Nvidia AI Bubble Cooked by 2027, AMD Morale Leak, RDNA 5 Comeback | Ed Zitron | Broken Silicon 365

So, I think what's going to happen is everything's just going to slow down. I think you're going to see data center bans here and there. I think Ohio just banned the tax breaks, which is going to make them way less appetizing.

>> And welcome to Broken Silicon, a gaming hardware podcast. I am your host Tom. Today's sponsors are Wondershare Recover and CD Key Offer. Check them out through the links below. And today, um, it's funny. I've been kind of, I don't know, say, let's say it, I've been hyping you up that you've been coming on the show. I've been saying, you know, Ed's coming on soon. So, you know, get ready to ask questions. And as we get closer, I'm like, h, ah, you know, there's different guests that could come on. It's like, you know, there could be a PlayStation-centric guest in Nvidia and Intel. But the closer we got to our agreed-to time recording slot, I was like, no, this is more and more feeling like perfect timing to have you on to talk about AI. And I wasn't, I wasn't sure if at Computex how much AI there would be. I was relieved to see there was some non-AI talk at AMD, but there was ample AI. Um, but please, please introduce yourself and and I am honestly curious like, what is your history and and what, what would you say you do professionally and and why do you do it?

>> So, my history was my first job was at a PC gaming magazine called PC Zone in London, which I worked until I was 22. Then I moved to America into a PR job and I stayed in tech PR until mostly like the beginning of this year, end of last year. But even then, the PR business is kind of, it's still going in that I have one or two clients that I deal with very specific reporters for. But my main thing now is Better Offline podcast with iHeartRadio and my newsletter, which is actually the primary source of income. Yeah, I haven't done, I haven't done the PR as the main squeeze in a minute. And that's because I got COVID in 2020. I started writing a newsletter and then over the course of a few years, oddly, just became an analyst of sorts. I just started, and I don't have a history in finance or anything like that. I just picked it up. I just kind of like for the sheer love of the game. I didn't charge for my newsletter until the be, uh, sorry, the beginning, the middle of last year.

>> Mhm.

>> And that was, that was kind of where the PR firm started winding down is the wrong thing, 'cause I still got one or two clients, but they're like, so separate to everything. It's all consultancy stuff. But yeah, so pretty much my job is newsletter and podcast.

>> Mhm. Which is fun and I love it. I, I genuinely adore doing this job. I love writing it every week. I love doing the pod. I love it. Like, it, it fills me full of vigor, especially because there's so many [ __ ] and also just AI is one of the weird, even the tech industry at large. It's one of the weirdest things 'cause in your neck of the woods, for example, when you look at the journalists covering Nvidia today, I genuinely don't know how many of them know this company made graphics cards for games.

>> Yeah. I don't think any of these mothers knew what a 3DFX card is, for example. Like, no one, no appreciation for any of that [ __ ]. They knew AMD as an AI compute company. They, they don't realize that any of these companies have any other business lines. Now, I get that because in Nvidia's case, it's like 90% of their revenue is data center, but still, these companies do many other things. And so, it's been fun kind of being willing to be a bit daring and to learn this stuff and to really get into it. And then sometime last year, I just really got deep into the financials. I got really in the guts of it. I don't know what about it did it, but I was just, I think what it was was the actually, probably end of 2024 as well. I was like, "Okay, is any of these companies making profits? Any of them?" And I thought I would look and I would see like, "Oh, Google's nope. They don't even tell you what AI revenues they make." And so, yeah, just a bit of brain madness that turned into a natural career. You know, there's something I'm sure you'll sympathize with considering your gaming kind of, [clears throat] you know, journalism background. Um, and and it's that I feel like, like there's like two main types of coverage of gaming stuff. It's like the people that cover the hardware and they're like, "Oh, the gigahertz, how many cores this has." You know, they argued over teraflops for argue stupid reasons this generation between the consoles. Um, and then there's the other people that just like cover the games and they're excited about the games, but it's rare that you find someone that is like, "Hey, is this even possible?" 'Cause it, it's not just AI. Like I think of this, this is basically what I do as well, but also like I know people like Colin Morardi do this where he was like very at the forefront of like Game Pass for Xbox, just, just it isn't going to work. Like the math is not mathing. And why is anyone thinking like buying a $70 billion Activision Blizzard is going to result in making less games? Weird. That was such a strange acquisition.

>> They have to make games for PlayStation if they're going to spend $70 billion because otherwise they won't make their money back and

>> while also cutting Xbox own revenue when it comes to the Call of Duty titles. Just insane. Sorry.

>> No, but yeah, I was, I was just thinking like, what do you think about this like dissonance and and I think it is changing, right? But this dissonance in coverage where it's like, either I, I, it's, it's like just access journalism. I hope Intel will send us the CPU first so we can say good things about it. I'm not really sure what our job is at that point if that's all we do. To the people that are just like hyped about games but don't think about the hardware. Like I think it's changing. I think there's people like you, honestly, have people like Hardware Unboxed or Gamers Nexus ask good questions quite often. There's still a lot of these legacy large websites, whether it's covering the games being released or the hardware that seem to just make their whole business around getting 5,000 views a YouTube video because AMD sent them something. I, I wonder what you think about that because I think this bleeds into how people treat these AI companies. It's the same mentality. So, I love Steve's been on the show quite a few times. Steve Burke from Gamers Nexus, he's fantastic. And it's interesting because so back in the games journalism days, I had the same problem. I was a really pissy little 16-year-old when I started. And at the time, I was very confused because people would get in trouble for giving too bad a score. If you scored something too low, they'd be like, "Well, the publisher won't send us stuff." And I remember getting into multiple arguments where I'm like, "Right, but this costs someone money." And I think that that's where a big kind of interference happens because if you have access and you're not paying for stuff, unless you're extremely intentional and you're like, "Okay, I'm putting myself in the mind of someone spending $300, $1,000, whatever particular piece of hardware," you kind of lose the plot. Similarly, I don't think many journalists think very hard about what their job is.

>> I think they think whatever their boss tells them, what's big right now, I will say what's big right now. And then they are happy, sad, they feel absolutely nothing. I don't know. When it comes to gaming stuff, I mean, there's just this massive influencer thing going on. Well, gaming is weird as well because like gaming hardware, there's a ton of influencer stuff. PC hardware, ton of influencer stuff. And it's not really, I'm surprised it gets traffic, but it does that there's people who are just like, "I found a laptop. Wow, look how, look at screen. Ooh." And it's, you can find a happy medium. I really like Engadget for example. They had a few weird years, but like Sherin Lowe's done good work there. But I think the problem is is that journalists don't see themselves as having a customer. They don't see themselves as having a responsibility to someone at the other end. They're not thinking, "If I tell this story in such a way that someone gets the wrong opinion, someone makes bad decisions, that's my responsibility." And that's because often times it isn't. It is just, it's just really easy to forget about that. Now, this is kind of crushing together a lot of industries. When it comes to AI coverage, I think it's just that the vast majority of journalists don't give a [ __ ].

>> Mhm.

>> I don't think they [ __ ] care. I think that they, I think they want to see the companies win. I think that they believe that capitalism is the ultimate form of society and that they can only ever celebrate capital's victories rather than people's victories. People pushing against data centers, which I know we're going to get into later, they frame those as weirdos like campaigning against the progress of AI. They frame AI data centers, many of which will never get built, as this thing where it's inevitable and necessary, so that AI can happen. Now, you're probably wondering, do they ever ask why? Like, what does more data center do? God, no. They don't even know these [ __ ] things work. It's the put the put the story in the bag, ship the bag, money go up, number go up, raw economy, blah, blah. It's all the same [ __ ]. And it's frustrating because the AI bubble was inflated using this to its ugliest successes. Like this is the worst of access journalism, the worst of thoughtless journalism, the worst of this irresponsible journalism all kind of wrapped into one thing.

>> [snorts]

>> So I wonder how much of it you think is like the access journalism and how much of it is just because I have to watch myself. Like, I, it's very easy to get caught in a pattern of work. Like I do the podcast every week. Some contact at AMD or Intel or Nvidia tells me something no one else knows. Good. There's my 20-minute video. Let me just get that done. I, I don't want to work Friday. Like, it's very easy. And then sometimes I have to remind myself, like, this is actually pretty big. Maybe I should take a week to look over this. Maybe like, but if, if it's hard for me when my work is very all over the place, I can imagine someone that reviews games or reviews, like you said, that just every laptop's good. Like, I got the thing, I do the thing, I go home on the weekend. How much do you think is just they get stuck in a pattern and then also they see everything as inevitable? Like, it's not just that, like they're rooting for capitalism so much as they just think whatever the companies are going to do is what they're going to do and you can't do anything. And, and I say this because I don't think that's actually true. It's only true if they let it be true. Like recently AMD was making this bizarre decision to not bring the latest FSR to like, really most of their products and everyone got mad. It was actually very funny. Um, every comment on the AMD YouTube channel and Twitter, like, even if it didn't have to do with gaming, it, the example I give is they'd be like, "We're partnering with BMW for AI, whatever." Every comment is, "Could you partner with BMW to bring FSR to RDNA3?"

>> That rocks gamers. Gamers rise up. And now they are. And everyone's acting like DLSS 5, which is just an AI post-filter, is going to be inevitable. And I'm like, it isn't. It isn't inevitable. If we don't want it, they will be forced to change. And how many journalists do you think have seen so many things just keep going whether they like it or not, and they don't realize it really isn't inevitable unless they let it be inevitable.

