Better Offline - The AI Bubble Is Bursting

Episode Date: April 12, 2024

Every major tech firm is betting billions of dollars that the generative AI revolution will change society - yet when you look under the hood, the reality of generative AI might be far grimmer. Ed Zitron walks you through the many signs that we're on the verge of the AI bubble popping - and what the consequences might be if it does.See omnystudio.com/listener for privacy information.

This is an IHeart podcast. Guaranteed Human. Run a business and not thinking about podcasting. Think again. More Americans listen to podcasts than adds supported streaming music from Spotify and Pandora. And as the number one podcaster, IHearts twice as large as the next two combined. Learn how podcasting can help your business. Call 844-844-I-Hart.

Another podcast from some SNL late-night comedy guy, not quite. Unhumor me with Robert Smygel and friends. Me and hilarious guests from Bob Odenkirk to David Letterman help make you. you funnier. This week, my guest, SNL's Mikey Day and head writer, Streeter Seidel, help an a cappella band with their between songs banter. Where does your group perform? We do some retirement homes. Those people are starving for banter. Listen to humor me with Robert Smigel and friends on the I-Heart Radio app, Apple Podcasts, or wherever you get your podcasts. Your 20s can be so exciting, but they can also be really overwhelming, confusing, and honestly, just kind of lonely.

May is Mental Health Awareness Month and the psychology of your 20s is breaking down the science behind the biggest roadblocks we face. I was six years into my career, the 80-hour weeks and just the first one in, the last one out, and I ended up burning out. There was a large chunk of my 20s that I was just so

wanting to be out of that phase out of my skin and I just really regret not living in the present more. You don't need to have everything figured out right now. You just need to understand yourself a little bit better. Listen to the psychology of your 20s on the IHeartRadio app, Apple Podcasts, or wherever you get your podcasts. AllZone Media. Hi, I'm Ed Zittron and welcome back to Better Offline.

As I've discussed in my last episode, there are four intractable problems that are going to stop Generative AI from going much further. Its massive energy demands, its massive computation demands, its hallucinations when it authoritatively tells you something that isn't true or makes horrible mistakes in images, and the fact that these large language models have this insane. sensationalable need for more training data. Yet I think what might pop this bubble is a far simpler problem. Generative AI simply does not deliver the magical automation that everybody has been fantasizing about. And I don't think consumers or enterprises are actually impressed. A year and a

half after launch, it seems like the kind of immediate and unquestioning infatuation with Chabbt GBT, and for that matter other generative AIs, has softened. Instead, there's this rising undercurrent of apathy and mistrust and, of course, failure that's kind of hard to ignore. In June 2020, traffic to Chat GPT's website, where people access the ChatGBTGPT bot in a web browser, fell for the first time since launch, starting a trend that's continued for five of the following eight months, according to data from similar web. People are becoming more aware of the technology's limitations, like as I mentioned, hallucinations, which, as I know, is when ChatGPT confidently asserts things that aren't true,

which can be in writing when it gives you an incorrect fact, or in an image when it gives a dog 18 legs. To make matters worse, according to data from data.AI, which used to be known as App Annie, chatGPT's downloads in iOS have begun to drop from a high of just over 700,000 a week to a plateau of around 450,000 to 500,000 a week, since early 2023, which sounds impressive.

until you hear that only 7.35% of people who downloaded ChatGPT in January 2024 actually used the app again 30 days after they downloaded it. Cratering from a high of 28% a month after the app launched in June 20203. In fact, things immediately appeared to have fallen apart. In July 2023, only two months after launch, only 4.59% of users open the app for a second time. numbers like these tell the story of a buzzing new application that isn't actually providing users with much utility. I think the generative AI engine has started to sputter

for customers, for businesses, and indeed for the startups that create them. That's bad news for any industry that's yet to reach profitability or indeed sustainability, and especially for generative AI, which remains relying on an indefinite supply of cash to operate. Back in April 2020, Dillon Patel, chief analyst at semi-analysis, calculated that GPT3,

the previous generation of Czech GPT, current ones known as GPT4, cost around $700,000 a data run. It's about $21 million a month or $250 million a year. In October 20203, Richard Windsor, the research director, at large, of Counterpoint Research, which is one of the more reliable analyst houses,

hypothesized that OpenAI's monthly cash burn was in the region of $1.1 billion a month. Based on them having to raise $13 billion from Microsoft, most of it, as have noted, in credits for its Azure cloud computing service to run their models. It could be more, it could be less. As a private company, only investors and other insiders can possibly know what's going on in OpenAI. However, four months later, Reuters would report that Open AI made about $2 billion in revenue in 2023, A remarkable sum that, much like every other story about OpenAI, never mentions profit.

In fact, I can't find a single reporter that appears to have asked Sam Norman about how much profit open AI makes, only breathless hype with no consideration of its sustainability. Even if Open AI burns a tenth of Windsor's Estima, about $100 million a month, that's still far more money than they're making. There's not a single story out there talking about them making a profit, and I don't think they're making one. Here's one thing we can be certain of, though. Things are getting more expensive. Progress in generative AI means increasingly complex models,

and as I previously mentioned, OpenAI attempted, God damn it, a Rackus model, one built specifically to wow Microsoft by making ChatGPT more efficient, failed to actually make it more efficient. Their attempts to make this a better company are not working. Windsor, the aforementioned analyst, in a separate blog, also pointed out that there's nothing really sticky about these companies. There's nothing stopping someone from switching from, say, chat GPT to Anthropics Claude 2 model.

