AI bubble: ‘It’s approaching vindication hour for me’ | Ed Zitron

[00:00:00]Generative AI does not have a business model. Like forget viable, forget sustainable. They don't have one. They don't. They have to pay their customers to use AI. If they'd have raised prices, I'd have been like, "Okay, this makes sense. People are going to be angry."

[00:00:15]But lowering them means people must be they must be being told by companies we're not planning to spend anything. It's less than 3 months and their customers are already taking hostages. And I think it's hilarious. I think it is genuinely it is the it is approaching the vindication hour for me today on the tech report. I'm joined by the writer of Where's Your Edet and the host of the Better Offline podcast, Ed Zitron. Thanks for coming on.

[00:00:48]>> Thanks for having me. Welcome back, Isaac.

[00:00:50]>> Thank you very much. Before we get into some reports that OpenAI and Anthropic are considering cutting their prices, we kind of have to talk a little bit about SpaceX's record-breaking IPO today. As of recording, they've kind of just launched that it's I think at just over $150 or it's going up slightly. And considering the the number of hail Marys that SpaceX stock prices requires to be come true like data centers in space for example, how much of this market reaction do you think is from the fear of missing out the sort of the hype around it?

[00:01:24]>> I think that's all that's driving SpaceX cuz SpaceX used to be a rocket company with an internet company attached like a satellite internet company. The only really profitable part of the company is Starink. Then they were like, let's add a really bad social network full of CSAM generating AI to it. And then they did that and now that's dragging down losses by billions and billions of dollars. So the answer is this is just exit liquidity for venture capitalists and Google who owns 50 I think 5% of shares or something. It's a genuinely dark day that this company was given the valuation it was. And Goldman Sachs and JP Morgan, the underwriters, should be ashamed of themselves. Everyone who works there should be ashamed of themselves. Anyone who touched this IPO should be because this is an unstable company, a kind of Frankenstein's monster of bad ideas attached to the occasional good one that is now allowing Elon Musk to become a trillionaire based on a company which claims to have a total addressable market of $50 trillion.

[00:02:26]Data centers in space are a pipe dream.

[00:02:28]No one has ever done them. Like because the I forget exactly the science of it.

[00:02:33]I'm not a scientist myself, but the people I've talked to about it have said that basically the amount of power it would take to actually get the data from space to Earth and also run GPUs in space and the massive cooling required would mean that it was some astronomical figure for like a single megawatt of data center. And I mean, it's funny, we're having trouble building them on Earth, yet people are people like, "Yes, put them in space. That will work." So yeah, I don't know what SpaceX does long term. I think it is going to I'm not a stock analyst. I don't really don't really follow the markets in that manner. Uh I will say I can't imagine it stays much above the 20 30% range today because why? Why? Other than I guess they could hope it becomes a meme stock, but Tesla only really worked as a meme stock because they sell cars. They sell a physical thing. SpaceX is far murkier, but I think he wants to merge them. I think that that's the ultimate plan. And it's horrible. It's all bad. Everyone involved should feel bad. Goldman Sachs, I think, said they were going to 300x their AI revenue by 2030. JP Morgan said similar things. Makes you wonder if we even have financial regulators anymore, whether they're just sitting at work kind of snoozing at their desk or smoking big cigars. It's a genuinely dark day. One thing I always wondered about data centers in space was how do you change out the GPUs?

[00:04:01]>> Well, you what happens is you get someone in a rocket which costs a ton of money for your low margin business. You get someone in a rocket. That person then goes up there. They change out the GPUs which will cost so much money. You you're spending countless amounts of money. And also the the other problem is okay, you have to get a rocket to go up there. What's one thing SpaceX rockets are known for? That's right. Exploding.

[00:04:26]just blowing up. So, do you I hope that that never happens again because if it does and you change out the GPUs, I don't know. It just makes me wonder how I I thought that there were regulations about companies not just making stuff up in the to mislead the public, but I guess that doesn't happen.

[00:04:44]It It's just very strange, very dark, and very sad.

