Transcription
Dario Ammedday and Sam Orman are grifting the economy. Because if you say all of this chaos within SAS is caused by the glory of AI, you can cling to the dream that AI is just another way where hyperrowth can go. It's just the next thing that will grow everything. It it is exploiting the ignorance and desperation of a software industry on the decline and it will suck for our markets and our economy. But I think this era is necessary to wash out an era of rot.
On the tech report with me today is writer of Where's Your Edit and the host of the Better Offline podcast, Ezong. Thanks for coming on.
>> Hey, thanks for having me. So you gave me a bit of a sneak early peak at your newsletter about the sort of the state of the software industry right now and you make the point that the huge supposedly AI fueled software selloff that we saw uh the other month was actually a symptom of an industry that has been declining for a while now. And in your newsletter you go into all the details. So I'll just kick straight into it with so you you make the point that software as a service businesses between 2018 and 2020 were getting lots and lots of investment but now four years later you now are calling them something zombie SAS or zombie software as a service. Can you just explain what that means?
So yeah, uh Apollo Global's co-president John Zito made the point the other month that between 2018 and 2022, 30 to 40% of private equity deals were for software companies. Because software as a service is this business model where to simplify a bit, you pay us instead of pay paying for one thing up front in a box, you pay a monthly fee to get access to the software. Now about 2005 through 2018, these were the hottest thing in town because it was an incredibly obvious business bubble. If you invested in it, you get to a certain amount of revenue. You can say, "Oh, wow. They're growing 30% year-over-year. Fantastic." This led to a massive boom both in investment and in startups being created in the software business. Now the problem is after 2018 which is also the year where private equity and uh venture capital started slowing down as well. It stopped being quite as easy to sell these things to other people or take them public. So nevertheless private equity doesn't really know what it's doing. So they bought an absolute ton of these companies hundreds of billions of dollars of software companies pump them pumping them full of debt pumping themselves the private equity firms full of debt to buy them and then saying okay now the easy part. We'll just sell them to someone else for five far far more. The zombie SAS generation is this series of hundreds of billions of dollars of software companies sitting in the coffers of venture capitalists and private equity firms that are worth less than they were invested in at or leverage buy out or leverage bought even. So you've got all these software companies that can't really sell, aren't really going anywhere. And the wider software industry inside and outside of this is slowing down too because there's a limited amount of software you can actually sell and a limited amount of people you can sell it to.
>> What has changed in the last I mean if we take the end of that period 2022 to to now four years later what what what shifted to make what was like you say things that were creating 30% gains returns to now being something which is kind of dead dead and walking still.
So the sad thing is is that nothing actually changed other than the assumptions. The assumption was that every single software company would grow in perpetuity. The only thing that grows forever is cancer. Nevertheless, there was this assumption that all of these companies would just keep growing like gang busters because you had examples like Salesforce or Adobe even that were growing and growing and growing. But when you look at actually the charts of their growth, everything kind of slows down in 2022. There was this thing called the zero interestf free policy era after COVID and the great well great financial crisis than COVID. They eased interest rates in a very simple level. It's just made it easier to get debt which meant the venture capitalists had more money to invest and private equity firms had easier access to debt that they could buy software companies. Now sensible people might hear this and say yeah but didn't they check whether the companies were actually growing? And the answer is no. The business idiots that run these firms don't really check these things. The assumption is because something happened in the past, it will happen again and again and again in the future. Now, this wouldn't be a problem if these companies were growing to like 50 $100 million a year looking for sustainability. Problem is, venture capital and private equity don't give a damn about a company that's just growing nicely or just growing profitably. No, no, no. Everything has to be making 300, 400, a billion dollars a year or at least look like they're doing so. The problem is software doesn't grow forever and the assumption is that it was. In fact, all of the values of public software companies come from this idea that they will grow in perpetuity. When the AI software selloff happened, this was kind of it was kind of underhanded. The assumption was, oh, these companies are stumbling cuz the glory of AI when actually what was happening is they were stumbling cuz there's only so much you can grow. AI isn't replacing these companies. I don't even think AI is a threat. Look at Adobe for example. People say, "Wow, Adobee's being changed by AI." When you look beneath the surface, Adobe, a company that makes about $6 billion a quarter, made, and I'm not kidding, about well, I think it works out to about 30 something million a month from AI first software additions. It's nothing. And that's the thing. Everything about this is sold on a lie that AI is scaring software and AI is the reason that software is stumbling. No, software is stumbling because we're at the end of the hyperrowth era of software.