>> I actually think you're really on to something with that because I, I am quite critical of a lot of tech and business journalism, but I actually think you're right in that I don't think it's people sitting there being like, "Oh, yeah, I want the companies to win." There are some of them, don't get me wrong. I think it's a lot of people who have, like you said, built habits and they're like, "Well, this shit's going to take off 'cause they'll shove enough money into it and it will work eventually and who gives a [ __ ], put the article in the bag." But I think that there is also a level of editorial laziness. I think the editorial crust doesn't support them, doesn't teach them good habits. I don't think a lot of tech editors are very good. I don't think they know what they're talking about. And also on the financial side, I think a lot of journalists are just very lazy with this stuff. I think that they just go, "F it's all numbers. I couldn't possibly addition and subtraction. It's not difficult." I get if you don't want to read an S1 or a 10Q or what have you. I don't, you don't have to be a real data pervert. However, addition and subtraction and by, but the thing that really gets me and I think it does come back to the inevitability point you made is even if you think it's inevitable, the things I've been saying for two years are now mainstream of like, "Hey, these companies lose billions of dollars. How's that work?" And I've been talking to people for years who have been saying it's just like Uber, just like Amazon Web Services. It isn't. I've proven this many times, but people don't even want to check their homework. They've come up with the the preloaded myths that they want to attach to. And I think that the reason that AI has taken off is it exploits everything we're talking about. It's just it exploits the inevitability. It exploits the willingness to do access, even quite simplified access where you're not even getting very much where it's just, "If I piss off OpenAI, they're not going to send me the same press releases they sent 30 other people." And I think that, I think it's sad because the AI industry could have been stopped. They could have been stopped in 2024 had had there actually been a willingness within the tech and the business media to scrutinize these companies. This was a, a media-driven hype cycle. Without it, business, if I don't know, journalists had said, "Yeah, these things don't really do what people are saying." I don't know, businesses might not have bought them. But why would we do that when there's all this access happening when, and actually, there is another thing which is

>> before 2023 things were quite bleak for the tech media. There wasn't a ton for them to get excited about. So even something that [ __ ] sucks was something for them to get excited about. It's a mixture of things and none of them great.

>> Um, Jam Wap writes in and says, "Obviously AI has some good uses and might boost worker productivity, but people forget that worker productivity has already been increasing over the past several decades. In part because of advancements in computer hardware, even if you don't put an AI sticker on it. A consumer PC from 2026 can smoke a high-end workstation from just 10 years ago. We also have laptops with 20-hour battery life that don't break your back or burn your lap. No magical buzzwords are required. So why are these big tech CEOs so hellbent on selling us on some revolution? Am I crazy?" I just failed to see the light. And I agree because I've always found something interesting to talk about in tech. And yet, it's if it's not a revolution, like I kind of built my channel like on Intel's kind of falling apart and comeback and AMD's resurgence and I, and to this day, like from what I'm seeing for Zen 6 launching next year and I guess to an extent this year, it looks way better than Zen 5, better than Zen 4. It looks incredible, but it's not Zen 1 or two which is revolutionizing the industry and making Intel go out of business. So people aren't always as hyped. And I'm like, guys, this actually looks incredible what they're achieving. Why does it need to be a revolution every time? So for the media and for the markets, it must be a revolution or it's nothing. So I have this theory called the Rockcomb bubble, which is the reason they're doing AI is because they've run out of hypergrowth ideas. They don't have a smartphone. They don't have, uh, cloud computing. They don't even really have containerization. They just have like AI and AI capex and APIs and [ __ ]. They don't have another thing to look at, another thing to say, "This is going to be what makes Google a $15 trillion market cap company. This is going to be the next Microsoft 365 or the next Google search or the next Amazon Web Services." So something that's really exciting and cool for you, that's not going to move the, that's not going to do enough. That's not going to be an MI450 Instinct sold 10,000 times to produce some sort of growth in the future because guess what? Lisa Su doesn't give a [ __ ] anymore. She was smart for a few years and now she's fully AI-pilled.

>> And so what it is is

>> She, she, not, I'm sorry to cut you off, but she comes off as the least AI-pilled out of them.

>> Oh god, no, mate. I'm sorry. No. Lisa Su is full of [ __ ]. No. The reason I say that is in October of last year, Lisa Su announced that OpenAI was buying 6 gigawatts worth of AMD chips

>> and uh they would be able to buy 10% of AMD stocks. That just never happened.

>> Just never happened. Never came up in revenue guidance. Never happened. She went up with that, that [ __ ] ghoul Greg Brockman at CES and burbled out a bunch of meaningless talking points. That's why I don't like her anymore.

>> You know, actually, I, I want to take a chance to bring up something else as well. This was something I was going to bring up on the live stream. By the time this comes out, it would have been last week, but something that is kind of like a pseudo-leak or really a warning, an open letter to like the middle and upper, especially upper management at AMD, but I got reached out to by people both at AMD and Nvidia, a few of their partners, because I don't want to say specific companies, but

>> a specific issue where they're like, "Actually, I'm getting really annoyed with AI right now because the at the executive level, they are pushing it super, super hard and it's not just that it like can make errors or mess with our work. This person said, "Actually, if and one other person at another company echoed this, if you're like really, really, really good at your job, like fully understand the code and the process that you do, AI, as long as you're allowed to burn tens of thousands of dollars in tokens, can actually double your productivity if you're that person because you will always see the error or know how to use the AI to not get the error. But that's not most people. And this person's definitely concerned, definitely concerned that what they're seeing is a lot of their employees that they feel aren't the most competent ones come off as competent because they're kind of pseudo doing their job quicker, but then they spend all this time fixing bugs in the product and that they're very concerned that over time nobody will really know who the good employee is and five years will have gone by and half their products won't work. And they're just, they're praying executives stop pushing AI. They're happy to use AI, but don't make us use it when we don't want to."

>> Yeah. So it's like imbecile evangelism. So large language models do an impression of work. That's the best way of describing them. They can spit out functional-looking code and if you bonk it on the head enough, if you're able to just do as many turns as it takes for the code to run, you can feel like you know what you're doing even if you can't read a line of the code. So it has allowed, and there's a guy called Nick Sach, who writes an excellent book called "Look Dick." He wrote a piece called "I'll [ __ ] Pile Drive You If You Mention AI Again." He makes a really good point, which is you can, these things are more dangerous with imbeciles than anyone else because instead of being like, "Yeah, this guy can't do [ __ ]." This person can half-ass their way through writing code that runs, but it's heavily verbose. No one really knows it was written without intention 'cause these models don't think. And then, yeah, this person like you said, can kind of hang around the organization without being fired, being a decent enough diplomat while being [ __ ] at their job and leave someone else to fix the problems. And because LLMs write so much code, they have no consideration of brevity or anything like that or any consideration for that matter, these people, these like secret imbeciles like the scrolls hiding around organizations waiting to pop up to [ __ ] things up 'cause that's what's happening. AWS went down twice because of LLM code. And the problem, the other problem is, is that organizations are actively encouraging non-coders to use them. I know spoken to multiple people at Meta who all confirm that tons of code is shipped abroad by non-coders.

>> The organization, the morons are running the the companies, the business idiots. I know we have a question around that, like how they are controlling everything. And the problem is, is that there's a coddling that happens with LLMs both for the user and the greater organization where it's like, "Well, it's not that you can't do something and thus should learn, just use the LLM."

>> And if the LLM can fart something out, kind of approximate good-looking great, the problem is that the competent people now have way more code to read

>> if they need to fix it. They have way less time to do so 'cause people get being [ __ ] laid off. So it creates this situation which I think is going to destroy at least one hyperscaler. I think Meta could die in the next decade. I think that Meta will go down at some point for a couple days. I think that's

>> Do you believe they're pushing the use of AI more than the others? I mean, they all basically are, but is that why or why them?

>> Uh, they are because Mark Zuckerberg is fully AI-pilled. He's just a, a, the world's largest.

>> Just like he was VR-pilled, now he's AI.

>> Exactly. It's a guy who has no ideas found an idea and they found an idea machine. They built that guy an LLM just for him. I want, I want access to that LLM so bad. I want to talk to Mark Zuckerberg's chatbot so bad. I want to fit, I want to make it speak to him in like, uh, what was it? A new, a new kid, like the angel language that drove Alistair Crowley insane. Like, I want to like get in Mark Zuckerberg's head and be like, "Mark, Mark, grab the officer's service, revolt. Mark, it's time. It's, it's time to do it, Mark."

>> You know, one thing that worries me, um, is because this plea from, and again, it, I found this across several companies, but specifically a few people at AMD is this sort of a thing. It's not the same as what happened at Intel at the base level 'cause, you know, 10 years ago, Intel didn't have AI, but it's kind of the same thing. Intel became a monopoly-ish. They got really lazy a lot. I kept always hearing from contacts that some people would do good work, some people wouldn't. And no one would communicate. And then they would ship, I mean, Sapphire Rapids was delayed how many years? It had like a bazillion stepping because that kept not working every time. And it's because they're just using old tools and people just go home and not really check if their stuff works. I'm worried this sort of thing could then, because of AI, happen at AMD and actually happen at a lot of companies.

>> Yeah, I think it's a lot of companies. I, like, I just, I'm going to say this with complete certainty. Microsoft is one of the most middle manager heavy. It's like the middle manager borderlands. It's every, I speak to people at Micro, this is one of the, this is a personal thing. I get the occasional criticism. It's like, he doesn't speak to anyone in the tech industry. I speak to tech workers all day. They reach out. I love them. Eitron76 on Signal. Reach out. Have a chat with me because when I speak to people at Microsoft for like the last two years, they've been horrified. But also the people running the place are middle managers. It's all middle managers. It's the middle managers who are now using LLMs and going, "Well, I know how to code. Why don't you do code? You need to do way more code." And like you said, these companies that have some sort of monopoly or some sort of presence in the market that they think makes them strong, I think are going to just start falling apart. I don't think they'll be beaten by anyone because everyone's got poison with this technology, but I think Microsoft's in real trouble. I think Meta is. Google's interesting, but I just talked to someone yesterday who said that Google is fully AI-pilled. And putting aside the financials here, the culture of these companies has been hollowed out by AI. The, because it's like a fight between the competent people and the incompetent people. And the incompetent people outnumber them because they're hired by VPs and middle managers who basically check their emails, go to lunch, and leave lunch. That's, that's most of their day.