[00:04:48]>> Before we move on, just quickly as a little side note, what what do you think of someone having a net worth of a trillion dollars? I don't think it should happen. I I don't think we need billionaires. I don't think we need trillionaires. I think that there is a certain level of when you look at someone with this much money, even though it's mostly on paper, you know, the amount of good they could do. I think what they work the the World Health Organization, I think it was, worked out, it's like $6 billion would solve world hunger. And he was like, "Yes, uh, send me a plan for that and I'll do it immediately." And he just ignored it and didn't do anything. He could just they he could fund every school in America like click of his fingers. It would be fine. But he doesn't. He doesn't have any. Musk has no responsibility to anyone. None of these billionaires do. And our first trillionaire has to be this guy, doesn't it? Has to be him. Can't be can't be some shadowy figure that we never hear from. It has to be a guy being like, "Oh, it's epic and based. It's epic and based if we're going to put data centers in space." It just makes me so it's so frustrating because the whole idea of startups and tech and this flows into everything else as well is that it's meant to be meritocratic. It's meant to be about the survival of the best, the smartest, the fittest. And I don't think you can say Musk is the smartest best or fittest in anything. Definitely not fittest. And I think that it kind of and they realize this is kind of a wider topic. I don't blame Gen Z for being a little nihilistic seeing this because getting a college degree doesn't mean that you're guaranteed a job anymore. Uh you look around and you look at the people who ostensively may like work the hardest and they don't make very much money. You see the people who spend 90% of their life on X the everything app being like the the the left are trying to they tried to steal my dog and you see that guy become rich.

[00:06:38]Yeah, I'd be nihilistic too growing up in this. I and it's it's very frustrating because Musk has actually tied very dangerous things to SpaceX. XAI is a really bad business, very capex and opexheavy. And even the funny thing is with Musk is that he is not actually very liquid. He has huge wealth on paper, but his actual liquid capital is very low and he has a bunch of margin loans on his Tesla stock and I imagine he will soon on his SpaceX stock. Thing is, SpaceX has attached a very bad AI company to itself. The costs are not coming down. They're renting out that people also are arguing, well, they're renting out the space to Anthropic as well, so that's going to help them. Anthropic, another unprofitable company that loses billions of dollars.

[00:07:29]At some point something is going to break. I don't know where it is. I I truly don't know when. I think he'll keep tap dancing for years, but I think that there is a point when SpaceX becomes Musk's albatross.

[00:07:40]>> So, seeing how the SpaceX IPO is going, how do you expect OpenAI and Anthropics IPO to to go, assuming they can make it through the financial security to get there in the first place?

[00:07:51]>> So, I I may eat my words here because I'm I'm just looking at the stock tracker as we talk. SpaceX appears to be up between like 22 25% on the day. This is a to quote Matt Lavine Bloomberg, it's the perfect range. 20% means that you got a lot out of it, but you didn't leave any money on the table. You couldn't have priced higher. But nevertheless, this is not a huge pop.

[00:08:16]This is not a giant offering. Maybe later in the day as we record this, it may change, but this is not like a huge one. So, there is some investor skepticism. There's clearly that there is there isn't so much it's going 30 40 50 60% it's not running in a crazy way.

[00:08:33]This makes me feel that any anthropic and open AI IPO and um remember neither of them have actually gone public. Yeah.

[00:08:40]None of them have actually publicly shared their numbers.

[00:08:43]I think it might actually show a little skepticism towards those IPOs as well because this is a good offering for Musk. This is not an amazing one. It's one of the largest raises of all time.

[00:08:56]But the money for this had to come from somewhere. Selling off other equities, for example. The money for OpenAI or Anthropic will also have to come from somewhere. So, it makes me think that the investor excitement is not quite there or if it is there, it's not sufficient to help turn this into just some insane offering, some 50% 60 70% offering. Not that those are common, but that's kind of what the AI bubble needed. They needed this to be bigger than ever. Still a success for Musk, which I hate to admit. I really don't like it. Makes me sad.

[00:09:32]>> Let's move on to the price cuts then.

[00:09:34]You've got a great free piece coming out on it next week, but just quickly, the Wall Street Journal says Open Air is considering massive price cuts to contend with Anthropic, who are reportedly looking at doing the same. I mean, can these companies afford to cut their prices or given the enterprise backlash over tokenbased billing, can they afford not to cut their prices?

[00:09:57]>> So, this is the funny thing. And when I heard about this story, I'll be honest, I'm surprised I didn't predict this, but that's because it's so stupid. So, Midwits, the recently led poisoned concussion sufferers, they will say, well, they're able to do this because the margins are so good on token.