Is the this kind of constant hiking of subscription prices in exchange for sort of useless and often bloated updates. You call it sassflation.
>> Is this kind of the last move of these companies to try and keep up this pursuit of infinite growth? I mean, yes. And what comes next if when it presumably fails?
>> So, think of it like this. As a SAS company, you pretty much only have three options if you want to keep growing. Number one, make new things. No, no, no. We couldn't possibly. And that's me being somewhat sarcastic. There's only so many things you could do. Take Dropbox for example. Dropbox is a theoretically meant to be a place where you have files, you put them in there, you share with people. except now they've got signatures. Now they've got collaborations. Now they've got all of these other diddly features. Uh we don't like them. But what the SAS companies like to do is option two, which is increase the prices. And it doesn't mean the software is better. In fact, in many cases, the price just goes up without anything changing. Uh Microsoft loves to do this by being like, well, guess what? We're giving you AI now. Do you want AI? Well, you don't want AI, but you're going to take it. Take the AI now. So that's option two. Raise the prices. And it's happening across the world. It's happened every single year. I think SASflation, which is just last year, if you were paying a dollar, you're now paying a$120. It's about 12%. And it's happening every year because they don't really have any other way of growing. They don't want to say, "Well, we've kind of hit the wall of what we can do." And then there's option three, which is just buy another company and staple them onto the side. Salesforce is the best at this. They just recently said they had nearly $3 billion of annualized revenue in AI. Now 800 million, sorry, no, over a billion dollars of that came from a company called Informatica that they bought in November of last year. They bought Slack. They've bought all of these companies so that they can keep saying, "Look, look, we're growing." Are they growing? Not really. They just stapled another company onto the side. This is an era of desperation. AI was meant to be the thing that would would kickstart growth again within software. The problem is is that AI features are not that useful and everyone has the same ones cuz large language models are very limited in what they can do.
>> What about the the business idiot as you affectionately seem to call them the sort of the executives that will be buying this software or I suppose to to the point renting this software or subscribing to the software and then kind of forcing their users to use it anyway. Is that not going to kind of keep this is is that the the thing that's giving them undeath if we're going to extend the the zombie sass analogy?
>> I mean, yeah, to an extent. So, it's important to understand how software is bought and sold. You may think of it as a consumer as I'm I'm here with my credit card company. Here you go. My credit card, I get software. Enterprise software is sold in a very bizarre kind of Plato's cave situation where it's like a guy comes to your boss and says, "Do you want to get dinner? I will buy the dinner. I love you. You're the best. Wouldn't it be amazing if your company did this now?" And bought Monte Carlo. Do you know what Monte Carlo is? Do you care about data observability? No. But now that you've read on LinkedIn that data observability is a big deal and you heard it at a conference, you got to have Monte Carlo. This is quoting an excellent piece by a guy called Nick Sesh called brainwash an executive today. It is how software is sold. Billions of dollars of software is sold from someone that doesn't use it to someone that doesn't use it and then it's dumped onto the the business itself. AI features are the kind of extrapolation of that away where it's well now that's happening to the consumer. a a software company that doesn't really know what they're doing is scared and they hear AI and there's AI under the bed and there's AI in the newspaper. Well, I'll just add some AI features and then tell the pigs that pay me that they have to pay more because we've got AI now. And thus, it's the natural kind of undeath is to use your term of a software industry that stopped serving the customer and always served the business idiots up top. And it's frustrating because watching them do this is watching the end of a great generation because the idea, the magic of software is kind of cool. It was cool that you had software for everything. Software that could solve distinct problems except the people funding these companies are not funding software companies to solve problems. They're funding software companies to take over industries. And that's it is the ugliest form of rot economic capitalism. H well one of them I guess war is probably a lot uglier but nevertheless this is what happens when the software industry starts being about something other than making useful software. Do you think we're again not to not to go too much onto taking exact lessons from history and assuming the same thing will happen again, but we're seeing like you say with SAS there is this uh sort of the business model that is kind of fundamentally flawed of trying to achieve in well uh long-term forever growth if not infinite growth >> with a little smattering of debt to fuel it. And it sounds very reminiscent of what we're seeing in AI at the moment and and sort of I suppose do you think chat GPT Gemini those sorts of chatbots are headed in the same direction because they're largely also taking on stereounts of debt to sort of gamble on the idea that AI will grow to something which we cannot even imagine.