>> Mhm.

>> Um, well, let's get into this now. Girth Maxer, which that name is definitely an inside joke for the channel.

>> Uh, as a big fan of Ed's work, can Ed explain the role of business idiots in the AI bubble and how little work the executives push AI actually do?

>> Okay, so the term business idiot is what I used to refer to, the middle manager and CEO sect, I think I'd call them, of I mean, just we've all had bosses, and I think that the majority of bosses I have don't do anything. They, they have great consternation on their face. They're always hanging around meetings or lunches with clients, but you don't, and they're like, "Oh, I handle strategy." But when you talk to them, it's very clear they don't have any idea what's going on or contribute. I think this is most of the economy. And put aside [ __ ] jobs. I think [ __ ] jobs, great idea, but I think it misses the problem that there are useful knowledge work careers. There are, but we've created these layers of management everywhere that are weakening every business. And the problem is MBAs control the world. Sundar Pichai of Google, MBA, McKinsey, uh, Andy Jassy, MBA, Amazon, uh, Satya Nadella, MBA. The guy who replaced Andy Jassy, I forget his name. Matt Garman, I think his name, MBA. All these people have masters in business and they're, they're [ __ ] morons. They just, they know how to say the right things to the right people to get hired. They know how to say the right things in the right meetings to seem like they're in control. They know the right things to say or do at a particular time to appear like they are the success story behind a team that did the real work. And these people are [ __ ] parasites and they're everywhere. AI is their dream come true because these people are all ideas guys. We're ideas guys. All that's between me

>> I'm a big picture guy. I'm not going to get into the details.

>> And all that's between me and greatness is this messy [ __ ] called effort and knowledge. I, I got idea. I got big ideas here. I'm a genius. All I need is just a machine to do all the knowledge and the work for me. These people believe that they are genuinely geniuses that are just misunderstood and they never have time to learn things. They couldn't possibly learn a new thing because that might take them away from meetings in which they claim other people's work as their own. We really should run MBAs out of every business. I actually think MBAs should be exiled from society. I think that they are a dangerous career. Uh, sorry. They're dangerous parts of anyone's career. Every MBA I've run into for the most part has been an impedance to an organization. I'm sure someone will hear this and be like, "I've got an MBA. I'm not useless." Great. Thank you. Thank you for being the exception. Tell your friends. Make it, let's make MBAs useful. I don't know what to tell you. The majority of them I've met are people who have just got an MBA to get a job that involves telling other people what to do. And I also think that there's one abstraction higher, which is the dominance of middle managers.

>> Mhm.

>> I can think of every manager I've ever had for the most part, with the exception of maybe, well, with definitely Robert and Sophie at Cool Zone with iHeart. They are amazing. But pretty much every manager I've had is useless. Just completely [ __ ] benign. A person that yells at me to do stuff, then does nothing. That makes me do everything. I've talked to so many people about their managers. Same experience. [ __ ] my manager didn't do this, my manager didn't do that. I asked my manager for something, they told me to do it. I said, "I need something." My manager said, "Go ask them." What does a manager do? Nothing. So, I think we are in like the middle manager apocalypse. And that is what the business idiot thesis is really about. It's that we've allowed these people to poison our society and then they released a tool built for middle managers. Like a, a LLM are the dream of the middle management, just telling someone what to do, but without a messy little person to have to undermine or steal from.

>> So I, I have to ask though, do you think this is a bit localized to tech and maybe journalism? Because I worked at General Motors and maybe I got lucky. Well, so yeah, I guess I worked at General Motors in 2014 and I would say my manager, she was actually incredible. Having said that, she was brought in because her predecessor was the opposite and she had to clean up his mess. And we would make fun of like which teams had real managers and not. But maybe I got lucky, but and then I was an account manager at a supplier to like, you know, GM and other companies. And I'd say my manager was very weird, but he worked very, very hard. I, I think of my girlfriend. She works in the music industry. Uh, her manager sounds incredible. I've met her. So I don't, in my, maybe I'm lucky, but in my, and uh, I'm thinking too, like my brother who works in like genetics, his manager is great. But I, it hasn't been my experience that all middle managers are bad, though. But of course, I haven't been at the mid, upper middle level, I'll say that right, where maybe there's a difference between the people managing a team that works on seats at General Motors and the person managing all the people that manage all the teams that build all the things. I, I do you think there's a chance some of this though, like harshness is because you have been exposed to a lot of it? Or and it's not a bad thing. You're looking for where there needs to be improvement and so you get a lot of, you see a lot of the bad because, well, really your job is to find the bad because I'm curious though. So I can see that. I think, yeah, sure, I am definitely tech and journalism heavy. But even speaking to friends of mine in service industries, they feel the same way.

>> Mhm.

>> Uh, the people that own the businesses are useless, feckless layabouts. The higher up someone is in pretty much every organization, the less likely they are to do work. I, I have seen this across the board. Yes, I am definitely data-biased to the tech industry and journalism. I sure am. I know plenty of people in other industries, people in the film industry who have producers that don't do a [ __ ] thing.

>> Well, I think everyone here can easily believe that though.

>> Yeah. Yeah. But it's people whose actual jobs require them to do stuff who just don't do stuff. Like, it's really, it's a frustrating thing because the term manager is broken. Like that. This is a, and maybe this is just knowledge work focus. Maybe, maybe this is the only thing I can speak to with depth because I'm very lucky to have avoided what I would refer to as real work where I have to like stand up all day and lift things.

>> Mhm.

>> But I, I still feel like there is a problem where we have elevated this term manager or management to mean not working. When in fact, a [clears throat] manager should be someone who actually moved, moves things out the way and gets you the things you need and actually plans for you so that you can work. A nurse, right? A nurse can mean someone that went to a four-year school or a nurse can mean someone that, I don't know if you knew that, can mean someone that maybe did 12 months of training and

>> Yeah, well, like a vocational nurse, a registered nurse.

>> They all, yeah, versus they all nurse, but actually one of them has quite a bit more.

>> Even a vocational nurse is still quite a qualification. But nevertheless, with it, with that in mind, it's kind of like the term coach.

>> A coach can mean someone that you occasionally talk to, gives you an idea, and then you do a thing, or it could be someone that naturally actively mentors you and says, "Okay, today here's what you're doing." Bam, bam, bam. And I think human beings are naturally going to choose the one that means le, the least amount of effort.

>> Like that's, that really is going to be it. Like, it's, it's going to be whoever has, who I think it's just, let me rephrase that point.

>> I think some people see management as an escape from the pearls. They don't see it as a responsibility. They see it as a chance to section themselves off from the messy world of work.

>> Yeah. Well, and I, I think that there's also a thing too where, you know, this is always a worry. I mean, when was this? This is probably the worst actually in, I don't know, the Soviet Union where like all managers basically did nothing and just sat there and, you know, maybe this is a thing that has become worse and worse lately as well. Um, but actually, have you seen the movie "Good Fortune"? Because they, they actually

>> so it has, uh, Keanu Reeves is like an angel who

>> stops people from texting while driving and he wants to do more than that and Seth Rogen and Aziz Ansari, sorry, he's, sorry, he's like an Amazon driver. Seth Rogen is one of these manager types that has a nice house in LA that you're talking about and they make them switch lives and when Aziz as Seth Rogen goes to a meeting on a busy day, they get sushi, talk about work for five minutes and they're like, "All right, well, let's call it a day." [laughter]

>> I feel like that movie made fun of kind of what you're getting at with some types of managers.

>> Of course. Of course. And I think that many people know this. I think it is a pop culture thing. I think that these people are, these people have always been in these positions. And I think LLMs allow them to feign a level of, um,

>> feign a level of competence that I don't think they have. And also do the thing where they say, "Well, okay, if they like, okay, if I can do this, you must be able to do this 'cause you have all the time." Now, what am I doing? I'm a manager. I, I have management to do. I have to go and ask someone else what they're doing. If I don't do that, well, I mean, then nothing would happen because I wouldn't ask them to do it. And then there's, I've, I've also had some real [ __ ] house managers, I'll admit.

>> Mhm.

>> Look, let's be honest. Storage is really expensive this year. And so if you can avoid buying more of it until prices come down, you really should. And that includes if you have a catastrophic data loss event or it looks like your drive is breaking, instead of going out and buying a new piece of storage immediately, you should really try to recover that data and recover your drive. And you can do that with today's sponsor, Wondershare. Recoverit. Look, it is honestly one of the worst things that can happen to you as a person or even to just someone like me who's a creator. Maybe you have a drive with tons of pictures on it that you've just left there. Maybe you didn't think about them all being on one thing that can fail, or you're a creator like me. You, you know, you shoot a bunch of stuff and then suddenly something happens and it feels like everything's gone. But it doesn't necessarily need to be gone. And that's where Recoverit comes in as a natural solution to this worry. You see, if you think you've lost data, what you should really do is just stop using that drive immediately to avoid overwriting the files first and then open up Wondershare Recoverit. Recoverit can scan for lost or corrupted files from solid-state drives, hard drives, SD cards, and even formatted storage, going into that formatted storage and trying to recover things that you thought maybe you accidentally deleted or someone else did. It can often recover those even. Also, it is incredibly easy to use. Just check Recoverit's high Trustpilot score and you'll see that it's at least partly due to the fact that it is very easy to use and everybody says so. And so that's really what you want though, right? After a data crisis that you know, you're, you're going to be stressed out already if you thought you lost everything, so you should really have the most easy-to-use interface that can restore data standing by so that you can restore it with one click. And well, again, that's Wondershare Recoverit. So one more time, before you go out and buy new storage after a data loss, it's definitely worth trying Recoverit first. And you can do so below for free. You can go down and download the app for free to try it out. And just remember for the future that it's free to do so. You don't need to buy some other app and you don't need to buy a new SSD right away. You can download Wondershare Recoverit for free. Scan your drive and see what files are recoverable before you pay somewhere else.

