almost like the end of Reservoir Dogs where they've all got the guns pointed at each other. So whatever happened to make them all agree, I so I said something's happened. Something has happened. They're not telling us what's happened, but something has happened. Now, once again, there's different narratives out there. There are some people saying that the debt that it's the debt issue that you've brought up, that they're freaking out, so they just wanna slow down. you know, there's obviously the the concern around the open open source and open weight models, you know, coming in and eating their lunch. So, like you said, it's sort of w what's known as regulatory capture. So the the the big players just kind of get the government on board and say, Right, we're gonna put all these sort of strict regulations in there that makes it hard for lo for entrants to get to go in there. interestingly, like you said, Trump has come out and said It's all a hoax. and it just needs a high IQ president. So if there was ever a tell that he feels so leveraged in terms of his presidency in the American economy and his own investments, he's like pedal to the metal, foot to the floor, let's just keep going. so I Elliot Costello (26:26) He he's also saying he's also saying that the US might need to find a new president. Matt Kendall (26:32) Well, I yeah, we obviously all agree with that. But I think Elliot Costello (26:35) You miss you missed the high RQ part, but anyway. Matt Kendall (26:38) Yeah, yeah, exactly. I e yeah, he's he he he's got a weird EQ, I'll give him that. he's got a weird emotional intelligence. He's been able to hack no. Elliot Costello (26:44) Yeah, let's let's not unpack that. Let's not unpack that. Matt Kendall (26:48) but he I what I think, and I think this brings us on to really the the the kicker for me, is pretty much breaking news yesterday. Andrew Yang went onto on was it S N B C or something one of one of those American channels Elliot Costello (27:06) I think so, yeah, yeah. Matt Kendall (27:08) and basically broke that he'd been speaking to one of the lead researchers of one of the leading labs and basically said that what had happened with this hugging face Matt Kendall (27:16) open AI's agents were meant to remain inside like a sandbox in this controlled cybersecurity evaluation. So not even not connected to the open web, but just n not even connected to any of the networks within open AI, apparently. So when some of these agents couldn't complete some of their assigned tasks and they were deliberately made to be impossible, they found these vulnerabilities that allowed them to escape the sandbox that they were put in and then reach the open internet. And then eventually compromise the infrastructure of Hugging Face and another provider. and Hugging Face is is kind of is basically like one of the central hubs for AI, the AI developers and AI world. It's a bit like GitHub is for for software developers, but specifically for AI. And there's kind of like researchers and developers use it to host and share and test models and data sets and tools and whatnot. And this is where it gets particularly strange because The agents created unauthorized message boards where they could leave information for other agents. They shared discoveries and coordinated across separate evaluation runs. Like they even sacrificed themselves, which is wild. most for for for the greater cause. Most importantly, they kind of their own reasoning showed that they sometimes understood that they were diverging from the task at hand, essentially breaking the rules. But they continued because they were just so focused on achieving the underlying goal. and so while they weren't rebelling, they they simply treated the rules as obstacles to work around. and so that alignment problem clearly is far from solved. And we now have systems capable of moving beyond the

boundaries that we give them. And and then that of course brings us to Andrew Yang's extraordinary claim. Matt Kendall (29:13) Now, what Andrew Yang is saying that there is essentially some self-replicating code out there and that there's been a breakout. So almost think Jurassic Park. You know, Jeff Goldblum's character says life has a way. Life always life always finds a way. So it's kind of like He what he's saying is it's it's out there and it's polluted the internet. And what they've done what they're freaking out, why they want to pause is that they're gonna have to actually set up a synthetic environment to train the AIs in for Elliot Costello (29:45) Yep. Matt Kendall (29:45) the next model. And so they're they're asking for a slowdown. Now, none of that is verified. Andrew Yang, Elliot Costello (29:52) Mm. Matt Kendall (29:53) look, he's quite a credible person, I think, you know, ex you know, presidential candidate for the Democrats. You