I think the AI bubble is a symptom of the grander collapse of software growth because all of the so another thing with all the debt within software there's a growth of this thing called venture debt. It's about tens of billions of dollars of this where it's exactly what it sounds like instead of buying instead of giving someone money and getting equity back you now are a venture capital firm that loans money. Good example one is called Hercules Capital. So they will give a loan to a startup. Now AI, the reason I say it's a symptom is AI has come along and like I said, it was this magical idea that wow, we have something we could plon onto our software that we can now use to sell it. The problem is is that they're trying to do the thing that SAS did with a completely different product. So just to simplify, the whole thing with a SAS company is you come up with, I don't know, a SAS for car washes and a venture capitalist gives you millions of dollars so that you can build up a sales team. the actual cost of the software is not super expensive. It's the servers, it's the getting the getting it built, the engineering talent, and then the sales team to sell it. The idea in theory was that you put all that money in, they grow a bit, you go, "Wow, they're growing. I'll give you some more money. They'll grow even more." You get more salespeople, you get more engineers, you grow more. And as you grow, your costs scale nicely. Like, it doesn't become every customer isn't super expensive. And as a result, you could charge a monthly fee for that software. The problem with AI is you do that except the cost for each user because of large language models and how they work is not reliable or replicable. And actually the more a user uses it, the more expensive that user becomes. This is not generally how SAS works. But because the tech industry does not know how to do new things, they created these products and they said we'll charge a monthly fee for them. SAS companies. Oh, we'll just charge another monthly fee on top. The problem is, well, other than the fact that the features aren't very useful, every customer is so expensive to you. As I mentioned previously on another episode, anthropic for their clawed code subscriptions, they're spending anywhere between $8 and $135 per dollar they make. This is the problem of AI. And because the tech industry is incapable of having new ideas and of adapting, they just tried to do the same thing again, shove them for the money. Except when you used to fund a SAS company in the second or third rounds, you were funding for growth. You were saying, "Okay, going to shovel the money into your mouth. You're going to get more sales people. You're going to build bigger." With an AI company, half or more of that money you're investing is just going into the tokens to open AI and anthropic. It's a lifeline. And yes, funding SAS companies was always a lifeline to an extent, but with AI companies, the ongoing costs of running models are so severe that you're effectively constantly having to well fight for your life and constantly having to raise money.
>> Well, I don't I don't want to go into the whole of your newsletter, so I won't spoil anymore, and people can there'll be a link in the description of this video if people want to go and have a read. But just just before we move on, what do you make of Sam Alman who came out saying that he thinks that AI will eventually be sold as a utility like water or electricity?
>> Yeah, there's a a term we'd use when I was growing up for people like that. It's a stupid I'm sorry to be sorry to be blunt, but that's just a dumb thing to say. That's not how utilities work. It shows a complete misunderstanding of everything. Tokens are not a utility. That would be like saying that I don't know, access to Amazon Web Services is like water. That's not how this works. Uh GPUs also are obsolete within 3 to 6 years. It just doesn't make sense. It's a self-serving statement made by a Made by a guy who's just saying stuff now because nobody has told him to shut up cuz no one on the stage with him has said, "Sam, what are you talking about? Sam, do you have a gas leak in your home? Sam, did you hit your head while walking on stage? Because what you just said sounds stupid. Humoring these things is I'm not saying that you are but the continual humoring of these business idiots of these people who truly do not know what they are talking about leads to things like the AI bubble and it they will keep happening as long as people on stage don't pull out like a little spray bottle like you'd spray a cat with when they get on top of something. It's I'm not even being sarcastic. Utility. What are you talking about? That's just him wish casting that the governments of the world will pay to run endless GPUs, which will not happen. There is no bailout coming. There is no hope for what Sam is talking about.
>> On that note, let's let's move on to Oracle, one of our our favorite topics.
>> Oh, yeah.
>> Earlier earlier this week, they beat their earnings report. But a little bit deeper, several US banks have pulled out of their future lending towards Oracle. And this is there been some rumors now that they're gonna have to cut up uh cut off or lay off rather up to about 3,000 uh 30,000 staff.