>> Um, Chris Rich writes in and says, "It's well established in computer science that simple metrics of developer productivity such as how many lines of code they wrote are useless and counterproductive." Yeah, someone in the Moore's Law team reminded me that like, I think IBM paid people per line of code or tried to incentivize that. And I'm like, I can't think of a dumber incentive structure.

>> Oh, that's silly.

>> He says, "Yet with AI writing code, all I see is boasts about the amount of code written yet again or the number of tokens used, which is basically the same thing. Aren't we just repeating mistakes from the previous eras? Are there any signs of these problems being taken seriously?"

>> Well, I think right now the AI ROI conversation is where this is happening.

>> Mhm.

>> Like I think that the f So what happened was in sometime, I think it was March, but it really bubbled to the surface in April. Both OpenAI and Anthropic moved their enterprise customers from these subsidized subscriptions where you could kind of burn as much as you want within reason, within like these rate limits, to token-based billing. And suddenly organizations which had been saying, "Yeah, just use as much AI as possible. Burn as many tokens as possible. That's how we know you're doing the good shit." They are, well, they've now found that, "Oh [ __ ], telling people to burn as many tokens as possible makes them burn a bunch of money." And the lines of the thing is, again, goes back to the manager thing. When you're a manager that doesn't have any actual experience with the work and doesn't really check, I don't know, is the software good? Are we hitting, do we have KPIs we've hit? Do we actually have KPIs? Have we set up timelines? Are we meeting those timelines? Is the software stable? If you don't have a manager who's directly connected to the process and productivity itself and production itself, you just have these varying metrics of like, how many lines of code did you ship? Have you used AI a lot or a little? Do you use AI every day?

>> And I think what it is is the person's question is correct. It's like we're repeating previous mistakes but also compounding them as a result of those mistakes existing by not having strong managerial culture which actually measures by having practical experience because managers more often than not don't do stuff. The rare exceptions are there. But

>> See, that's the funny thing. I have to say, when I worked at General Motors, my boss and her, I would guess you call them senior engineers or like in the military, they would have been her lieutenants. They had their own programs that they worked on. Maybe not quite as many as me, but they did. And then they'd also make sure everyone else was. And then the other, the supplier I worked for, he would fly out with me or handle things when one of the account managers was too busy. So, he was forced to do the work and thus understood the work being done, right? And same with my, my girlfriend's boss. She manages a lot of the the money at this company along with her employees that also do some of the accounting stuff.

>> So, that's how it should work.

>> It seems like, yeah, the manager has to, even if it's not the same amount, half of it is delegation, half of it is making sure you do some work. And actually, the companies I worked at, every other manager was the same. If you're not doing some of the work yourself, you're not understanding the work your people are doing, and you don't know how to measure their productivity.

>> Exactly. And if you're, you just using AI to do stuff and you have a manager that doesn't really understand what you're doing, or worse still, uses an LLM to evaluate it. Yeah. You're never going to actually measure anything. All you're going to measure is burn and lines of code. So it's just an exploitation of the worst excesses. It's the, it's the worst possible situation by accident. I don't think that this was planned. I don't think anyone planned a single thing here. I think that it's just kind of the dominoes fell where they would and finally we found a technology that kind of caused problems within the world of management that were always kind of there. And I think remote work also terrified the managers because when everyone was remote, it suddenly became about, "What do you actually do?" And, uh, a lot of managers freaked out. I don't know if you remember back then, but you would read articles in the newspaper all the time like, "Remote work is bad. We need serendipity in the office." And it was always a manager, without fail. It was managers just having a [ __ ] fair because they had no reason to exist in many cases. And I'm glad to hear that there are good managers. I genuinely am, 'cause a great manager, the rare ones I've had are quite magical. And then there, there are the other ones like my editor at PC Zone, Will Poor. [ __ ] brilliant. Amazing guy, great advice, great editor. Played games, knew what I was talking about, made sure I didn't make mistakes because I was quite young at the time. And then there are other people who I've worked with outside of PC Zone. So, I'm actually just thinking of this now, like how much of this do you think

Is the uh mercenaryness, the cowboyness of a lot of these, not the lower level. Well, I guess it applies to everyone to a degree, but especially if it's mostly managers.

Because when I think of like when I worked at General Motors, you know, I went to the college Michigan Tech. Actually, my boss did as well. Um, and everyone there went to like Michigan, Illinois, Wisconsin tech colleges, and like, you know, we all went to these colleges where it was more about getting the degree, actually working. And so, if you go there, it's like you're you're doing the work, you've done the work. Even if you jump from GM to Ford, you're probably doing the same thing. Doesn't really matter.

Engineering like it's actually it's not something you can fudge.

But but it's something a sort of manager I noticed and I met some of them actually at Hot Chips once when I went there. I I got into this VIP hot wings with hot chips thing that was very very interesting. It was like Jim Keller and Roger Kadori eating wings while answering questions and and the engineers in the audience would ask really good questions. And then the middle manager markets, especially marketing people, would ask the most ridiculous softball questions that almost didn't make sense. And how much of it is like, oh well this marketing person, you'll check their bio. Oh, she works at Intel and marketing now and she used to work at Postmates. And it's like, well, why do you know how to market? Like, how much of it is the flexibility of people jumping from company to company? They're not It's like going from GM to Ford. you're you're going from entirely different businesses and you're not learning anything because you never your job isn't learning to do things. It's it's marketing or managing a team even though you never were on the team you're managing.

Yeah. I think that again it gets back to the bigger picture which is a good CEO who understood the CMO role or the marketing role or the VP whoever hired this marketing person would make sure that they knew what they were talking about. Yeah. That never happens. Especially in marketing roles. They just I for my time in PR and marketing, I will tell you the vast majority of marketing people I've spoken to that barely [ __ ] understand the things they're marketing. I mean, especially in the space. I mean, especially in games. It's got better in games, but especially in games. Game tech as well though. Some of the tech PR people [ __ ] kill me. Just [ __ ] just oh my god. And it's frustrating because tech PR could be quite an interesting role when you know what you're talking about. Marketing could be an interesting marketing challenge. I mean Steve Steve over at Gamers Nexus, he's done things with MVD engineers, like really interesting things that would be defined as PR, but it's because the PR person in question and the people at Nvidia, people will like you go, "Oh yeah, this would be an interesting thing and I'll get what I want out." And the PR person on the other end in theory could say, "Okay, we're going to get something great out of this and in return we'll give you something. We give you access to someone useful or we won't muzzle them. We'll let them talk in an interesting way." All of this requires knowing stuff and thinking and you just put the emails in the bag, mate. Just [ __ ] send the [snorts] who gives a [ __ ]

I I Yeah, I worked at three big companies. That's why I got hired. I worked at I worked at Mastercard. I worked at Postmates and I I worked at >> Gopath and now you know exactly like just random companies but they're big enough that you know the names and the amount of managers I've met from like three disparate companies who just very clearly do not know what they're talking about little bit.

See that's what I wonder if that's a lot of this problem is this new culture of engineers jump between AMD and Intel but they are doing kind of the same work managers that jump Not from Intel to AMD, which there are a lot recently, but the ones that jumped from Postmates to AMD. Why did we hire just cuz you could you I think I I'm the more we talk, the more I'm wondering how much of this assumption that they can do the same job at every company is just a false assumption, right?

It is like that's that's all of it. That's the that's literally the entire it's just that people assume that because you have because you have worked at big company and not been fired for a period of time you are capable of doing any job anywhere and the HR people and the people interviewing the person don't have the domain expertise to say >> hey is this person full of [ __ ]

The amount back in the PR thing when I used to work at various agencies I'd interview people and sometimes I remember long time ago I interviewed someone and they were like yeah I know this journalist and they wax lyrical and I DM' sorry AOL instant messaged I think the journalist I'm like hey do you know this person they were like no I'm like are you sure this is their name and like and they were like no I definitely don't I'm like you sure because they're saying you do they're like yeah I definitely don't just turn the laptop around went hey so journalist says they don't know you what's that about and they to their credit just got up and >> [laughter] >> They just got They just got their [ __ ] up, suddenly walked out the door. Honestly, I respect that. I respect that far more cuz it's just like, "Yeah, [ __ ] it. Who gives a [ __ ] [ __ ] You got me, mate." But it's it's frustrating, but it's also a lot of this comes down to people trying to survive in the world in whatever way they can, but it's it does cause problems at scale.

Yeah. But if people wouldn't let them get away with it, some of them, maybe not all of them, they would learn how to do it.

Yes. And if and if we made it shameful to do that, yes, that would it would go away. Like it but we don't because there is a certain degree of there's a certain degree of kindness, not kindness, friendliness I think, that exists in the workplace that people are scared of breaking. And also there are far more people that are just making [ __ ] up and doing nothing than there are people who are working. I maintain that across organizations. It's not something where I think all organizations are full of people that don't do any work and there's only a few that do work. I just think the vast majority are moving stuff around spreadsheets and moving people around to do the work for them. I'm happy to hear there are some managers that actually do work. That's nice.

Um, all right. Carbon CR writes in and says, "As you had keep hammering down and as is obvious from inventory numbers of IDMs, data center construction really hasn't meaningfully accelerated contrary to bombastic announcements of gigawatt projects. So is the worry about the natural resources consumed by AI data centers buildup also overblown."

I think that I think it sorry it's not overblown. It's definitely not like it it like the environmental problems are there. Even though these people keep talking about closed loop cooling, a lot of them are still using evaporative cooling AC. I know Google's definitely using evaporative cooling across the the across the world. I also think the problem isn't so much they're using the water. It's they take the water, they remove all the nutrients, they dump it back into the ecosystem, making it a problem. the heat cores, the G. A lot of these data centers are also using because of the the backlogs of generators and the backlogs of transformers in particular, they're using these old gas belching turbines behind the meter stuff.

