Enshittification With Ed Zitron at the Seattle Public Library  ## Metadata - Author: [[Podcast – Cory Doctorow's craphound.com]] - Full Title: Enshittification With Ed Zitron at the Seattle Public Library - Category: #podcasts - URL: https://share.snipd.com/episode/f68a20de-a4e7-43ee-a48f-a273914d3f8e ## Document Notes ## Highlights - **Search Degraded On Purpose** - Google deliberately degraded search quality to increase ad impressions and revenue. - Internal fights showed business priorities beat technologists' quality concerns. Transcript: Cory Doctorow Do you want to start by talking about Prabhagar Raghavan? Ed Zitron Prabhagar Raghavan. Love saying that motherfucker's name. So I don't know if any of you have read The Man Who Killed Google Search, which was a piece that I wrote by accident. I found a bunch of notes from the Department of Justice. Anyway, long story short, around the early part of 2020, there was a thing called a code yellow at Google, which just meant there was, I'm going to use some financial things, but there Was query weakness in Google, which means that not enough people were just asking Google things. Now, you may think, well, Google ideally gets you an answer, right? No, no, no, you pleb. How dare you? What do you mean? You get an answer? Fuck you. You need to look at ads. Sundar Peshai needs, and all of his McKinsey friends need to do their seance or whatever. Nevertheless, there was an internal argument at Google. There was a guy called Ben Gomes who ran search, who literally said, I'm worried that all Google cares about his growth. Turns out he was right. Around June that year, a guy called Prabhagar Raghavan took over. He was formerly the head of ads, and he pushed Ben Gomes to change things to make Google search worse to keep people on Google search. So the decay of Google search really started then. It was actually a few years beforehand. They were slowly making it, so it was harder to tell sponsored links and ads from regular search results to the point it's pretty much impossible now, and most of the results are pretty Bad as well. But long story short, Google deliberately made things worse, and then the guy who was called Prabhupada Raghavan, remember that name, arsewipe, he mandated basically that, I can't Prove it, but it matches up, that he just reduced the quality of results. They allowed websites that had been previously pushed down to come back, spammy results, so that you ended up farting around on Google more, seeing more impressions, and bing bong, Number go up. Propagor, sadly, he sadly moved on. He's alive. He is now the chief technologist at Google. I would say 30% my work was the problem. The funny thing is that Propagall was riding high, taking a massive, massive growth of Google. He was doing so well. And then he got made to take over Gemini, Google's AI. And then there was a thing in, I think it was early 2024, where this, well, maybe it was 2025, where Google Gemini was generating Chinese George Washingtons. And the conservatives were going, and there was, there's a black Nazi, there's a black, there's a Chinese George Washington. I can't, my, my tiny little brain is vibrating at a thousand miles a minute. And for some reason, Prabhupada Raghavan had to take responsibility. And this is not a charming man. This isn't a man who's like a political navigator. He's just rude. He's very good at being rude. In the end, sadly, Propagar's work has been done. Google sucks now. ([Time 0:06:18](https://share.snipd.com/snip/d29dab73-4e90-4372-9114-16f01a18522b)) - **Policy Created Platform Rot** - Platform rot arises from policy choices that enabled serial acquisitions and market capture. - Big firms grew safe from competition and regulatory discipline, creating perverse incentives. Transcript: Cory Doctorow And the question that I try to interrogate in shitification is what gave rise to it, right? What created the in shit a scene? And my conclusion is it wasn't because you shopped wrong, right? It's not because, oh, you didn't pay for the product, so you became the product. You know, farmers who buy half million dollar tractors are exploited by John Deere, which won't let their repairs go live until they pay a $200 call-out fee for a John Deere person to Come out and type an unlock code into the keyboard. Now, that is not a free ad-supported tractor, right? That's a tractor they paid six large for, and it won't work until they pay ransom money. So it's not because you shopped wrong. It's also not because these guys are the wrong guys to be running the company. They are terrible people. But the reality is that these Zuckermuskian mediocrities that run these companies are not smart enough to be causes. They must be effects, right? They're responding to an enxietogenic environment created by policy. And that policy, in the case of Google, is the policy that oversaw for decades Google's serial acquisition of both vertical and horizontal competitors, so that a company that had only Made one really successful consumer-facing product, which they made a millennium ago, right, at their search engine. And that had almost, with that exception, failed to launch anything internally, except for things they bought from other people in anti-competitive acquisitions, to the point where They bought all the shelf space, right? Nowhere else a search engine could take root. They're bribing Apple to the tune of more than $20 billion a year not to enter the search market and have a direct competitor that