>> I mean banks pulling out cutting staff just to make uh I think I saw a figure which was maximum about $10 billion which we'll get into is not really enough to to to cover the difference. I mean is is Oracle in crisis? Is is there anything more that they could do which would be more of a symbol of crisis for that company right now?
So Oracle is a company in decline too and Oracle if you look at their graph of growth their Oracle's main business line when you put aside AI has been databases licensing and suing their customers and I'm not kidding suing their customers when they don't renew suing their customers if they don't pay for every license they need to really aggressive company but their software and licensing business has plateaued has for a long time or Oracle has then what they've done is said, "Well, what we'll do is we'll start up AI compute and we'll spend at this point over $100 billion on building AI data centers with uh in some cases negative 100% gross margins." Good stuff. So, Oracle's earnings were fascinating because oh, they beat on all of these numbers. Yeah, but they had negative $24 billion of cash burn. They claimed that for for this fiscal year they'd spend $50 billion in capex, which is going to be pretty bloody difficult because they've already spent 40 billion and they've got at least another quarter to go. Unclear what Oracle is going to do there. But the thing to remember is that Oracle is in this they're kind of in a kabayashimaru. They're they're in an unwinable situation. They have to build these data centers now. They have to because if they at any point go, "Uh, we're pulling back." Everyone on the street will go, "Wait a minute. you said that this was the what are you doing and they'll scream at them and throw rocks at them. So Oracle has to do this. The problem is is that Oracle does not have enough money to pay for this. They have to they've had to take on 50 60 70 billion worth of debt. They're going to have they did a share sale of $25 billion. They have to do all of these things to build data centers that are way behind schedule. Oracle screams till they're blue in the face that they are not behind. Except Stargate Abene, which is eight buildings, was meant to be done. Well, they were meant to be done at the beginning of the year and the middle of 2026. Now, the end of 2026, they're not going to be done by that. They're just not going to be done in time. And now OpenAI has chosen not to extend to the other two buildings. Now, reports you will hear will say, "Well, Microsoft's going to take it. Well, Meta's going to take it. Nvidia put down an $150 million deposit just to stop Crusoe from cutting the building entirely, the extension. But putting all that aside, Oracle is building six more buildings in Abalene. They've got two built full of Blackwell GPUs that by the time they turn them on will be two or three years obsolete for a customer open AI that is running out of money perpetually, that burns billions of dollars, loses billions of dollars a year. And when just assume that OpenAI can actually pay them, which they will not be able to, they will then um probably still lose money. They'll probably still be losing money on the deal. So Oracle is building 4 and a half gaws of data centers for an unprofitable startup that cannot afford them. Oracle cannot afford to build the data centers for themselves. So they're taking on a bunch of debt. And as you mentioned, banks are a little bit nervous about this debt, which is why we're yet to hear about the status of Stargate Michigan, which last year Blue Al pulled out of the loosest legs in BDC's and private equity. Now, we we haven't heard much about that recently, and people's assumption is, well, everything will be okay. I bloody well hope so because if it isn't okay, these projects are going to go into the toilet and get flushed aggressively. And Oracle's not really going to have any growth story because most of their growth has come from this near or totally unprofitable AI compute sold to Nvidia, Meta, Bite Dance for Tik Tok and Open AI. It's a disaster. It hits exactly that this imbalance that I think we've been seeing for a little while of these well a minimum build time of say a year probably more like two years for a data center to put in GPUs which have a threeyear shelf life and by that point you've got only a third of a GPU essentially what what you paid for anyway.
>> Do you know why Oracle is building essentially outofdate data centers with money they don't have and their customers like OpenAI like you say don't have.