Oh, and that is a hu like gas turbines are [ __ ] horrible. And because gas turbines, new ones are sold out 8 years back, I think they're using older ones, less efficient ones, whatever caterpillar crap they can get. And so, yeah, the environmental problem is there. And ironically, I think it's probably worse than if things were being built on time, cuz if things were being hooked up to the grid with regularity, not that I think they should be, there would be less need for behind the meter temporary power, but instead you have Meta literally standing up GPUs in tents and using and using generators there. So, yeah, it's a very horrible situation all around.

It's by >> I find it interesting how many generators they're using. And I mean, I guess I'll be a full disclosure, I just installed solar on my house, but it's because it I didn't want power outages anymore. Number one, especially while I'm recording. Uh, and number two, it is getting to a point where it's cheaper than every other form of power. So, I'm Why aren't they like cuz I Why aren't why wouldn't they be using more Some of them must be using solar, but why aren't they using more solar? Because I actually think solar is cheaper than gas and it doesn't Well, gas is annoying to deal with. Let's just be honest.

Well, I think that because solar can't produce enough power >> for their amounts. Yeah, >> for the amounts for the amounts they need. Like solar can produce a [ __ ] ton of power, but I there's something with batteries as well. I'm not a power expert, but I just know that solar cannot provide the amount of power they need or at least not as consistently as they need it.

Mhm. >> Is these things are power, >> right? Consistency can be a concern if especially if you're hyper >> intensive work. Um >> Yeah. And the land that they have needs to be taken up with I mean transmission lines and actual data centers and the halls of the data centers and the cooling rigs and all that. They run out of land when they build everything. God, what a [ __ ] mess.

Um Payeyton S writes in and says, "Hi Tom and Ed. We have seen a wave of state and local governments taking actions either for or against data centers. How do you see things playing out given the unpopularity of AI with the general public?"

going to be part of what pops the bubble. Well, yeah. And I've been pretty outspoken like anyone who thinks the amounts of data centers they say they're going to build is inevitable is uh doesn't talk to humans because yes, everyone hates data centers. And in fact, I just had Wendle from Level One Tech on and he was talking about how they're not building them remotely as fast as they said they would. Number one, >> and number two, they have entire warehouses of GPUs that aren't being plugged in. So it's like everyone acts like they've already built them. They haven't.

>> That's the weird That's the weird thing. Everyone is talking about these like they mention a a gigawatt data center like it just sprung up. Even and this is the thing that really gets me the even like Project Ria out in Indiana. That's uh that's Amazon's [ __ ] pile. That thing has been announced as fully completed as of September last year, October of last year. It isn't. It's like a quarter completed.

Mhm. But people are just like, "Ah, [ __ ] it. Put the article in the bag. Don't need to think about that too much." I think local opposition is going to be a huge problem. And it's already a problem.

>> I think it's going to make it's already stopping debt deals. Debt deals like the data center in Sale, Michigan for Oracle. >> That deal was a mess because of the amount of local opposition. They got around it eventually, but it made the terms of the deal worse for Oracle. I think you're just going to see debt deals die. I think you're going to see data centers die. Uh Mr. Dog [ __ ] Kevin Olirri he his big data center in Utah was going to be 9 gawatt apparently going to be indeterminately smaller now because of public pressure. The actual heroes out there are the people going to these meetings and raising hell. Like there was one in Wisconsin that got killed. There was a bloke who was like I've been reading Zitron. I want to hug that man. I couldn't hunt. I've been trying to find him. I couldn't. But these people that go out there and say that they have real power. And also, local officials don't like this [ __ ] And there's not really a positive spin.

>> There's not really a thing you can say, well, this is good. They don't create jobs. They create a bunch of noise. They're huge. They're ugly. They suck. No one enjoys looking at them. It's also not clear what building more of them does. Like, that's the biggest thing. It's like, more data centers. Why? Ah, AI, I guess. And if they say they're going to perfect their LLM like in a year or two, what do you need more data centers for? It sounds like you're almost there. What are you going to do with all of them?

>> Exactly. Like no one can actually answer the question of what more compute means?

>> No one. No one has an answer. And then there's the thing about the warehouse uh GPUs. I think they're in Taiwan. I think they're with the ODMs like uh Honghai, Foxcon, um Quanta, Wian, Wistron. There's another one I'm forgetting, but nevertheless, the server manufacturers for the hyperscalers, I think that what's happening is they are buying them from Nvidia, warehousing them, and waiting for these data centers to get built. And I researched this a few like a month or two ago, maybe was a few weeks ago. Time's crushing. They Microsoft has not finished a single data center started building since 2023. Every single one they've announced that's broke ground is not complete. Several of them haven't even started construction. It's actually kind of bonkers. So, I think what's going to happen is everything's just going to slow down. I don't think I think you're going to see data center bans here and there. I think Ohio just banned the tax breaks, which is going to make them way less appetizing. And I think JB Pritska >> Mhm. >> in Illinois. I think he too was he's cutting the tax break. I think this is great. You know what? If you [ __ ] want to build these things, you have to pay your taxes. You have to pay your bills. I saw someone, I forget who it was, government official suggesting a 45% extra power charge for these companies for data centers. Yeah, brilliant.

>> If that happens, by the way, the AI bubble burst too because they cannot they already like these things are already not profitable. See, because if you could realistically convince me that they're going to put like this this like multiple football fields long data center and the whole roof's going to be solar and when they turn off the data center that solar powers the grid or they're going to turn on the what the the one island nuclear and out in the northeast or they're going to turn on this one hydroelectric plant that used to be used for Bitcoin mining now it's not. We're going to refurbish it and when we leave it'll power this town. Like if they could convince me all of them would net on net do that instantly the argument is well who cares if the data centers are profitable they're building us free energy but it sounds like you're saying they're just they're just not they're not building any >> they're building themselves energy >> right that's >> like gas energy not sustainable not energy that'll last 10 years >> and they they say they are upgrading the infrastructure they say that they're going to build their own stuff but there's a lot of will here there's a lot of we will I'm not seeing a a lot of have [snorts] and I think that these companies have acted in an untrustworthy manner around this to the point that I don't trust that they'll do any of that and also all of them are using behind the meter. All of them are using gas turbines. I think a really funny and actually necessary thing would be to stop them. You can do this but you cannot use gas turbines. You can't use behind the meter. You have to build power infrastructure. And they will whine and they will cry and they will say this will slow down the progress and we tell them that's not not up buttercup like well and I want to be clear to people too because they'll say a look at this woke guy talking about solar and it's like no I bring up solar because it it just would last like 30 50 years if they built it. And these Caterpillar Generrack gas generators that you seem to be s saying they're just piling up and using those don't last 30 years. Th those will not last. Those are not they're not building a new coal plant, people. Even that we could talk about they're just using portable gas power that has no longevity. It's not when I say sustainable, I don't mean for the environment. I mean just like this could stay here for a long time and provide power.

>> It's also just a gas generator. What the [ __ ] are you doing? But also, are people really politic Well, I know they are politicizing solar, but it's like, [snorts] yeah, if someone >> But I'm just saying if someone >> Yeah, I I get what you mean. I it's if someone is trying to build something this big that is ostensibly going to replace workers, they have to provide something in return. That is the only way this makes sense and it already doesn't make sense on its face. It just none of these people are being made to do anything. So I think local interventions are going to be what stops them because good lord do these companies not have any political nuance. They have they have never been challenged in their entire lives. They do not know what to do. You've got random disheveled looking fellas out in random states who are doing more to these companies than I or Eddie Journalist combined. They are the ones who are going out there and actually going to these meetings and bollocking them just being like, "No, no, what do you mean? What do you mean? How many jobs? How many jobs? When?" Just like the things that need to be asked and the companies fold. And local officials love these companies. They I assume that they take them out for dinner several times or what have you, but they also really don't like being yelled at.

>> And I think the yelling thing is really getting to them. I think it's genuinely like I think it's genuinely freaking out local officials. I do like that we're seeing things being like, "Yeah, AI is going to be on the AI is going to be on the ballot." It's like, who's the pro AI candidate and what are they offering?

>> Yeah. Aval rating of the US Congress like and I said that a couple times I'm like is that true? And it's like oh no it is. Even though Congress approval went down AI has also kept going down. It's kind of comparable.

>> AI is like the lowest right now which is really funny.

>> Um all right I've got a voicemail here.

>> Hi Tom and Ed. Finally amazing to see you both doing this. So the elephant in the room the straight from is closed off and oil execs are warning in next two months things work very bad. Even as the Donald mentioned that may may reopen by Labor Day feels to me this is going to cause a stock market crash which will 99% sure pop the AI data center stop bubble as soon as this year. Any likelihood to this? Also to add there's the 18-year real estate/line cycle theory out there and we are due for it causing a crash on its own this year since the last one was in 2008. it would it would be absurd for it not to pop any circumstances by later this year and no longer especially as I think the strip closure could last in my view later into the year.

So, yeah, he's basically asking, you know, and and I've think of wondering this too, like the intersection of like the closure of the Strait of Hormuz, which I do not think will be fully open tomorrow, you know, and like I feels like everyone's acting like gas will go back to normal tomorrow when even if it was open, it's not going to probably for like 3 years actually is what most experts say. But how much do you think that could influence the popping of this bubble as well?

So I think the whole moose situation, I will be clear, I'm not an expert on that.

Um, I think it's got right now it's in the rubber band stage. Like every 3 hours we get a story. It's like Donald Trump says Iran is about to open it up. They're going to give us $70 billion each. It's going to be beautiful. Then two hours later it's like Iran is never opening the strait. They're never doing it.