They're all trained on similar datasets and they all produce very similar answers. And while one model might be better one thing than another, they're fundamentally very, very similar. There's also nothing stopping someone from simply giving up on generative AI altogether. it doesn't seem to be the plug-and-play automation god that everybody's been repping it to be. And judging by the plateauing chat GPT user numbers, I think that might already be happening. It's also important to remember that while generative AI is shiny and new, artificial intelligence is absolutely not. And over the past decade, it's found a number of homes from expensive security apps

that detect when a hacker is trying to break into a corporate network to spam filters, proofreading tools like Grammally, plenty of things, even Siri on your iPhone. In these contexts, AI is either a small component of a larger product or something that directly builds on human efforts. This stuff is actually valuable. AI-based spam filters are typically better than those reliant on hand-coded rules, for example, but it's also, from a marketing perspective, kind of boring. Generative AI is a law is that it can supplant humans, either partially or entirely,

producing entire creative works that otherwise would have taken hours and carried a real financial cost. But behind this glitzy technology and media hype, the unspoken truth is that generative AI holds sway over the financial markets because it's regarded as a tool to eliminate an entire swath of jobs in the creative and knowledge economies. It's a ghastly promise and it underpins the vast market value of otherwise commercially unviable generative AI companies like OpenAI and Anthropic, and it's what is driving, I believe, the multi-billion dollar investments we've seen from Microsoft, Amazon and Google. Yet I see no evidence of mass adoption of generative AI, and my research suggests that enterprise

adoption, which is the meat of what would actually make these companies' money, it just isn't there. Deep within the earnings reports and the quotes of every major cloud provider claiming that the AI revolution is here is a deeply worrying trend. The AI revenue really isn't contributing to the AI revenue really isn't contributing to the company. much to the bottom line, outside of vacuous media coverage. And I think the internal story is going to be much bleaker. Another podcast from some SNL, late-night comedy guy, not quite. Unhumor me with Robert Smygel and friends, me and hilarious guests from Jim Gaffigan

to Bob Odenkirk, to David Letterman, help make you funnier. This week, my guest, SNL's Mikey Day and head writer Streeter Seidel, help an a cappella band with their between songs banter. There's the worst singer in the group. The worst? Yeah. Me. Is there anything to the idea that because you're from Harvard,

you only got in because your parents made a huge donation. The group. The yard birds, right? That's the name. The Harvard Yard. But they're open. Do you have a name suggestion? We're open.

Since you guys are middle aged. One erection. Listen to humor me with Robert Smigel and Friends on the I-Heart Radio app, Apple Podcasts, or wherever. you get your podcast. Humor me. I need some jokes to make me seem funny. Run a business and not thinking about podcasting, think again.

More Americans listen to podcasts than ad-supported streaming music from Spotify and Pandora. And as the number one podcaster, IHeart's twice as large as the next two combined. So whatever your customers listen to, they'll hear your message. Plus, only I-Hart can extend your message to audiences across broadcast radio. Think podcasting can help your business. Think I-Hart. streaming, radio, and podcasting. Call 844-844-I-Hart to get started.

That's 844-8-4-I-Hart. Agency, the ability to know that we're the experts in our own body. On the podcast, cultivating her space, Dr. Dom and Terry Lomax create a space where black women can show up fully and be heard. I wholeheartedly think, you know, you hit 30. You shouldn't have to share one with anybody. Mm-hmm.

From navigating friendships and healing to setting boundaries, and prioritizing your mental health. These are real honest conversations. We don't always get to have out loud. Totally unreasonable with different parts of life, right? Like, oh, have all three meals and make sure you're mindful during all of them? Absolutely not. During one meal, I'm standing.

I'm standing and handing my children food. Because healing, empowerment, and resilience aren't just ideas. Their practices. And this Mental Health Awareness Month, there's no better time to pour back into yourself. Listen to cultivating her space on the IHeartRadio app, Apple Podcasts, or wherever you get your podcast. In early March, the information published the story about Amazon and Google tamping down generative AI expectations, with these companies dousing their salespeople's excitement about the capabilities of the tech they're selling. A tech executive is quoted in the article, saying that customers are beginning to struggle with questions,

simple questions like, is AI actually providing value? and how do I evaluate how AI is doing? And a Gartner analyst told Amazon Web Services, sales staff, that the AI industry was at the peak of the hype cycle around large language models and other generative AI, which is a somewhat specific code for, it's not going to get much better anytime soon. This article confirms many of my suspicions

that, and I quote the information here, other software companies that have touted generative AI as a boon to enterprises are still waiting for revenue to emerge, citing the example of professional services firm KPMG buying 47,000 subscriptions to Microsoft's co-pilot AI at a significant discount on the $30 a seat sticker price. Except KPMG bought these subscriptions without really having gauged whether their employees actually got anything out of it. They bought it, and I'm not kidding you, entirely so that if any KPMG customers ask questions about AI, they'll be able to answer them. We're so clearly in the bow. Oh my God. Anyway, as I've hinted, it's also not obvious how much AI

actually contributes to the bottom line. In Microsoft's Q4 2023 earnings report, chief financial officer Amy Hood reported that six points of revenue growth in its Azure and Cloud Services division was attributed to AI services. I went around the web and I read every bloody article about their earnings. I looked and I looked and everyone was saying, oh, this is really good. I found someone who said it was 6% of their revenue and I went, that sounds like complete bollocks to me. So I went and spoke with Jordan Nevette, who's covered Microsoft for many years. He's a great cloud reporter over at CNBC and he actually covered Microsoft's earnings for CNBC itself. And he confirmed what this means is that AI contributed 6% of the 30% of year-over-year

growth in Microsoft's Azure cloud services. That is a percentage of a percentage. So, by the way, that means so 30% growth year over year, so 6% year-over-year growth is from AI could be good, but also all of the rest of it came not from new products, just from the natural growth of the company. It's unclear how much money that really is, but 6% of the year-over-year growth isn't really that exciting. Anyway, elsewhere, Amazon CEO Andy J.