So I have this theory called the beginning of history which is a Fukuyama joke. Not exactly but it's this idea that we need to stop thinking about and doing things in the terms of what worked before. Oracle is doing exactly that which is they are saying well in the past what worked for tech firms shoving a bunch of money and building a bunch of things then money came out right did it well it kind it feels nice to believe that so we're just going to do that so Oracle is just going through the motions they don't really have a choice but to build a data center full of obsolete GPUs this is the problem with the AI bubble data centers don't take a year to build. They take years and you can't say, "Well, I'll build for the future generation because Nvidia changes it up every year, year or two." Right now, they're currently with the first generation of their next GPU. So, we're on Blackwell right now. Next one's Vera Rubin. The first VR ones will fit into the same racks as Blackwell. After that, they move to Kyber racks, which are giant new expensive racks that will not work with your data center at all. So you have to build for what's available today because you actually don't know the future. If you did, you wouldn't be building them at all. But anyway, so Oracle has no choice. They have to do this. And every delay to a building project, walking around London, you can see how silly it is to believe the building gets finished on time. They have to do this, but every delay means that it's more obsolete. But this is the ultimate problem with AI as well. Every data center you see is with full of obsolete gear, no matter when it's built because it takes so long to build the data center and you have to buy the GPUs up front. Nvidia won't sell to you if you don't buy them in advance. Don't expect Nvidia to have them in stock. And this is the this is the problem with having a monopoly as well. But yeah, Oracle has to do this. And Oracle when they when they started this project, which was before OpenAI was involved in it, it was originally built for Elon Musk. Then Musk decided to build his own thing, Colossus, which is its own mess. Back then, everyone was extremely horny for AI. Everyone was saying that AI was going to be the biggest thing ever. Every bank, JP Morgan, who backed the joint venture for Abene, they were like, "Yeah, yeah, sure. How much do you need? You You want more than that, mate? I could give you more than that." Back then, money was easy. Today, it's not so easy. And what's crazy is I think the AI bubble could be burst far before Abalene is built.
>> So just going back to Oracle specifically, I mean if they cut tens of thousands of their staff, I I think the the estimate was it would be around 18% of their staff >> to free up as much as $10 billion in cash flow. Would that even be enough to address this imbalance between their between what they're spending and what they're earning?
>> God, no. Not even close. But I mean I it's about uh $14 million per megawatt just for construction. I mean they're going to build like half of a data center, one of the five they're promising. Like I it's it's very much rearranging the deck chairs at the Titanic at this point. And I think that what makes the media incapable of meeting this moment is that to imagine that someone like Oracle, decades old company, their first contract was with the CIA. I'm being serious. One of the they're in governments. They nearly made the Birmingham City Council bankrupt. Again, that's true. I swear to God. Oracle is an esteemed entity. So, it's difficult to imagine that someone like that could collapse. But the problem is they feel that way, too. They can't imagine their own collapse. So they just think drill baby drill. We'll just build as bazillion data centers and uh yeah, money will come out. Maybe I don't know because this is the thing. The tech industry is incapable of adaptation. They talk about being innovative and disruptive, but really since like 2018, the tech industry has been stagnant. Software companies stagnant. Oracle stagnant. They've missed on earnings a ton of times. The fact they beat on this one is because they're pumping money into AI and selling things cheap as well. Selling compute cheap. So what will happen to these data center projects will be dictated when Oracle stops paying builders because you can OpenAI has net 360 terms with Core. Like software startups take a while to pay anyway. However, you stop paying a builder, they fold their arms and they'll stand there and stare at you until one of you dies. builders won't work for free. I hear that Stargate Abene is paid up construction-wise, so maybe that'll keep going. But the data centers in Port Washington, uh, sorry, yeah, port uh, Wisconsin. Let's just say Shackleford, Texas as well. Those two data centers are in construction. They have a steel beam up in one of them. I'm not even kidding. There was a post from uh, OpenAI saying, "I've got a steel beam up." It's an Allen Partridge thing. Very weird. These projects will more than likely be construction hell. They will just sit there unbuilt. It will be horrible for the communities because of all of this upheaval. All of those things will stop. And honestly, we've never seen anything like this in America. You kind of saw it in the Dubai construction boom and the great financial crisis, like kind of lots full of luxury cars, unbuilt buildings. But I think this could be all over America. I don't think these data centers get finished. If I'm wrong, we will see possibly the most successful construction boom in history. I'm not going to be wrong. I'm not going to be wrong because the other thing that no one seems to want to talk about is these are construction projects that borrowed money. Have you ever heard of a construction project that came in on budget? if these data centers do end up going over or falling short of what needs to be done and I mean Oracle is already I think almost 110 billion dollars in debt and they're obligated to spend another 156ish billion dollars on data centers in the future. I mean what what are those still investing in Oracle seeing that? I suppose we're not seeing is what I want to know.