[00:10:14]They're generating the tokens. We don't know that. We have no proof of that. We have no proof that inference is profitable. What I will say is that OpenAI planning drastic price cuts to compete with Anthropic who is allegedly planning as well journal Kegy and Burbagin reported that over at the journal. It's a funny thing. Um let me tell you with complete certainty nothing happened in the last two months that made their services cheaper. There was no breakthrough. We'd know. They would tell us. They would be like, "Wow, we had a huge breakthrough that massively reduced our cost." No, what happened was like two, three months ago, everyone got moved to tokenbased billing. And it's very clear because the only reason they're doing these discounts is because it's very clear that customer churn is happening. It's obvious that customers are burning through their budgets and going, "Yeah, I'm not doing that again.

[00:11:03]I'm not doing this again." What's insane about this is cutting their prices for OpenAI and Anthropic is likely to increase costs and lower revenues because let's assume that and this is also a big assumption. Let's assume that people use as many tokens as before unless the unless the price cuts are so drastic and that will increase token burn. But you have to understand that if they cut the prices by half and they double the amount of tokens spent, it's just going to massively burst their opex. It's not going to be like that the revenue is not going to make up for it because they already lost money. It's just a crazy thing that they're doing.

[00:11:43]To your point, yeah, they might not be able to afford to not drop prices. They may have it's less than 3 months and their customers are already taking hostages. And so I think that this is a sign that the underlying economics of AI do not work. The only reason they have to do price cuts and they did, just to be clear, they did it almost immediately after their users started paying the real costs, it's a sign the underlying economics don't work. It means that at scale, OpenAI and Anthropics customers have massive price sensitivity. And at scale, these companies can't grow with their current pricing. And if they lower the prices, the economics get so much worse. So very much worse. But then there's the other thing of let's say they cut their prices by half. Does that necessarily mean people are going to use as many tokens? Because the point I've been making again and again is that you can't really measure the price the cost of a task. You can't say if I do this with it will be linearly this much. It could change the hallucinations happen.

[00:12:43]A different harness like clawed code or open code might change it. Chat GPT prompts are processed differently what have you. So you can't really measure a task. You also can't really measure the return on investment of AI. It's been over. So, okay, now you can't measure how much a cost task uh task costs, pardon me, or indeed how much value you're getting from it. So, why does that even if you cut it by half, does that mean people are still going to be happy? Because think about it, you're spending millions of dollars. Okay, you cut that in half, you're still spending millions of dollars on something that you can't measure the ROI of. I don't think that OpenAI and Anthropic are doing this because they have a plan. I think they're doing it because they're desperate. And I think it's hilarious. I think it is genuinely it is the it is approaching the vindication hour for me because wow, it really took them so little time before they had to consider this. It it must be their customers must be churning brutally. Well, like you say, we we were talking about this three less than three months ago, and it's not it's taken less than three months to get to a point where like, well, we we've messed something up here. We got to give them discounts. And you say 50% as a as a as a >> completely random number, just to be clear, >> but would 50% even be enough to I mean, what I'm thinking is Uber blew through a year's worth of credits in a quarter of a year, so maybe they need to slash them by 75%. I was just wondering what kind of figure would actually kind of put customers at ease here.

[00:14:16]>> Yeah, that's a good point as well. Like Uber, let's say Uber but burned in 3 months their entire annual budget. Yeah, you'd have to cut it by three quarters to make the budget work. Cool. So you just reduced you basically made the same amount of money as anthropic and open AI while getting just as much burn. It's just it's they must have seen they even if they have not actually said they're doing this yet. This is just a reporting.

[00:14:45]But they must have seen something really bad. They must have had some customers turn around and put a gun to their head or something. Like the customers must be revoling over this one because this is I can't think of a major tech product that has ever cut its prices this fast. And they usually do so because of some sort of hardware breakthrough. We have no hardware breakthrough. I saw a genuinely insane business insider piece today saying, "Wow, the advent of Blackwell GPUs will bring down the costs." News flash, OpenAI has hundreds of thousands of those things and their their costs are still very high. Anthropic has access to them. Anthropic has a ton of TPUs. Like it's I think the AI boosters are going to try and claim that this is temporary.

[00:15:34]that they're going to, oh, it's a temporary price cut and then the prices are going to go up. What do you think happens? Let's say that that's true.