Gassi, who took over from Bezos a few years ago and was the chief of Amazon Web Services, said that generative AI revenue was still relatively small. But don't worry, you said it would drive tens of billions of dollars of revenue over the next several years, adding that virtually every consumer business Amazon operated in already had or would have generative AI offerings. Now, they can just say that stuff. I really want you to know, you can say what you want on earnings calls. As long as you're not just outright lying, like saying we have $100 billion in cash, but you have $50. That is a lie.

You can't do that. But you can be like, oh yeah, mate, we've got all sorts of AI and everything now. It's bloody magical. You don't take a look, but it's there, I promise you. They can just say what they want.

But don't worry, they're not the only ones. Salesforce chief financial officer, Amy Weaver, said in their most recent earnings call that Salesforce was not factoring in material contribution from Salesforce's numerous AI products in its financial year 2025 guidance. Software company Adobe's shares slid in their last earnings as the company failed to generate meaningful revenue from its masses of AI products, with analysts now worried about its ability to actually monetize any of these generative products.

Service now claimed in its earnings that generative AI was meaningfully contributing to its bottom line, yet a story from the information quotes their chief financial officer, as saying that from a revenue contribution perspective, it's not going to be huge. I'm going to be a bit honest. I'm feeling a little insane with this stuff. I feel crazy every time I think about these stories because elsewhere in the media, so many people are saying

how big and successful the generative AI revolution is and it's going to be. Yeah, every time I look at the actual places where they write down how much money it makes, any of the actual signs of growth and significance and utility and adoption it's just not there. It's just breathless hype with this kind of whisper of stagnation and non-existent adoption.

And while there are startups beginning to mine usefulness out of generative AI, and they do inside by automating internal queries and customer support questions, these are integrations rather than revolutions, and they're far from the substance of a true movement. Maybe the darker truth of the generative AI boom is that it's a, feature, not a product, and that these features might be built entirely off the back of large language models, which are unsustainable to run, grow, or even make better. What if AI only drives a couple percentage points of real revenue growth for these companies?

What if what we're seeing today is the upper limit, not the beginning? Honestly, I'm beginning to believe that a large part of the AI boom is just hot air, and it's being pumped up through a combination of executive bullshittery and a very compliant media that's so happy to write stories imagining what AI can do, yet seems unable to check what it can do or what it's doing. It's so weird. Now, there's a bloke over at the Wall Street Journal called Chip Cutter. He should really look into if you want to know why your boss keeps asking you to go back to the office.

Wall Street Journal's Chip Cutter, he loves to write things about how bosses are good and how returning to the office is good. He wrote a piece in March about how AI is being integrated into the office, and most of it was just hundreds of words as him guessing about what people might do. But when he gets to the bottom, and he starts talking about companies using it, it's almost entirely examples of people saying, yeah, it makes too many mistakes for us to rely on it and we're just experimenting at. Elsewhere in the media, the New York Times talked with Salesforce's head of AI, Clara Shear.

And in this, I think it was, 600, 700-word article, didn't really get her to say much of anything about AI or what their products do. All she said was that their Einstein Trust layer handles data, and you may think I'm being facetious here, that's all she said about that. And then she added that it would be transformational for jobs the way the Internet was. What? What does that mean? Why am I reading this in the newspaper?

Why is this what I read in the newspaper? How is this helping? I know I rant a lot on this podcast, and I'm going to keep doing it. You're stuck with me, all right? It's free, okay? You don't pay for this, unless you do cool as though media, which you should pay for. Anyway, I know I'm ranting, but the reason that this stuff really infuriates me is it's misinformation on some level. I know it's kind of dramatic to say, oh, they're misinforming people by suggesting that AI can do stuff, but it is. It is misinformation. When you're letting corporate executives go in the newspaper and talk about how amazing their products will be, without asking them what they can do today, you're just giving them free press. You're not giving them credit for stuff they've done.

You're giving them credit for things they're making up on the spot. And when you do that, you make the rich, richer and the poor poorer. You centralize power in the hands of assholes. People who are excited. people who are borderline masturbatory jumping around saying, oh God, I can't wait till I can replace humans with fucking computers. Good news is they're not going to be able to, but that's what they're excited about,

and that's what they're getting media coverage around. The media has been fooled, just like they were with the Metaverse, by this specious promise train of the generative AI generation and these worthless executives championing it. These half-truths and this magical thinking has spread far faster due to the fact that AI actually exists and is doing something. And it's actually much easier to imagine how it might change our lives, unlike the Metaverse.

Even if the way it might do so is somewhere between improbable and impossible, it is easy to think about how it might work. You know, you could use an AI to automate date entry or boring busy work. Surely all of this you can automate, right? And when you use chat GPT, you can almost kind of sort of somewhat see how it might happen. even if when you open up chat GPT and try and make it do something, it's always a bit off. Never seems to quite do it.