>> It's the beginning of history ma'am. They're not seeing anything. They are looking backwards and they are saying what worked in the past. Well, it's convenient for me to believe to confirm my biases to say, well, what worked before was spending a lot of money. It's nothing like Amazon Web Services. I found out I actually went and looked Amazon Web Services in the 11 years following 2003. So, the lifetime that it took for Amazon Web Services to get built and turn profitable was about $38.9 billion. That's not very much money compared to the 40 bill 42 billion OpenAI raised in 2025 alone. Oracle everyone's convenient thing is saying well the com bubble there was a lot of spend except AI data centers are not fiber. They're not. They're not as useful and they don't last as long. Um I think that people are just dancing with the ghosts of the past. I think that they think that because things worked out before things will work out today. Vegas loves people like this, by the way. People that just see a few successful dice rolls and think they're going to make a billion dollars. Adore this is this is the market and this is the society and this is the realm where grifters feast. Because think about it, Sam Alman, all he needs to say is, "I need you to build an absolute crap ton of data centers and when you do that, we're going to make an absolute crap ton of money." But is that true? No. But it can justify and rationalize endless irrational and horrible and expensive and ruinous financial moves because it confirms people's biases. And I think that a good question to ask right now is why is it with this so take anthropic for example. Anthropic had this big burst of users following the big claude code push and all this. I don't see Anthropic building any more data centers. None of them came online, but they made all this extra money. Perhaps it's not true that the more compute they have, the more money they make. Perhaps that was a madeup thing. Because that's all this is. Dario Ammedday and Sam Orman are grifting the economy. They are grifting business idiots, which is a lot of the people with the money in running software companies that are facing what is annihilation. Not saying these companies will die, but the growth story in software is coming to an end. And they're do the grift is amazing. The grift is, hey, your growth's slowing. You want to do the thing you love doing? Spend money. You could spend money. And hey, while you're spending money, maybe you could lay some people off. You love that, right? We love doing we love laying people off. It it is exploiting the ignorance and desperation of a software industry on the decline. and it will suck for our markets and our economy. But I think this era is necessary to wash out an era of rot. to bring things full circle on on both stories that we've talked about. The the line that Oracle is going with that they're tied to sort of building these data centers whether they are out of date or not is directly linked to this uh this story that is also being sold by Sam Alman and and the like to make the most of the decline of the software industry.
>> Yes. So Oracle is actually a really good example of this because as I mentioned their core business, the one that made them famous is plateauing. It's stagnating. It has been for years. You just go and look at their revenues. When you cut out AI compute, the everything is kind of flat. AI and what Samman and Dario Amade promise is this panacea. It's this thing of if you plug AI into it, well, it will do everything for you. It will maybe replace employees. I guess it tells everyone what they want to hear. It tells them that not only will you have revenue growth, you will also have less costs because you can lay people off. Is this true? No. God know. You'll have more costs. AI actually increases costs. And nobody seems to be making money off of selling AI services. But the con is on and it plays on that kind of intellectual fragility and desperation to cling to the past because they say all you need to do is what worked before. Spend a lot of money. Say the right words. The business idiots you sell to will buy from you. Except the problem is is that even even a dumbass, even a buying something needs to get some utility out of them. and they definitely don't want that utility to be with spending more money on operational costs. So we're in this situation where everyone is desperately clinging to this idea of AI restarting growth in software except what it's doing is it is growing something. It's growing operational expenses and it's growing capital expenditures in the case of Oracle. It's really genuinely bad and it's only going to accelerate the SAS apocalypse which is absolutely not, I must be explicit, created by anything to do with what AI can actually do. Software is not being replaced. Claude code is not typing make me an enterprised resource planning platform. That doesn't work. That doesn't happen. But it's a much more comfortable lie for the media, for CNBC, for Bloomberg to say this is caused by AI. Because if you say all of this chaos within SAS is caused by the glory of AI, you can cling to the dream that AI is just another way where hyperrowth can go. It's just the next thing that will grow everything. Because if you have to accept the truth that there is no hyperrowth thing coming that we're we live inside a dream and the dreamer is Sam Alman is Darday another grifter grifting an industry of software that's been run by grifters private equity firms buying grifters companies to sell to other grifters we have left the era of software being useful and being sold as a fair exchange of value and the natural end to that is a slowdown down. And when I say collapse, I don't mean the end of software at large, but I think a revaluation of tech stocks writ large. And it's impossible to avoid unless they come up with something useful. And based on the last decade, I don't rate their chances.
On that note, Ezatron, thanks for taking the time.
>> Thanks for having me.
>> If you like today's episode and you want to hear more of the tech report, please consider liking and subscribing. If you're not aware, you can also get episodes of the tech report wherever you get your podcasts.