And we live in this kind of weird thing where the markets are aggressively avoiding it. They just they're not thinking about ignoring it. They're ignoring it. eventually something I don't know what it would be something is going to rattle them something.

>> well they think our oil stores in a lot of countries are almost empty right maybe it's that >> yes when that becomes impossible to avoid that's going to genuinely be that's going to be what shakes them but it's I think the AI bubble bursting is likely to be a combination of factors that's that kind of leads to the stock market tanking in a big way like it's just going to be like maybe this I don't know if the SpaceX IPO goes poorly Like I I truly don't know. I think Broadcom's [ __ ] ass earnings, even though they weren't terrible, probably didn't help. But it's once people stop believing that this [ __ ] is inevitable in the markets and the markets stop dumping, that's when people will start saying, "Oh yeah, is AI any any good? Really important? Does this help?" But what's really going to do it is data center debt. And the the caller kind of made this point where it's data centers are going to be the thing that moves it. And I think it's going to be when the bankers start going, "No, no, I'm done. I'm not giving you another $3 billion for GPUs. You're [ __ ] crazy." And eventually, it's going to be the pressure on the hyperscalers because the I think like 54% of Nvidia's revenue came from three clients last quarter.

>> One of those clients is eventually going to go, "Not for me, thanks." I think it's Microsoft. M

>> I I [clears throat] think it's Microsoft or Meta is the most likely one. I think Meta is going to keep [ __ ] that chicken a little longer, but Microsoft is the one where Amy Hood, the CFO, has been got a target on Sachin Nadella's back for a while. But it really is just it's going to be a slowdown than a pop. It's not going to be like one big aggressive thing that convinces everyone. going to be data center debt stops being issued and then the rest kind of flows from it cuz Google just signed a massive deal with SpaceX to begin October 2026. So may may as well be in the year 3000 at this point to rent a bunch of GPUs which suggests that data centers are just not getting built.

>> suggest that they're that they that they're just nothing's happening. Why else would you buy from Elon Musk? And so I think it's just going to be that new debt is not issued. Nvidia cannot sell as many GPUs. Marvel and SanDisk and SKH Highix and all them don't get the revenue they expect. Everything kind of starts shaking apart. I mean a large chunk of Korea's future tax revenue is basically coming from Samsung and SKHEX and they are budgeting based on the idea that these companies will be able to pay it. If that doesn't happen, that's an economic problem. I mean the KSPI the Korean stock market shhat its pants today as well. It may be what we're watching I mean this will come out after this day so who knows but what we like today Friday uh June 5th they are the stock market is tanking and no one's really sure what to do about it because no one has a good story. Nvidia's already done their earnings got a couple months more from that. We're going to have Oracle next week that's going to be bad. So, it's kind of hard to answer the question with specificity, but Hormuz could also just do something. A country could run out of oil and that might really [ __ ] up the various economies around it. America is somewhat insulated cuz a lot of stuff comes from just in America. It's going to hurt that. I think Europe is where you're going to see it as well. Jesse here really can't take it when she sees rabbits out in the wild. I'm not kidding. She wants to chase them so much that she starts visibly shaking while she stares at them. And honestly, I'm kind of the same when it comes to overpriced Microsoft software. I can't believe that some people, especially in these times where many component costs are inflated, would waste hundreds of dollars overpaying for Microsoft Office or Windows. And you don't have to if you go to cdkeyoffer.com. That's right. cdoffer.com is a sponsor of this piece of content and they are I believe the longest running main like large sponsor for Moors laws dead going all the way back when I lived in a different state and that's because they provide reliable keys for Microsoft Office professional and for Windows 11 and you know they also have an offering of a ton of different games that you can get on a discount as well and if you want to support the channel then go to cdoffer.com today through the link in the description just clicking on that link helps but then also So get yourself these keys if you need them and save by using offer code broken silicon to save 25% on all Microsoft software and then die shrink to save 3% on everything else on the website. Once again, thanks to cdqoffer.com for sponsoring yet another maybe what a thousandth piece of content that they've sponsored and thank you all for supporting a sponsor that supports Morslaw is dead.

All right, so I'm going to read two questions because they're really kind of just leading to the same overall thing I want to ask. But Karen Nougata writes in and says, "The following is my current best guess as to how the AI bubble will burst. First, general purpose AI will be fundamentally too expensive and providers will no longer afford to subsidize the costs. So are moving to per token billing or similar. And then for the general user, this will make AI too expensive. On top of both costs and results being unpredictable and then those users will try to move to providers with subsidized and predictable costs. When all providers though in a domino effect all have to stop subsidizing cost, demand drops off significantly. How plausible do you think this is? What other alternative scenarios do you have?" And then Valcomv basically just asks, "Are we sure it will be a crash or will it be a temporary slowdown? And then when's your best estimate of when the bubble will pop? Which I also have to ask. I know this is a thing that's impossible to ask with 100% certainty, but come on. We're on a podcast. Let's pretend we know everything like the everyone else does. When do you think the bubble will pop?"

by when >> I think by the end of the year. I think it's hard to say when cuz it could be next week, could be in four months, could be in a year, like whatever. I think that what they're describing I've me I've called this the rats rats in the sinking ship situation where it's going to be oh well I can't I can't use this plat. Well, I mean when GitHub copilot went to tokenbased billing the tons of posts in the Reddit were like yeah what's another platform I can use? What's another one that will let me do? And they it's like the it's like the dinosaurs looking at the meteor and being like if I just run that way I should be fine. And I think what it is is you're going to see like the like the question said yeah people are going to jump to one platform. They're going to overwhelm their cost. That platform is going to have to go to tokenbased billing where they're going to run to another one. I think anthropic is the last man standing on this one.

>> Do you think OpenAI is the first to go?

>> Oh yeah.

>> See cuz I do I don't know about anthroic. Uh, I think you can see Gemini maybe ekking out something to make search more efficient because their incentives are different than Well, I'm just think >> I think I think everyone's in the world pants [ __ ] in competition. I don't think they have any plan. I think Google's Google's AI search is awful. I mean, I think and this may or may not be because I've seen some stuff that might really, really, really reaffirm this belief that I'll be breaking in a couple weeks.

>> I think [clears throat] OpenAI is a broken company economically speaking.

>> I >> agree with that. I think they're the most exposed by far, right?

>> I think I actually think Anthropic is just as bad. I think anthropic is just better at manipulating the media. However, due to some things I may or may not know. Yeah, I think OpenAI is a dog [ __ ] sandwich. I think that they are just burning capital. Sam Wman the other day said, "Yeah, at the beginning of the year, no one really cared about cost, but now they think it's a huge issue." It's like >> they What do you mean by what do you mean by that, Sammy? But the other thing is with Anthropic as well is both of these companies have to keep growing. They cannot just >> stay at whatever they're at right now. They're both unprofitable. They cannot keep they cannot slow down. Anthropic is projected to make $174 billion by the end of 2029. OpenAI $284 billion by the end of 2030. That is the only way they will meet their compute commitments. Anthropic I think is 300 I think between them they have over $1.1 trillion worth of commitments. Anthropic at least 300. OpenAI claims 600 but it's over 700 if you actually add them up which no journalist bothers to do. They cannot afford these commitments which they have made. They have signed the contracts. They are part of the the remaining performance obligations of multiple companies. They cannot afford them. They cannot afford them. They just can't do it. So unless they become like what they are at now, however amazed or not amazed you are right now is insufficient.

Mhm. compared to what they had promised. They cannot they cannot survive if they do not become 10 times the size of what they are right now. And I really I'm not using hyperbole. This is not >> sure I mean what Nvidia is worth like $8 trillion now. That doesn't make sense to me. So I don't think there is hyper anymore.

>> They will also have to make more money than Nvidia. like, oh, and also for Nvidia >> to keep growing at the rate that makes the stock brokers hard, they need to be selling a4 trillion dollars worth of GPUs. Uh, sorry, quarter trillion per quarter in like a year and a half, I think. And that's just this is just to keep up with the growth rate, otherwise their growth will slow. And to be clear, they could be making 150 billion and the markets will be mad because they're meant to be making a quarter trillion. And again, they can only do that if the GPUs that you get [ __ ] installed somewhere, which no one has done. Like it's it's all really weird when you say it like this. It's all really weird, right? And I think it's just a confluence of everything coming together. Were the stock market to not have all these externalities from like the straight of four moves and all these other things. Maybe you could see them eek it out. If they were actually able to plug in the GPUs, if people actually kind of liked AI and they weren't protesting. I mean, because I personally think just straight up politically people aren't going to let them build any more data centers just >> I think people people are going to obstruct them >> then it's like there there's there's nowhere for them to go but so so you think by 2027 so by January you think this thing's already deflating?

>> Yeah. I think that there might be a story that comes out before their IPO about OpenAI's numbers that looks so much worse than people realize. Maybe the losses are tens of billions versus >> you actually think because I saw the S&P 500 is not going to allow them to like they change they're not going to change the rules.

>> I think they're going to try. I think Anthropic is going to try.

>> I think both of them are going to try some really loony math. I think they're going to try some community adjusted EBIT action. I think it is going to be a race to see who files first fully like as in not confidential filing who actually gets to the stage of IPOing. And it's weird that OpenAI has not filed yet. Maybe it's because their numbers are too bad and they know that they'll get crushed. Maybe it's that Anthropic is more confident in their numbers. I don't know >> because it's all about the narrative. And I could see Anthropic with shitty numbers getting away with it. I I don't think OpenAI can. I think people hate >> Clammy Sammy's Samuel Orman is washed. He doesn't have the juice. Dario Amade is a master manipulator. He's really good at manipulating the press. really good at manipulating business idiots, really good with stock people as well. I could see him shuffling this turd onto the market and it dying a death. I think Anthropic could be dead in a few years. I think Open AAI might copy it in the next year. I think that their real numbers come out. I think that people people genuinely see when those numbers come out that OpenAI looks like [ __ ] that OpenAI is just a terrible company and then go, "Well, [ __ ] I would be really shocked if Anthropic wasn't much worse because Daniela Amade said the other day that she that they need the public markets to raise more capital for training >> and it's like >> that sounds just like Open AI. Yeah, >> sounds like exactly the same company. Exactly the same. And yeah, I don't imagine their numbers are going to feel much better than open AIS. And I get the feeling that open AI I get the very strong feeling that open AIS are going to look very very bad.