I'm in a very, my day job, my PR firm. I'm in a spreadsheet and document heavy business. Of all the people who this could help, you think it would be me? A lot of my work is, hey, all of these things I need them in a spreadsheet. It can't bloody do it. And I'm sure you listeners will probably email me and say, oh, I've used chat GPT for this. Don't care.

I really mean that. This thing is not changing the world. And actually, I think far more of you have already shared. Thank you, by the way, easy at better offline.com. You can email me your ideas and your angry comments, you can go to the Reddit to complain. But the thing I'm hearing for most people is, yeah, I've tried it and it didn't do enough.

I tried it and there were too many mistakes. There was a Wall Street Journal article back in February about how Amazon and Google were having trouble selling AI services because, well, when they went to sell them, these companies, these financial services companies in particular, they were saying, yeah, but these hallucinations could actually get the SEC man at us. And the answer that they had was, yeah, what if we just made it so that the models would sometimes say they don't know stuff? Every time you get to a reckoning with AI where you want it to be better, where you're like, hey, AI, how will this actually be fixed, these hallucination problems, for example,

they come up with the most mealy-mouthed ship. And I truly believe it's because there is no answer to these problems, as I said in the previous episode. And I think that's why I can't find any companies that have integrated generative AI in a way that's truly improved their bottom line, other than Clana, which allows you to do 0% interest-free loans on almost anything. It's a very worrying company. Anyway, they claimed that their AI-powered support,

bot was estimated to drive a $40 million amount in profit improvement in 2024, which does not, by the way, despite it being trumpeted by members of the media otherwise, mean that they made $40 million in profit. I actually can't find what profit improvement refers to. And this is the classic AI boom story, by the way. There's always this weird verbal judo going on where they're like, yeah, it's $40 million in profit improvement upwards and downwards and side to side, mate. It's really good. And I think it's just headline grabbing. I think it's just buzz.

And despite fears to the contrary, AI doesn't appear to be replacing a large amount of workers. And when it has, the results have been pretty terrible. Like when Microsoft replace MSN.com's editorial team with a series of AI bots that have spread misinformation and conspiracy theories, things like Joe Biden falling asleep, it's so weird. Interestingly, there was also a study from Boston Consultancy Group, and just as a note, if anyone would love the opportunity to just replace workers with robots, it's BCG, McKinsey, Accenture, all these companies would absolutely, they would be giving however much OpenAI wanted to do that, and then they were charged $50 million for an integration that didn't work, which I guess makes AI perfect for them.

Putting that aside, in a study from BCG, they found that consultants that solved business problems with OpenAI's GPT4 model performed 23% worse than those who didn't use it, even when the consultant was warned about the limitations of generative AI and the risk of hallucinations. Yeah, really great stuff. To be clear, I am not advocating for the replacement of workers with AI. However, I'm saying that if it was actually capable of replacing human outputs, even if it was even anywhere near doing so,

any number of these massive, horrifying firms would be doing so at scale, and planning to do so more as the models improved, they'd be funneling cash right up open AI's ass. It would be incredible. But the reality of the AI boom is kind of a little more boring. It recently came out that Amazon's cashless just walk out technology. In some of their stores, you could walk in, scan a QR code, and you could just grab your rouse tomato sauce and your condoms or whatever,

your weird magazines. I don't know what they're selling there. I'm not giving any more money to Amazon than I need to. anyways, everyone thought, oh, it's just AI, you could just walk in, it's magic. The cameras would tell you through computer vision what you had bought. It would be great. Now, it turns out that there's a thousand workers in India that were monitoring these cameras and approving transactions. Worse still, Open AI used Kenyan workers who were paid less than $2 an hour to train chat GPT's outputs. And they currently pay $15 an hour, I think, for American contractors.

No benefits, of course. You know, fuck work. Lookers, right? That's the thing. Underneath this whole thing is just this undercurrent of disrespect for human beings and it pisses me off. And I realized I'm pissed off about a lot of things. You'd be listening for like an hour now. Oh, half an hour in this episode. Anyway, I'll keep going. But yeah, like I said, if AI was changing things, if AI was actually capable of replacing a person, it would have happened. It would be happening right now. It'd be happening at scale. It would be so much worse than things feel now? Unless, of course, it just wasn't possible.

What if what we're seeing today is not a glimpse of the future, but actually the new terms of the present? What if generative AI isn't actually capable of doing much more than what we're seeing today? What if there's not really a clear timeline when it will actually be able to do more? What if this entire hype cycle has been built, goosed and propped up by this compliant media they're ready and willing to take whatever these career embellishing bullshitters have to say. What if this is just another metaverse but with a little bit more product?

Every single time I've read about the amazing things that artificial intelligence can do, I just see somebody attempting to add fuel to a fire that's close to going out. When the Wall Street Journal's Joanna Stern wrote about Sora, OpenAI's yet to be released video generating chatbot, she talked about how its photorealistic clips were good enough to freak her out. And I get it. At first glance, these do look like people. These images do, they look like something approaching a video.

They look almost real. Kind of like texts from chat GPT is almost right, or it's right fully, but it doesn't feel right. But much like the rest of these outputs, you look a little closer, and they have these weird errors like cars disappearing in and out of the shot or a different car coming out from behind something, or completely different images between frames, or these strange, unrealistic moments of lighting. And they're never much longer than 30 seconds.