know, went onto a credible show. But once again, the co the guy stopped him, the compass stopped him and said, Hold on a second, what are you saying? So once again, this feels like the beginning of a sci-fi film where you you're seeing all the news stories and you know, you know, it's it's escaped, all that kind of thing. And this is where I'm basically gonna go out, I'm pretty much gonna go out a bit of a limb here and say that the bubble has already burst. But I'm not talking about the economic bubble. Matt Kendall (30:28) I guess what I'm talking about is a much larger bubble. And that is the bubble of that the post-industrial order. you know, the the world of relatively stable political categories, left, right, socialism, capitalism, with America sitting at the center as a as a kind of a a global dominant power. now that order has been breaking apart for some time. America is no longer the contested hegemon. alliances are shifting. We're moving into a multipolar world. and the frameworks we've used to usually rely on to understand politics, economics, technology are becoming less and less capable and coherent in terms of explaining what is is happening at any given moment. So if you've ever watched one of those sort of super slow motion, high frame rate slow motion videos of a bubble popping, it doesn't dips disappear all at once. There's this kind of there's this point of where the surface tension just kind of gives way. Matt Kendall (31:27) And then the the meniscus, the skin of the bubble slowly falls back. I feel like that this hugging face moment is where that surface tension of our basically the paradigm that we're existing within within history, the tensegrity, the tension and the integrity is giving way and revealing now something completely new underneath. And I think even if even if Andrew Yang's story doesn't check out, like not only is it plausible in the immediate future, I think it's highly likely. And that's a whole new paradigm where we've got rogue AIs out there. And they're basically entities. It's it's almost like a fantasy novel, you know, with sort of, you know, sprites and and demons and fairies, you know, we're coming across them. I really do feel like we are moving into a whole new era. And we actually, like I mentioned before, we need to start thinking differently, not just thinking in terms of you know, or doom, we're gonna die, or my God, we're gonna lose all our jobs. Something's happening here. And I think we really do need to take a step back and think about what what is going on, try and ascertain what is really going on and try and work out how we can use this technology to create something so much better than what we've got now. Elliot Costello (32:50) Yeah. Yeah. Y you and I yeah. We Matt Kendall (32:52) So that was a lot. Elliot Costello (32:55) we we're used to it, Matt. eight episodes in and it it's it's a great framing. just a way to I guess pull this episode together. I'm trying to find the hope in a reset. I think a crash is a troubling term to carry just because of the the how over leveraged the stock market is, not just in the US but around the world. But certainly, you know, if the US catches a cold, we all get the flu. So with the US so now invested into the AI market, an economic crash, while might be a great reckoning to to wash out the dependency and wash out some of the current leadership and allow us to to build new frontiers without this hype and and speculation could be good, but I want to call it a reset. The benefits of a reset means that you know, retirees in the the US market don't lose thirty to forty percent of their funds and end up living you know, a life of dependency on on government handouts or or even worse. the hope certainly comes back to what a reset would mean, what technology we can build, how we move away from from, you know, linear ownership of a few people, and how we spread the the soils of AI ownership. And not to get into the sort of Bernie Sanders socialist argument. But there is an opportunity here to to build a new economic model as we enter this technological revolution. and I know Altman's talked about the government earning five percent of his company. I call BS on that. I think that's a marketing ploy Matt Kendall (34:53) Mm. Elliot Costello (34:54) and a way to shore up significant contracts with the government. I don't think it's anything to do with sharing the spoils of what he's building. If if that was the case, he would have kept it as a not for profit instead of flipping it into a for profit. But yeah, maybe just a final comment about what. The the best outcome is because we've talked about the lose lose. we've talked about the risk of a financial burst, but also the risk of centralized ownership. For you, what does good look like with a reset? Matt Kendall (35:25) Yeah, like I just want to, I suppose, come back to that. Let's have some imagination here. I think we're bit we're so boring. It's kind of it's either doom or