>> Carbon cry writes in. He says, "What will be Amday's next bombastic PR stunt universe is a simulation like from the show Devs and our AI threatens to break it if Mythos 2.0 is released? Will fish fly?"

So, and and I put this because we and you were talking offline before we started recording that like Jensen Wang for me has always been one of the least skevy and that's just because of like the top Elon Musk. I mean, it's quite a competition to be the most slimy and weird. And yet, when they revealed N1X or RTX Spark, the APU for their laptops, which I I've been hearing for years is be delayed and delayed and delayed because of how many bugs it had. like the way he talked about it for the first time. I'm like this guy sounds like Elon Musk now because he was like he didn't just say this is going to be an efficient APU cuz it's expensive and we spent a lot of money making it. He's like this is the like when the smartphone was unveiled. This is the PC. this is the new like he acted like they invented the PC and it turned into the reason we'll make so much money and I believe this might be the next PR stunt is isn't that a billion people are going to buy their AI PCs that cost $10,000 and he said this because everyone has their own private theater. He actually said that and then like realized that sounded out of touch, but um he said it's because you're going to buy these $10,000 PCs and then they're all going to have three of their own $10,000 PCs that they are telling what to do. And that is why the market is actually going to be billions, not millions. He actually that was his pitch. Wow. If I said that on the street, they'd institutionalize me. It's amazing what money can do with a person. if someone's got enough money that you'll believe whatever asinine [ __ ] they burp out a [ __ ] quit. That's the thing like he could have framed if if I had have heard that he suggested that this was Nvidia taking over the PC industry I'd kind of believe it. I feel like maybe I don't know >> if this was him basically saying this is our equivalent of max silicon of like the M processors fine I'd be like sure this is their attempt now that they've starved the industry of as of resources now they've driven up the price they will come in and swoop in and eat everything but no it sounds like he thinks people are going to buy PCs to do AI on when did he last talk to someone who didn't make >> $100 million Yeah.

>> When did he last talk to someone that he wasn't screaming at? Like

>> Well, well, he's talked to many people that aren't billionaires, but they're always um asking them to sign their their jackets and stuff.

>> I saw a video of a guy who handed Jensen his wallet to sign and Jensen took the money out and handed it to some woman and it was the f it's like I posted this in Twitter. It was like, yeah, the quite literal wallet inspector. like just you you are being you are being yoked, [ __ ] It's just really depressing because with all their resources, Nvidia could do something cool, but they're doing this instead. [snorts]

>> Dark helmet writes in, and so does AI generated fish, and they ask, and by the way, my fans like to put AI in all of their names to >> go along with the everything has to be AI thing. Do you think Nvidia's recent push into client devices with part uh RTX Spark and their work with SPAN to install many PC data centers on people's homes as part of their hedging strategy in case the data center buildout continues to fail to proceed at the expected rate? How much do you think those efforts could offset any slowdown? And then a generated fish asks if the bubble pops to use the gold rush analogy, do you see a scenario in which Nvidia is stuck with unsellable shovels?

>> Yes, I mean they already they're going to have to write off inventory. that's going to happen. Happened with the H20, it will happen again. I think that I think that there's a good chance that some point Nvidia has to file some sort of filing with the SEC and say, "Hey, we have a bunch of [ __ ] we can't sell." Like that's I think we'll probably see it happen with Super Micro first or Dell, but there's going to be a point when they realize they have bought too much crap and they have no one to sell it to. Maybe we'll see it with the ODMs first, but I think that

>> I think the problem is is it's really hard to point to where the bubble will burst because a lot of it will come down to willingness to disclose and Nvidia is going Colette Crest the CFO over there. She's going to she's going to do everything she can to delay anything she has to do like that. And I think that maybe we get another earnings or two with Nvidia before that happens. Maybe they're forced to quicker. Maybe something happens. I also think this Chinese GPU story keeps popping up and everyone keeps aggressively ignoring it. The markets are just like, "Nah, I couldn't possibly." There was the short seller that claimed that Nvidia 20% of their fiscal year 26 revenue came from China. Nothing happened. Nothing. Elizabeth Warren wants to invite him Jensen Wong for a [ __ ] chat, but the media just like n you know,

>> I've been thinking that too. Like cuz >> nobody's not aspicked.

>> Even if uh like Trump allows Nvidia to sell to China, China's like, "Oh, we don't want them actually. We want to make sure that we compete with you." And it's like, isn't that like a substantial part of their revenue?

>> Well, it isn't anymore. But also, we very specifically, very, very specifically have outlawed China from buying these GPUs. However you feel about China.

Mhm. >> We that that was illegal the whole time they did it.

>> Yeah. if they did it. We We don't have proof. I I had some other people.

>> It seems it seems very obvious that they did.

>> The joke I make on my podcast a lot is um you know like Yeah, those gamers in Singapore that need four 5090s. I mean, come on.

>> There's just no there's just this guy there with like a [ __ ] MVL 72 of GB200s and he's just running like 18 chrome tabs and crisis.

>> Yeah, that's what he's just [ __ ] and he's he doesn't need heating in his house. his house just like 100 degrees at all times. He's nearly dead constantly. I love that guy.

>> All day.

>> All right, so this is a question that I also have to ask here.

>> Um, how much of Nvidia's AI problem, and hear me out, is a kind of a good problem to have. Like I looked at their valuation versus AMD and it's absolutely absurd. Like they can buy AMD several times over. I think it may have been six times over. If the AI bubble pops and everyone loses value and Nvidia loses two, three, four times, like they're down at 25% of their value, as long as they're still the biggest swinging dick in town, will they really care? I mean, they'll care. They don't want to be worth less, but will they really care if everyone drops five times and they're still the biggest company and they made all this money that they can just throw in a war chest and and and kind of connected. Do you think AMD can maybe a bit more easier than Nvidia pivot at the end of this? Because everything they're making or most for AI like I've just leaked all the stuff about RDNA5. This like big GPU that's really powerful. It can be used for gaming. It's also a gaming chip. The chiplets are shared between the nextG Xbox, their nextG APUs. Literally like the same thing they will put in a laptop is the same thing they'll sell to gamers. It's the same thing they'll use in an AI thing. So, do you think AMD is positioned in a way that they can maybe more easily pivot their production to consumers who are starved for stuff right now because of AI? And do you think just in general some companies will be less screwed than others?

>> I think that Nvidia like Nvidia is not dying at the end of this. I just think they take an 80% haircut.

>> Yeah.

>> I think that Nvidia's biggest problem I don't think Nvidia will die. I'm not I used to say it was impossible. They have made like $85 billion or more worth of commitments to TSMC. Like they have to pay TSMC whether they got the money or not. And there just Daario, a fantastic uh writer, he found this. Their cash flow only increased $600 million last quarter after making like $82 billion in revenue because they've sunk all their money directly into NeoClouds and other companies and investing in OpenAI and Anthropic and all this [ __ ] So, I don't know. I think that AMD is probably best to weather a storm just because,

>> okay,

>> they've yet to really suck on the tee quite as hard. I think Broadcom is going to be in real trouble. I think Hawk Tan is a real [ __ ] Like just a real weird guy. Like a real like it seems quite unpleasant, but also they are heavily heavily I mean they're the reason the semis are selling off because their guidance was not high enough. So if and they were guiding to hundred billion dollars of revenue. So if that's not enough and things slow down from here, uh-oh. So I think AMD's in a decent position. I think Intel Lip Bhutan I think is probably put him in the best position possible. Can't get any [ __ ] worse than how it was a few years ago.

>> No,

>> I think that and I think that if timing works

out for Intel, they could actually take advantage of this to catch up [clears throat] because everyone else has been far up their [ __ ] with AI, but also Intel is also super AIDS. So, who knows?

It's just >> Well, they want to be AI pled. I've covered them. They they just keep failing to make any product that's compatible. Did you read that like was it Pat Gellzinger he was going to invest in open AI or something like one someone at Intel passed an open AI which [ __ ] can you AI bubble would have ended in two months just like chat GPT would never have they just were like yeah we tried to do it didn't work sorry like just we never would have there's an alternate world where the AI bubble just never happened.

You know this is funny though because it seems like you do agree with me because I you know I will say in the gaming space right now. I've never seen more people blackmailed. And I deal in like leaks, rumors, covering stuff that comes out the next year and they're like, well, why will I why should I care? And I'm like, yeah, I don't know. I mean, Nvidia's bet everything on AI and that a lot of the products like I almost wonder if they're skipping the generation that was supposed to come out late this year or next year.

I thought they were I thought they were already saying they were going to skip that.

Oh, I don't know. Maybe they've already admitted it that they're just screw it. We're we're skipping it. or the new Blackwell is going to be a four-year gaming gen and you know this leaves AMD in an opportunity here because all their stuff is shared like once they can get RAM they can just go and they're supposed to next year make PlayStation 6s and Helix and laptops and and so can Intel and everyone will be desperate to buy and I do wonder if there's an opportunity there where Nvidia's really pissed off their biggest fans. I know it's a rounding error, but it won't be someday for their earnings where AMD has the product that Nvidia just said they would skip and Intel has a product Nvidia can't compete with. And maybe there's actually a a pretty great consumer year in 2027.

It could be. I don't know if it's 2027, but it could be a weird reshuffleling where Nvidia is selling the main processor for laptops and Intel has moved into a gaming position. Don't know how much hope I have for Intel in that, but maybe AMD comes back. But the thing is, I worry about Lisa Sue now. I cannot trust her.