Stern, who, by the way, I deeply respect, isn't really afraid of what Sora can do, but what would happen if Open AI was able to fix the hallucination problems that makes these videos kind of unwatchable? While it's easy to imagine tools like Sora could eventually play a role in online disinformation campaigns churning out like lifelike videos of politicians saying or doing appalling things, we can all breathe a sigh of relief in knowing that the videos themselves are often so flawed, you can pretty much instantly see their AI generated. Also, Sora is not available to the public yet, and I don't even know if it ever will be. You just need to look at the hands or the backgrounds. Look at the people in the background of any AI-generated photo or video.

They often contain too many fingers or you can't see their faces. Or in Sora's videos, their legs don't look right. It's so weird. And I don't know how to put it perfectly, but they don't feel human. Just to be clear, though, Sora is dead on arrival. No one actually has access to it. It's unclear when it will come out. Every journalist that has quote-unquote used SORA has just given a prompt to OpenAI to run.

But there's also a very obvious problem that kind of relates to something I mentioned in the previous episode. Open AI and every generative AI company, they're all dependent on high-quality data to train the models. And video data is so much larger, more complex, and harder to find. There's less of it because it's visual media. And it's just a much bigger, more complex model and a much harder computational task to create video. Moving image is... It's actually putting aside my anger about Generative AI, it's amazing they've done even this. But to be clear, as amazing as it might look, it isn't enough to do anything.

It's just a kind of a do-hicky. And this data is so much more complex than the text-based data that OpenAI is running out of to make ChatGBT spit out words. Even if there were enough data, there's pretty good reason why Open AI is coy about when they'll release the model. Like I said, it's expensive and complex to run, and at no point, has anyone asked how the fuck this actually makes them any money, how they sell this? It's so weird. To be clear, when you use SORA, it turns text prompts into a video. You can't edit the video, you can't change the video, the video is what the video looks like.

There's no way to make SORA make the same thing multiple times, which makes the very basics of making film, which is multiple angles of the same thing, completely goddamn impossible. In fact, consistency between the same two prompts is impossible from these models, because they're all probabilistic. We've recently seen some of the first quote-unquote movies made with Sora, and the first one was called Airhead, which is about a minute long. It's this man with a balloon head walking around and it's got this, it's very twee. It sucks. It's just putting aside the AI part, it's just crappy. And it's got a guy being like, yeah, having a balloon head is difficult. Yeah, it's weird.

I hate having a balloon head. I hope I don't get popped. It sucks. It's really bad filmmaking. But also, each shot, and there's multiple shots of this guy with a yellow balloon head looks completely different. It's a different balloon every goddamn time. It's a different balloon. And it's so funny because you have these guys in Twitter going like, oh my God, oh my God, I'm crying and pissing myself. This is the best thing I've ever seen. But it isn't. It's so close yet so far away. And the only reason it's impressive is people are willing to sit there and say, but what if it wasn't shit?

But it is. It really is. Like every other generative output, it's superficially impressive. kind of sort of lifelike. But once you look at it for more than a moment, it's just flawed, terribly, irrevocably flawed. It's time to wake up. We are not in the early days of AI.

We're decades in, and we're approaching the top of the S-curve of innovation. There are products being built, don't worry. But it's all things like Claude Author, which creator Matt Schumer calls a chain of AI systems that will write an entire book for you in minutes, and I call a new kind of asshole that can shit more than you'd ever believe. Generative AI is the ugliest creation of the rock economy,

and its main selling point is that it can generate a great deal of passable material. Images generated from Generative AI models like OpenAI's Dali, all have the same kind of eerie feel to them, as they're mostly trained on the same data, some of it licensed from shot stock, some of it outright plagiarized from hundreds of artists. Without sounding too wanky and philosophical, everything created by Generative AI feels soulless.

And that's because it is. No matter how detailed the prompt, no matter how well-trained the model, no matter how well-intentioned the person writing that prompt, these are still mathematical solutions to the emotional problem of creation. One cannot recreate the subtle fuck-ups and delightful little neurological errors

that make writing a book or a newsletter or a podcast special. While this podcast is, admittedly, trying to generate what I believe AI might do in the future, it's not generative and it's not generated as a result of me mathematically considering how likely an outcome is. My fury is not generated by an algorithm telling me that this is the right thing to be angry at. I'm pissed off because I feel like we're all being lied to and treated like idiots. What makes things created by humans special isn't doing the right thing or the best thing, but the outputs that result in us fighting past our own imperfections and maladies. Like the strep infection I've been fighting for the last few days.

And like, look, to my knowledge, you can't give a generative AI strep throat, but if I ever find out it's possible, I will make it my damn mission to give it to chat GPT. Another podcast from some SNL late-night comedy guide, not quite. Unhumor me with Robert Smygel and friends. Me and hilarious guests from Jim Gaffigan to Bob Odenkirk to David Letterman, help make you funnier. This week, my guest, SNL's Mikey Day and head writer Streeter, Adele help an a cappella band with their between songs banter.

There's the worst singer in the group. The worst? Yeah. Me. Is there anything to the idea that because you're from Harvard, you only got in because your parents made a huge donation. The group. The yard birds, right?

That's the name. The Harvard yard, but they're open. Do you have a name suggestion? We're open. Since you guys are middle aged. One erection. Listen to humor me with Robert Smike. and friends on the IHeart radio app, Apple Podcasts, or wherever you get your podcast.