[00:15:42]They drop the prices, then a few months later they're like, "Yeah, they're expensive again." You think the customers are going to be like, "Oh, that's all right. Yeah, yeah, I'll I'll pay what I used to, the price that I hated before. I'll happily pay that now because of reasons." And so and they again they've not announced this so we can't fully do the touchdown dance yet but journal's very good at reporting but also things must be really dire for these companies to even consider this. I also don't if they cut token prices as well they'll do so for startups. So on top of the enterprise customers that are pay unless they just do the discount for enterprise customers which creates a whole other bucket of problems. Let's say you're a perplexity or a cursor and everything's suddenly cheaper. Those AI startups suddenly become a bit more economical. I don't think they'll become profitable, but they become more economical. Other than the fact that their customers will now burn more tokens, too, which will burn more for the AI startup and burn more for the AI lab. It's a genuinely it's one of those things that I had to do a double take when I saw it because it's just if they'd have raised prices, I'd have been like, "Okay, this makes sense. People are going to be angry." But lowering them means people must be they must be being told by companies we're not planning to spend anything. Now they could also do backdoor discounts. They that's possible as well. But the words of the journal were drastic price cuts which sounds drastic to me.

[00:17:15]>> You kind of said it earlier as well that cutting cutting the price would reduce revenue while increasing how much people use generally assumably.

[00:17:25]But it also in in researching for this, I was wondering if the very fact that people have now started looking for the invisible or non-existent return on investment of the AI adoption in their company, will that alone start to make these companies seriously question why they're bothering at all? Yeah, that's the I was kind of getting at this. Yeah, even if you discount it by 50%. You're still in a situation where people can't measure the ROI. Like it doesn't just because it's cheaper doesn't mean it's cheap. Just because it's less money doesn't mean it's value for money. And I don't think unless the discounts are 90% which will be economically unviable in a level that just kills the companies.

[00:18:11]I don't see how they square this circle.

[00:18:15]I think that it let's say we go back to the 50% number. Even then I'm not confident that that will actually make customers spend as much because their assumption I assume is that okay we discounted by 50% random number okay now they'll use it as much if not more and we'll make it up on volume making up on volume with a lo money losing product of course I think that there's a chance that people pay less and use it less which kind of does also does not help I also think the maybe I mean these companies not publicly listed yet Q3 must be a nightmare right now. They must be looking at the projections for the next quarter and saying, "Oh god, oh god, no." Because the way that Anthropic at least sells their sells to their enterprise clients is they sign up for a minimum spend without the year uh throughout the year even. So $10 million and they in many cases I've seen that they're charging more not less for tokens than the publicly available rates. I reckon what's been happening this just to guess I reckon people have hit those maximum those minimum spends and they're going no just like we're not like we're go we'll spend a bit more but we're capping everyone and we we don't and actually having a conversation with their account managers saying like hey this is not economically viable for us anymore. I mentioned this in the recent piece about AI slowing down but this is at the worst possible time for these companies the absolute worst. Because even if you're one of the ether sniffing people that believes AI is the biggest, most hugest thing ever, Anthropic and Open AAI need to make by 2029 upwards of $350 billion in annual revenue combined. They have $1.1 trillion in in compute commitments that they must keep up with. And the only way to keep up with those is to have massive revenue growth. Anthropic, they I think they're claiming they'll make uh $10 billion in Q2. That is not even close to enough. By 2030, these companies need to be making about10 billion dollars a month.

[00:20:22]>> It's the only way that they'll keep actually. Yeah. 2029 would have to be $10 billion a month.

[00:20:27]>> And they need How are they going to do that if they're discounting it? How are they going to do that with their customers threatening muting?

[00:20:32]>> But that um and I think that I think that this is just people people you go on Twitter and people are saying, "Oh, it's a strategy. Oh, they always plan this.

[00:20:45]or they have something efficient or they've got this. I need people to realize that the most likely answer is nobody has a plan.

[00:20:52]They I don't it's not oh they're going to get bailed out by the government.

[00:20:55]It's they never thought about it. They just we'll keep raising money and we'll keep spending money and then something will happen at some point. I think that that is the strategy. I think when they saw customers emailing them all caps saying I will kill you, they that's a joke. Um they said okay well we'll just lower prices. And the other problem is is even though they haven't announced it, this is now everywhere. How would you feel if you're signing up to spend money with either of them, if you saw a story saying price cuts are happening? I would personally take a step back and go, whoa, you reducing prices anytime soon? Can I can I get those reduced prices? And here's the thing. How do you sell against that? Because the mystique and the magic of AI has kind of gone.