I just seeing the last six months of AMD, I have serious doubts that that woman is suddenly going to have a somewhat some genius moment when she's like, "Oh, we need to change the But you know what? She's turned around AMD before."

Yeah. And I've heard kind of behind the scenes that she's a little annoyed by AI or at least it's an impression I get. Obviously, she's doing the obligatory, you know, whatever you have to do to get the investments, but I feel like when she's on stage, there's a little bit of a rolling of the eyes. Maybe it's maybe it's just an impression. I I don't see and she's not saying we're going to [laughter] have your PCs own three PCs, so she's not doing that.

I know, but the thing is she's still she's I just It's the open AI deal. It the open AI deal really just made me like I I cannot trust the web.

I was pretty surprised when they did that. And I was like, "Really? You're doing this, too?" But >> no, when I [snorts] definitely was quite pissed off about that cuz I was like I had up until that point been like AMD, they're doing a little bit now. They're paying they are paying Cruso $300 million to rent back their own chips. Like they they are the same as Nvidia.

Mhm. So yeah, it' probably just be luck like they just happen to have the products in development that >> it will be that they're just less leveraged. Like that's what like Nvidia is heavily leveraged in this. AMD, Intel will be less leveraged. Plus, they will have more flexibility. I just don't know what happen I'm I'm not a super fab expert here, but I don't know what happens to capacity when you can't pay for it. I truly don't.

All right, final question here by generated by Jensen GPT, who asks, "What consumerf facing generative AI technologies do you have a more positive outlook on?" A minor one for me is eBay letting me autofill product details from simply taking a picture when you're selling. The end result is almost always spot-on. Saves me a bunch of time. Otherwise, I'm generally pretty bearish on generative AI. I just don't see the average consumer wanting to spend a large amount of money on AI hardware so they can chat with their phone or smart devices rather than just using them as already existing and already easy ways that we've been able to do for a while. And I think I just also want to take this to add on, what do you think happens to like what lasts? What AI things do we end up still using in a few years? Like what gets built back up and maybe goes into generative direction again? like how how does this postbubble world look like and is there some good that will be left?

So start with the one use case I have for LLMs. The only use I've had for them is I think once I used Claude to drop a PDF in like a I think it was a 200page document be like can you find anything like this in there and then when it did so I opened the document myself and went to the pages.

Mhm. Like that's that's it. Like that's the thing that >> Gemini Pro did that for me with a solar uh station too and then and chat GBT argued with me that it wasn't in the document and and Gemini knew exactly where it was. I'm like, "Oh, GPT."

I love that you can't guarantee that they'll do anything. Um I think that there is a chance that Google's search hangs around with some LLM bits. I think there is a chance that there is a shell of chat JPT that exists, but right now every AI product is losing money.

Yeah, >> every single one. Google, Meta, Amazon, what have you. I think Meta is the last man standing, too. I think that they have a good chance of it. I think that Mark Zuckerberg is a mad king. Well, he can't be fired. He literally can't be fired due to the board him owning the controlling board seats. So I think that what survives could be either very little, very expensive. I think anthropic and open AI could be completely wiped off the map. I think that they could exist as arms of Amazon, Google or Microsoft. I also think that a lot of this AI stuff will just go away. I think there's a good chance because it's expensive. It's annoying and users hate it. Like there's there the people that like AI >> are very loud but remarkably minor. Most people who >> like AI are just like, "Yeah, it's fine. Yeah, I use it for this." Like, no one is sincerely excited by it.

They're like, "No, like other than people who sound like they have gas leaks." Like, it's always someone very craven who likes them. Other than the few coders I know who are super good coders who use them for like, "Oh, there's a script in one language that I use once, but this one thing." And that's how they it's always described in this very minor computer sciency way like software engineering is pretty much always they will take every advant I say this with love. Software engineers are data perverts. They they find whatever little edge they can find. They'll get whatever bit they can tweak to get a little bit more efficiency here or there and and everywhere. I think that that's a good thing. I think that that's cool that they're able to do that. I think they will use ondevice models. I think that there will be some degree of leftover hardware that they can adapt. I think that LLMs as we know them will not be prevalent at all. I think LLM services will be quite expensive. I think that there will be massive impairments across the tech industry. And I think that in the end we will look back on this era as the largest misallocation of capital in history and also it will make it impossible to trust certain people. I think if someone has been fully AI pilled, no matter who they are, there is no way to trust that person ever again. It's just there's no way to do it. It's just how would you possib someone who told you that AI is the future, how can you trust another word out of their mouth? How it's been like it's got marginally better, but it's not become anything close to what they promised.

On that note, I think we will end the conversation. I think that was a perfect way to end. Please, Ed, tell everybody where they can find you. Where's your edit at? Subscribe to my premium newsletter. It's how I make my money. It's how I survive. It's how I get I pay for for diet cokes of which I drink 11 a day. Listen to the podcast better offline.com. bt t e r o ff l i ne.com. Also on YouTube. Yeah. Listen to my stuff. Read my stuff. Thank you for having me. It's been fun.

Yep. I'll put those links down below. And of course everybody as usual like, subscribe, share, hype points, comment, join the Patreon to ask, ask questions, get access to die shrinks. One just came out about uh what defines a console generation. Quite a different discussion than today, but we have an eclectic set of things we talk about here. So if you're interested in that, join the Patreon. And otherwise, everybody, thank you for watching. Thank you for listening. Once again, thanks for joining me, Ed.

Thank you for having me. This podcast was brought to you by the YouTube channel and website Mors Law is Dead. Mors Law's Dead and Broken Silicon are trademarked by me, the creator, Tom. However, Mors Law's Dead is much more than just one man. It is a team with Broken Silicon co-hosted by my brother Dan. Audio editing by Gerard Cortez. Renders done by the industrial designer Jean Phipe Clare Maul and special assistance also provided by Carbon Cry and Carrie Nosugata. Find all of our information at www.morslawsd.com morslaws.com on the about support page in the event you want to hire me for consulting work or hire Gerard for audio work, Jean Philippe for industrial design work or interested in working with Carbon Cry or Carrie Nosugata as well, you can find all of our information there. You can also find our long-term sponsors there and show them some love for helping make this work possible. Additionally, if you are a fan and would like to mail Mors Dead some love in a letter or with hardware donations, please mail parcels to Moors Law Dead, PO Box 60632 in Nashville, Tennessee 37206. And speaking of fans, patrons are what makes MoSaw Dead content possible. Every month and every day actually depending on the person. Me, Gerard, Dan, Jean, Philipe, and more are working tirelessly to provide a steady stream of entertaining content and investigative journalism that aims to be as independent as possible from outside influence. Only direct funding from fans can make this possible in the modern era. And so, if you have any extra money every month, even just a dollar or two, consider supporting us directly on Patreon. All members of the Mors Lazad Patreon, even down to the $128 128 bit tier, get access to a catalog of dozens of essential die shrink episodes and the ability to ask guest questions ahead of Broken Silicon recordings. Additionally, higher tiers get access to the entire catalog of hundreds of episodes of Die Shrinks, and these include exclusive interviews with game devs and other guests, behindthe-scenes looks at Morslaw as Dead Leaks, and more niche coverage of subjects that really only the biggest gaming dorks like me and you would care about. They would not profit on YouTube, but they don't have to profit on Patreon. And so if you're interested in that, you can get access to all of that stuff and even more at higher tiers like ad free episodes of Broken Silicon, the ability to ask free questions and loose ends live streams ahead of recordings, thanks in the credits of videos and podcasts, and access to the Discord with thousands of fans there who would love to discuss gaming hardware with you. I am one of them. And hey, if you cannot afford to support us, please just share more lives and videos and podcasts with friends and family and on social media and Reddit. and give Broken Silicon a fivestar review on Apple Podcast or your preferred podcast app. All of that stuff does help a ton as well. But as I said, this podcast would not be possible without its patrons directly providing predictable and reliable support every month. And so now is the time to give a personal thanks to the greatest of the fans. The following supporters are at the 1,24-bit HPM3 or higher supporter levels. Brad Medlin, Zj Jits, Daniel D, Nicholas Buckner, Deak, Aaron Close, Jan Rer, MJB1, Daniel High, GZ, Ziggy, SNS, Charmers, Jan Farri, Valcomv, Jensen Wang, Sarcastro, Chris Rich, Evan Dingle, Halbuma, Shredder, Sammy Moss, Chris Anthos, Jonathan, Tommy, Matthew Cassell, Jake 223, Jake Martins, Zlicki, Win, Gamers Unite, Ian Clifford, Meat and Pork, Tim Rog, Travis Gooding, Zutsu Taylor, Mads, Greg, Michael McGee, Stephen Coats, Mark Mitchell, Amy Will Chief, James Anderson, Mark Raid Maker, Dan M, 1988, I should, Salty Internet Pirate, Henry Jen, Wesley Sier Cameron Michelle Pel Hexapuma Tik Autumn Little Germany Winstar James High Raiders Shay Milton Brett Jones Chuck Glidden Steven Dick MD Daffy Justin Busel Austin Hagerty Joe Brian Wright Brian Scanland Harland Slesbot Johnson the Forbidden Juice Bing Bong 3 John Swin Solarized 80 Kum Sputnik Brian P Smash Axley Corner 671 Pentoa Omega Doge Hugo Ander Kevi John Wick Boomstick Mr. Tiger AI Soldier 789 Penguins Love AI Porcelain Oats Dominic Paceri Kim El Mazri YG BS Manoskus Kevin Coun Jonathan Harris Bespoke EP Kuse Johnny Trans Fred Fredberger uh Zubir Ahmed RKO RK0 Hunter Temple Dudster Henry Day John Tiberius OMG Clover Roadkill Dog Rebellious Pilot Joseph Renfro and of course thank you to Sahara for the music.