Human me, I need some jokes to make me seem funny. Run a business and not thinking about podcasting, think again. More Americans listen to podcasts than ad-supported streaming music from Spotify and Pandora. And as the number one podcaster, IHearts twice as large as the next two combined. So whatever your customers listen to, they'll hear your message. Plus, only IHeart can extend your message to audiences across broadcast radio. Think podcasting can help your business. Think IHeart. Streaming, radio, and podcasting. Let us show you at IHeartadvertising.com. That's iHeartadvertising.com. Agency, the ability to

know that we're the experts in our own body. On the podcast, cultivating her space, Dr. Dom and Terry Lomax create a space where black women can show up fully and be heard. I wholeheartedly think, you know, you hit 30. You shouldn't have to share one with anybody. Mm-hmm. From navigating friendships and healing, to setting boundaries and prioritizing your mental health. These are real, honest conversations. We don't always get to have out loud. Totally unreasonable with different parts of life, right?

Like, oh, have all three meals and make sure you're mindful during all of them? Absolutely not. During one meal, I'm standing. I'm standing and handing my children food. Because healing, empowerment, and resilience aren't just ideas, their practices. and this mental health awareness month, there's no better time to pour back into yourself. Listen to cultivating her space on the IHeart Radio app, Apple Podcasts, or wherever you get your podcast. All of this hype is predicated on solving problems with artificial intelligence models that are only getting worse.

And Open AI's only answers to these problems are a combination of, we'll work it out eventually, trust me, and we need a technological breakthrough in both chips and energy. That's why Sam Altman has been trying to raise $7 trillion, and that's not a mistake, by the way, to make a new kind of AI chip, because there's no sign that this or even future generations of chips will actually fix anything. Generative AI's core problems, its hallucinations, its massive energy, and its massive, unprofitable compute demands are not close to being solved. I've now watched a frankly alarming amount of interviews with both OpenAI CEO Sam Altman and their CTO Miramirati and every time they're saying the same specious empty talking points promising that in the future chap GPT will do this and that as all evidence points to their models getting worse and through the last years by the way they've just said the same thing

in every interview they always talk about chat GPT being like something or help creatives they never really say how it's just kind of weird But yeah, generative AI models, they're expensive, they're compute intensive, and they don't seem to provide obvious, tangible mass market use cases. Murati and Altman's futures depend heavily on keeping the world believing that development and improvement of these models' capabilities will continue at this rapacious pace of progress, even though it's unquestionably slowed, with Open AI even admitting themselves that their latest model, GPT4, may actually be worse at some tasks. A study from UC Berkeley last year found that GPT4 was actually worse at coding them before and that chat GPT was at times refusing to do certain tasks.

Nobody wants to work anymore. I feel like I'm walking down the street telling people their houses are on fire, only to be told to stop insulting their new heating system. These models aren't intelligent, they're mathematical behemoths generating the best guess on training data and labeling, and that's they don't really know

what they're being asked to do. You can't fix that. You can't fix hallucinations. You can't just make these problems go away with more compute. You can mitigate them. The current philosophy, by the way, is that you can use another model

to look at another model's outputs, which, as I mentioned in the previous episodes, is very silly. But seriously, everyone telling you hallucinations are going away, look a little deeper, and look at when they actually failed to tell you how they were. It's just very silly. Look, every bit of excitement for this technology right now

is based on this idea of what it might do, as I've said. And that quickly gets conflated with what it could do, which allows Sam Altman, who, by the way, is far more of a marketing person than an engineer, his one startup looped was a failure. He's failed upwards. He was in Y Combinating he did this. It's actually ridiculous. He's so famous.

All of this bullshit, it allows him to sell. the dream of Open AI, and he's selling it based on the least specific promises I've seen since Mark Zuckerberg said we'd live in our bloody Oculus headsets. And it's frustrating, because this money and this attention could go to important things. We have real problems in society. I believe that Sam Altman, and pretty much anyone in a position of power and influence in the AI space, has been tap dancing this entire time, hoping that he could amass enough power and revenue that his success would be inevitable. Yet I think his hype campaign has been a little bit too successful. And it's deeply specious. And he, along with the rest of the AI industry,

has found himself suddenly having to deliver a future that he's not even close to developing. I am always scared of automation taking our jobs. I think it's always worth being scared of. But I don't think that's the thing the tech industry is working on right now. I don't think they're close. And I think there's something more imminent to fear. And that thing is the bottom falling out of generative AI, as companies realize that the best they're going to see is maybe a few digits of profit growth. Companies like Nvidia, Google, Amazon, Snowflick, and Microsoft, they have hundreds of billions of dollars of market capitalization, as well as expected revenue growth, tied into the idea that everyone's going to integrate AI into everything, and that they'll be doing more than they

are today. You can already see the desperation coming from these companies, like Microsoft, for example, which in March effectively absorbed a company called Inflection AI into itself, kind of an acquisition by stealth. Inflection AI is a public benefit company that portrays itself as a nicer, gentle version of Open AI. Its core product, a chat GPT-style chatbot, counts its empathetic tone, its humor and its emotional awareness. Inflection was created in 2022 with an all-star founding team that included Reed Hoffman, the founder, of LinkedIn and Mustafa Suleiman, the British-born co-founder of Deep Mind, which Google acquired in 2014. In mid-2020, Microsoft took part in a $1.3 billion funding round, which saw the company

acquire a significant stake in inflection alongside other AI players like NVIDIA. Inflection's core product has the same inherent underlying issues as every other generative AI product, hallucinations, for example. But it has an accomplished team that has taken a different approach to its competitors, Whereas ChatGPT and Claude 2 tend to be, or at least aspire to be functional, tools that provide information or complete tasks, inflection sought to make its product feel a bit more organic. For Microsoft, the appeal was obvious. It has so much riding on its AI ambitions, but in terms of money spent as well as its share price, that it can't really afford to appear stagnant or worse as though it made a bad bet.