[00:21:38]It's gone. I no one no one's excited about this anymore other than very strange people on X. Uh and so it's just a very strange time in the industry. I I genuinely didn't see price cuts coming.

[00:21:51]I thought that they'd do that much earlier. To do it at this stage is just desperation station.

[00:21:57]>> Given kind of what you just said and Cisco's G2 Patel recently saying that token costs are far higher than the actual value that they are producing at scale. How close do you think we are to enterprise contracts being cancelled or to your point being delayed until they're like maybe we can get a 75 50% discount if we just hold off on using AI for a few months. I think the next week or two will be very telling but I think we're already there.

[00:22:24]So Anthropic and Open AI are both and I'm not kidding giving out $1,000 worth of tokens to any any organization any business that uses Claude Code for the first time. Just $1,000 off the top.

[00:22:38]Check this out. What do you think?

[00:22:39]Here's some free money for some reason.

[00:22:41]I sure hope they don't book that as revenue because that would be illegal.

[00:22:45]But putting all that aside, they could actually sorry, they could book it as revenue then take it off as a sales and marketing cost. But man, is that man is that a dodgy business. Nevertheless, I think we must be already there or we're getting close to it because even the rumor even the rumor is it because remember the rumor is drastic price cuts from OpenAI in response to Anthropic doing the same thing. Even the rumor is the sign that they're freaking out that they're actually tr they may I don't know if it leaked. I don't believe this leaked. If I was Anthropic or OpenAI would rather die than having have this leaked. I think that this is just a result of desperation and I mean there was the the amount of they basically have to give away money to get people to pay them and you've seen this also with the we I I don't know if we want to talk about the subsidized rates as well but it's very clear that the underlying economics of AI don't work they never have. if they worked, they wouldn't have to do discounts the moment the people started paying token rates. If they if the economics made sense, they wouldn't have to subsidize anything. And I think that it's very vindicating and I'm laughing a great deal watching again and again is things that I've been saying for years come true. I know a lot of the listeners will be excited about this, too. We're all excited together. But it is, it's funny watching the entirety of Silicon Valley be like, "Huh, you're saying profit is revenue minus the cost of goods sold." Interesting. So, you want that to be a positive number. You don't want it to be a neg You don't want a negative profit, right? But what if my business only loses money? Oh, that's bad. H didn't really think of that one, but let me get back to you with a solution. And it's very funny like the G2 Patel quote straight up him just saying like yeah we they're way more way more expensive than the value like when you're Cisco.com bubble Cisco Cisco the most docomiest of them all is like yeah yeah mate this just doesn't work. What was really funny was I saw someone on Twitter who asked Claude Fable the new model how the business model would work for AI and it said yeah you need like a miracle of economics to work. It's not going to work though.

[00:25:00]It's another of these when you have your AI just being like, "Nah, mate. Nah, my company sucks, man. Yeah, it doesn't make any sense. They're going to run out of money." It I I'm paraphrasing here, but the quote was something like, "You would need an economic miracle that precisely times with accounting, which is just an insane it's insanely bad.

[00:25:21]It's not just like an economic miracle at some point. It has to happen very soon, you know, like I've been saying for months." So, yeah. I just I feel it's hard to say when this all ends because it's also insane, but I feel like we are you can hear they're playing the Oscars' music.

[00:25:37]They're telling them to get I would like them to wrap this up soon. It's It's silly for them to pretend.

[00:25:43]>> The the band on the Titanic comes to mind.

[00:25:45]>> Yeah. Yeah. Deck chairs everywhere. I do I do want to get into the subsidy stuff.

[00:25:49]And you again, you do go into this in your your free newsletter next week. I did some back of the back of the cigarette pack math off the top of my head as well. And just just for now, ChatGpt's $20 a month tier, you need 35 paying and completely idle users to subsidize just one person hitting their cap. And you need 20 uh sorry, that then doubles to one to 70 for the top tier.

[00:26:17]And similarly for anthropic you need 20 idle users on the lower tier again doubling to 40 on the top tier. I everything about we've been talking about enterprise customers on top of that and the tokenbased billing where where is the money going to come from to pay for functional business models is I guess the big >> well the answer is soft bank for now.