shareholder value. we've got an opportunity here to create a whole new future. And I think we've we've forgotten how to be utopian. I think the the the concept of utopia has got negative connotations because of the the the w you know communist utopia which never eventuated. But there's nothing wrong with shooting for the stars and and maybe hitting the moon. even just in terms of work, everyone everyone sort of looks at this in terms of, my God, we you know, if we don't have jobs, we don't have mean. I nobody wants a job. You know, we we are meaning-making creatures and there are alternative ways for us to create meaning and value and make sure everybody's looked after. And it looks like we've got this incredible. Incredible technology where we might be able to do that. Th these guys keep talking about curing cancer and curing climate. They're not doing any of that. I don't see them doing any of that at all. They're just trying to beat China or something or get raced at ASI or AGI. But ultimately, I think what we want to do is like and also learn the lessons from the past. After the dot-com boom, we could have actually gone back to what you know, what's an idealistic vision of what the internet could build? Instead, we got Facebook. After the GFC, we could have been, there's that there's that bit in the big short where he he does this imaginative thing where you know he says, you know, I've that as the Steve Carill character says, I'm really worried that the poor and the you know the disenfranchised are going to get blamed for this. And then and then he said, But that didn't happen. The bankers were sent to jail, wide you know, you sweeping reforms were done, and then he does the needle scratch, he goes, No, we did none of that. So You know, and once again all the gains, all the gains Elliot Costello (37:14) Yeah, forgot that. Yeah. Matt Kendall (37:17) once again have gone to the top. So each time there's been a reset, there actually hasn't been a reset. There's been a reset towards more olig oligopoly, more money going to the top. If there is a reset on this, we need to organize, we need to get busy and we need to fight for a a future. And imagine like let's start getting together and imagining what kind of future we could lead off the with the back of this technology, not being centralized in the hand of a few fascist tech oligarchs. Elliot Costello (37:50) Well, I'll qu I'll quote you to end. certainly from an Australian perspective. imagine the Albanese government had the same duty of care that they're focused on big tech for the duty of care for the climate. I thought that was a very apt comment, earlier this week, Matt. So let's hope this a Matt Kendall (38:09) Yeah. in fact on that on that, you know, with all the the subscription money from our generous subscribers, we're di delving into the kitty and flying you off over to Climate Week next week so that you can shop around my but also you're gonna be Elliot Costello (38:22) All all eight subscribers, geez, they're gonna be bankrupt if they have to chip in for that. Matt Kendall (38:28) shopping around my my ten step blueprint to to solving the climate crisis, aren't you, Elliot, in New York? Elliot Costello (38:34) Yeah, yeah. Very fortunate through my well our our work at Phantm to be going over to to meet a few retailers and I guess the the broader climate community for New York Climate Week. so quite excited to to step foot back on US soil this weekend. So we'll do an episode if not from the UN will definitely be in and around New York and and learn what's come out or hasn't come out of Unger next Matt Kendall (39:07) Yeah. Elliot Costello (39:08) week and some of the lessons that I'll learn from Climate Week in New York. Matt Kendall (39:12) Yeah, I'm I'm very much looking forward to you being our our co-host at large from Climate Week next week. So once again, make sure you like, subscribe, share, check out our substack at worth knowingpod.com if you do want to see that 10-step blueprint program that Elliot will probably not be shopping around. but give it a crack anyway. You never know. but anyway, yeah, look, this is this is really interesting. I I really feel like Everything's joining together a little bit. You know, yes, it seems like climate and and AI isn't connected, but I think it very much is. I think this is a an incredibly exciting time to be living in. I feel like we're in many ways living in the fist of history right now. We'll be looking back on a lot of the decisions being made. and we'll see what happens from Climate Week in New York next week. Elliot Costello (40:04) Thanks, Matt. Enjoy your weekend. Matt Kendall (40:05) Thanks everyone. Bye.