Acquiring inflection would help it maintain its image, especially with Idiot Wall Street analysts. But here's the problem. Microsoft already holds a massive stake in Open AI, and regulators, both in America and Europe, are wary of market consolidation. Acquiring inflection, it'd give them a little too much scrutiny. So Microsoft took a third nastier puff. Instead of buying the company, it bought the employees, with Suleiman and the majority of his co-workers jumping ship to found Microsoft's new AI division. It secured the talent, a subsequent $650 million licensing deal, yet another example. example of Microsoft basically paying itself, and then gave that deal to the shell of inflection,

you know, the one without any of the staff left, giving it access to the company's tech, its IP, and there's nothing regulators could do to stop it. To be clear, Microsoft is in a position where it could easily absorb the shockwave of a potential AI bubble burst. It still prints money from its other business units like Office and cloud computing, Microsoft Windows, and the Xbox gaming system. And the same is true for the other big name. like Google and Nvidia. They're well insulated for any slowdown in AI investment or from a growing apathy towards AI enterprise customers.

I will note, however, these massive investments in data centers, if they're all for naught, you will see a form of crash. I can't say the same for startups, though. Other companies aren't going to be so lucky. Stability AI, the developer of stable diffusion, a generative AI that can produce images from written prompts,

Innovative for the time is perhaps the canary and the coal mine of PKK. Stability AI rode the same waves as Open AI, especially in 2023, but now that money is tighter and skepticism is higher, it's struggling to stay afloom. Although the company raised $100 million in early 2023, it burned through nearly $8 million a month, and in a recent attempt to raise further cash, they failed. The company routinely missed payroll, and according to Forbes, amassed a sizable debt with the U.S. tax authorities that culminated in threats to seize the company's assets. They owe debts to Amazon, Google, and Corweave,

a niche compute provider that specializes in AI applications. With negligible revenues and rapid cash burn, combined with no obvious way to monetize the product, stability AII is now in turmoil, with its key talent leaving the company in March, followed by the company's CEO and co-founder Imadmastak. Its ongoing existence is in question, with the financial times writing in March,

that the company's future, despite once being seen as among the world's most promising, startups is in doubt. While it would be fair to say that stability AI was unique at its internal turmoil, its external pressures, the ability or lack thereof to monetize an expensive product, and its reliance on external funding to survive are much more common across the industry. Its survival depended on investors believing in a lofty future for AI, where it's integrated into every facet of our lives, and it plays a role in almost every industry, which, of course, we now know it doesn't. While that belief hasn't been shown,

shattered, or at least not yet, it's fair to say that expectations and aspirations are increasingly tempered. After reaching the apex of the AI piss fest, the tech industry is getting a hangover, and companies like stability can't survive the headache. But to be clear, I am not excited for the AI bubble to pop. And on some level, as weird as it sounds, I kind of hope it doesn't. Once it bursts, the AI bubble will hit far more than the venture capitalists that propped it up. This hype cycle has driven the global stock markets to their best first quarter in five years. And once the markets fully turn on the companies that falsely promised an AI revolution, it's going to lead to a massive market downturn and another merciless round of layoffs

throughout the tech sector led by Microsoft, Google and Amazon. This will in turn suppress tech labor and flood the market with tech talent. It's going to suck for everyone involved in software. A market crash led by the tech industry will only hurt innovation first. other, draining the already low amounts going into the hands of venture capitalists that control the dollars going into new startups. And once again, the entire industry will suffer because people don't want to build new things or try new ideas. No, they want to fund the same people doing the same things or similar things again and again because it feels good to be part of a consensus,

even if you're wrong. Silicon Valley will continually fail to innovate at scale until it learns to build real things again, things that people actually use and things that actually do something. I don't know if I want to be right or wrong here. If I'm wrong, Generative AI could replace millions of people's jobs, something that far too many people in the media are excited about, despite the fact that the media is the first industry that Open AI kind of wants to automate. If I'm right, we're going to face a dot-com bubble S town turn in tech, one that's far worse than what we saw in the last few years. In any case, I do wish the tech industry would get their heads out of their asses.

I'm tired. I'm tired of watching tech firms life for their teeth about the future. That will live in the metaverse, that our future will be decentralized and paid for in cryptocurrency, and that our world will be automated with chatbots. I truly think that these companies think regular people are stupid, which is why Microsoft put out a minute-long Super Bowl commercial for their co-pilot AI that featured several prompts like, write the code for my 3D Open World game, that don't actually do anything.

That prompt I just mentioned, go type it into copilot. It will give you a guide to coding a game, no code created. Also, in the commercial, he types in classic truck shop called Pauls. But none of these image generators can actually do words, so it just looks like scuba-blah-blub. Go and do it. Trust me, it's funny. But every time that these big tech booms happen, every time they say, oh, we're going to live in the metaverse, and ooh, we're going to be able to automate everything.