[00:26:41]Um, so I want to kind of simplify your example on the 200, sorry, the 200 buck a month chat GPT Pro subscription, you can burn up to 14,000 14,000 worth of tokens. But a really simple way of saying this, this came from Semi analysis, a big, they're AI boosters as well. They are very pro AI. This is not a hater.

[00:27:04]If a customer of any subscription tier, 20 buck, 100 buck, 200 buck a month, spend uses any more than 25% of their their you their rate limits within a month, they are unprofitable.

[00:27:21]They are they you get gross margins of at best negative 25%. They're called gross margins because they stink. And the funniest one is if somebody uses and this is by the way semi analysis is calculating this saying that the cost of tokens has a gross margin of 75%. It doesn't. I know I know this for a fact.

[00:27:42]You'll find out how soon. But anyway, even within this wacky mathematics, if someone uses 50% of their 200 buck a month chat GPT pro subscriptions utilization, they have a gross margin of negative 775%.

[00:28:00]Now, I ain't no I ain't no business genius or nothing, but this this is not a functional business. It's what I'm comes back to a to a one point which is the economics don't make sense.

[00:28:12]Generative AI does not have a business model. Like forget viable, forget sustainable. They don't have one. They don't. They have to pay their customers to use AI on the subscription level.

[00:28:25]They have to pay them. They have to just give them way more. And the reason they're do these are companies that are heavily growth hacked. They measure everything metrics. They measure how long people spend on the app. They measure how long people use the app for.

[00:28:37]And they're only doing this because they know that if they don't give them upwards, I think it's 400 bucks worth of tokens on the 20 buck a month, 2,000 on a 100 bucks, 8,000, you can 8 grand worth on a 200 buck a month Claude code subscription. They're doing this because it's very obvious that their customers will not pay on a monthly basis, will not pay 20 bucks or 200 bucks a month unless they get 20x in API tokens. And it's completely insane. And what's more insane is Anthropic signed with SpaceX to take their Colossus data center $1.25 $25 billion a month for one reason and this is based on their own blog to increase their rate limits to increase the rate limits to make them further unsustainable.

[00:29:28]AI has no business model. It's not getting cheaper. It's not getting more affordable for anyone. The only time they've ever made serious money is by beguiling business idiots. by basically doing a pay thing where they convince these companies to use AI as much as possible and just have these companies don't anthropic doesn't provide very deep um spend metrics. You have no real time way of seeing how much is being spent. They don't they obiscate the costs for their most expensive enterprise contracts and now those contracts are saying no I'm good like I I'm not spending this much. And the funny thing is is people are saying, well, they go to open- source models. I don't even know if that's the case at this point. I think that the other thing that these companies have not realized this, there's the obvious sticker shop problem. There's also cost fatigue.

[00:30:22]There's a certain degree of I can imagine there are some businesses that are saying, "Oh, we spent another 500 grand this month. Why are we doing this?

[00:30:29]I hate this." Like I could it the I talk to tech workers all the time. I hear from them daily all the time. And good lord are people tired of this. And I imagine even the dumbest executive who loves AI the most because they spend 11 hours a day on LinkedIn, even those people are probably like, I don't like the fact that I lose half a million or a million dollars a month in a way that's also borderline impossible to explain because there say you have 1500 engineers. Can you explain with any clarity why it costs that much? Andrew McDonald from Uber clear he said that it was hard to justify the cost because it's not like you can point to more useful features. It's not like you can really point to anything. In fact, the only thing you can point to is cost go up. So yeah, AI doesn't have a business model. It's they're doing price cuts because the previous one didn't work.

[00:31:24]The new one won't work. And at some point I think that they're just rushing to IPO so that Daario goddamn Amade can become a billionaire.

[00:31:33]and we have to hear I I want this era to end the most just so that I don't have to hear from him again. I'm I find him more noxious and annoying than Alman at this point.

[00:31:42]>> Well, on that note, Ed Zitram, thanks for taking the time.

[00:31:44]>> Thanks for having me.

[00:31:46]>> If you enjoyed today's episode and you want to hear more of the tech report, please consider liking and subscribing.

[00:31:51]Also, you can get episodes of the tech report wherever you get your podcasts.