Every time they lie, the world turns against the tech industry. And this particular boom is so craven in its falsehoods that I think it'll have a dramatic chilling effect on tech valuations if the bubble pops quite as severely as I expect. And Sam Mortman desperately needs you to believe the bubble won't pop. He needs you to believe that generative AI will be essential, inevitable and intractable. Because if you don't, you'll suddenly realize that trillions of dollars of market capitalization and revenue of being blown on something, it's kind of mediocre. If you focus on the present, what OpenAI's technology can do today and will likely do for

some time, you see in terrifying clarity that generative AI isn't really a society-altering technology. It's just another form of efficiency-driving cloud-computed software that benefits kind of a small amount of people. If you stop saying things like AI could do or AI will do, you have to start asking what AI can do. And the answer is not that much and probably not that much more in the future. Sora is not going to generate entire movies. It's going to continue making horrifying human adjacent creatures that walk like the

Atats from Star Wars and cartoons that look remarkably like SpongeBob SquarePants. ChatGPT isn't going to run your business because it can barely output a spreadsheet without fucking up the basic numbers if it even understands what you're asking it to do in the first place. I think that AI has maybe three quarters to prove itself worthwhile before the apocalypse really arrives. When it does, you're going to see it first in the real infrastructure companies, starting with Nvidia, who's grown to about $2 trillion in market capitalization because of the chips they make, which are pretty much the only ones that can power the AI revolution. There are other companies like AMD and Micron, but Nvidia is the one that's really grown.

If you watch any of their keynotes, they're insane. They're just full of fan fiction. Once Nvidia starts to see growth slow, an Oracle in particular, Oracle, massive data sentient company, massive database company as well, one of their largest customers, Microsoft building data sentence for them. Once that starts slowing down, that's when you should start worrying. But the real pain is going to come for Amazon, Microsoft and Google. when it's clear that there's not really that much revenue going into their clouds. Once that happens, once you start seeing Jim Kramer on CNBC saying, I don't think the AI boom is here, despite having saying it just was,

that's when things get nasty. And the knock-on effects will be horrible. It's going to be genuinely painful, worse than we've seen in the last few years. And it's all a result of the same problem. It's all a result of the growth at all-cost tech economy, when things are made to expand, when things are made to build more rather than build better, when you're building solutions to use compute power to sell cloud computing services rather than helping real people make their lives better.

Tech is not building for real people anymore, and the AI revolution, despite its spacious hype, is not really for us. It's not for you and me. It's for people at Sachin Adela of Microsoft to claim that they've increased growth, by 20%. It's for people like Sam Altman to buy another fucking Porsche. It's so that these people can feel important and be rich rather than improving society at all. Maybe I'm wrong. Maybe all of this is the future. Maybe everything will be automated. But I don't see the signs. This doesn't feel much different to the metaverse. There's a product, but in the end, what's it

really do? Just like the metaverse, I don't think many people are really using it. All signs point to this being an empty bubble. And I'm sure you're sick of this too. I'm sure that you're sick of the tech industry telling you the future's here when it's the present and it fucking sucks. And I'm swearing a lot and I'm angry. But I'm justified in this anger I feel. I'm not telling you how to think and I've heard from some of you saying, oh, don't tell me how to think. And I agree. I agree. I'm not here to tell you to be angry about anything. But I want to give you, at least my truth. And I want to give you what I see is happening. Because I don't feel like enough people are doing that in the tech industry.

And that's what Better Offline is going to continue to be. I really appreciate you listening. It's been about a month, month and a half since we started. It's only going to get better from here. Thank you. Thank you for listening to Better Offline. The editor and composer of the Better Offline theme song is Mattosowski. You can check out more of his music and audio projects at Mattosowski.com.

You can email me at easy at Better Offline.com or check out Better Offline.com. to find my newsletter and more links to this podcast. Thank you so much for listening. Better Offline is a production of Cool Zone Media. For more from Cool Zone Media, visit our website, coolzonemedia.com, or check us out on the IHeartRadio app, Apple Podcasts, or wherever you get your podcasts. Another podcast from some SNL late-night comedy guy, not quite. Unhumor me with Robert Smigel and friends, me and hilarious guests from Bob Odenkirk to David Letterman helped make you funnier. This week, my guest, Michael's Mikey Day and head writer Streeter Seidel help an a cappella band with their between songs banter.

Where does your group perform? We do some retirement homes. Those people are starving for banter. Listen to humor me with Robert Smigel and friends on the I-Heart Radio app, Apple Podcasts, or wherever you get your podcasts. Your 20s can be so exciting, but they can also be really overwhelming, confusing, and honestly, just kind of lonely. May is Mental Health Awareness Month,

and the psychology of your 20s is breaking down the science behind the biggest roadblocks we face. I was six years into my career, the 80-hour weeks and just the first one in, the last one out, and I ended up burning out. There was a large chunk of my 20s that I like was just so wanting to, like, be out of that phase out of my skin, and I just like really regret not living in the present more. You don't need to have everything figured out right now. You just need to understand yourself a little bit better. Listen to the psychology of your 20s on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts. Why are we all so obsessed with romance?

On the Radio 831 podcast, join us. Sanjana Basker and Tyler McCall, as we unpack all the trending tropes, fuzzy adaptations, book talk drama, and celebrity love stories with hot takes and sharp guests. Each episode digs into what these stories reveal about desire, fantasy, identity, and how we love now. Listen to the Radio 831 podcast on the IHeart Radio app,

Apple Podcasts, or wherever you get your podcasts. This is an IHeart podcast. Guaranteed Human.