would erode their margins. So they become too big to care. And this is really my thesis, right? ([Time 0:11:07](https://share.snipd.com/snip/106e0adb-a083-4fc7-9498-c367469427e1)) - **Show Up At Local Meetings** - Attend town halls and oppose harmful local AI/data center projects with facts. - Public pressure can block infrastructure that mainly serves speculative AI growth. Transcript: Ed Zitron So we live in a high information, low processing environment. The reason that regulators don't do much is not only because they're incentivized not to, but also they don't know shit about fuck. In fact, when you go around reading most ton of media, when you read a bunch of regulatory stuff, when you read what the government says about things, they don't know what they're talking About. I don't even mean slightly. I mean that the markets have been running away with this idea that generative AI is the future. The media has been reporting that generative AI is replacing jobs. Neither of these are true. They're not true. There are some jobs being eroded by LLMs, but it's jobs like translators that were already being eroded by machine translation. I forget the exact terms. And art directors. So real jobs, but contract labor. And also bosses that wouldn't pay real people anyway. The world itself is run on ignorance. Not just social ignorance, but just straight up factual ignorance about what AI can do. And the reason I pick AI is because this is where you can actually be disruptive. You hear a fucking data center popping up, go to the town hall meeting. Make your mayor upset. I'm deadly serious. I saw one brave soul at a Wisconsin data center meeting, he's quoted my work, love him, going to these places and actually not making a scene, just going in and saying, hey, there's only $60 billion worth of revenue in this industry. There's no growth outside of ChatGPT. Like these companies, OpenAI burned $9.2 billion in the first half of this year. That's fucking crazy. That's a completely crazy thing. There's no path to profitability for these things on top of there not being that much demand. I'm not just quoting things from my newsletter at you. Is the shit that you should be saying to your elected officials and show up at town hall meetings they hate it they do not want to see you but they have to so you show up and you tell them these Very simple things and they'll say what i read in the newspaper shut the fuck up man i want to hear from you man stanley i don't know or bowser i guess but Nevertheless, showing up and actually Just saying the very basic things and sticking to your points will do a lot. I know it sounds small, but really, that's where you get things done. Because these data centers are probably not getting built anyway. But if you can undermine them, then they're really not getting built. ([Time 0:21:15](https://share.snipd.com/snip/a77e0316-da34-4185-8176-c1d6713f2418)) - **Be A Citizen Not Just A Consumer** - Join or support grassroots digital-rights groups to influence policy and local rules. - Civic organizing changes structural outcomes more than individual consumer choices. Transcript: Cory Doctorow So as I've mentioned before, I work at the Electronic Frontier Foundation. We have a national network of grassroots groups, including several here in Seattle, called the Electronic Frontier Alliance. If you go to efa.eff you can find some of these local chapters. And they work on things like limiting police use of facial recognition, buying surveillance technology, privacy for abortion seekers, limitation on the use of digital infrastructure To track down people that ICE is chasing, right to repair laws. Washington's got a really good one. All of those things start at these grassroots levels. And it's getting involved as a polity that makes a difference. ([Time 0:25:34](https://share.snipd.com/snip/f5028190-0840-4345-a9cc-7dffc2d5ba32)) - **Unionize Tech Workers Now** - Tech workers should unionize to reclaim bargaining power and limit abuses. - Worker solidarity can constrain corporate worst impulses even without perfect regulators. Transcript: Cory Doctorow So, you know, the tech workers did have this chance to consolidate their power, and they missed it. So for a long time, one of the forces that constrained in shittification was workers themselves. Because tech workers are this uniquely constituted workforce historically. The tech sector has always had very low union density, but an enormous amount of worker power. And that's because tech workers were both very scarce and very, very valuable. The National Bureau of Economic Research estimates that the average tech worker in Silicon Valley in Seattle was adding a million dollars a year to their employer's bottom line. That's why they gave you free kombucha and massages and why they'd hire a surgeon to freeze your eggs so you could work through your fertile years. It wasn't because they loved you, right? It was because they were afraid of you getting a job across the street. Now, we know how tech bosses treat the workers. They're not afraid of losing, right? That's the Amazon driver who's got the AI camera that takes points off if they look away from the road to check something to one side or the other, and it decides that their eyeballs were In the wrong orientation. It's the warehouse workers that are injured at three times the rate of other warehouse workers in the sector. It's everyone who pees in a bottle. It's people who assemble iPhones in China and have suicide nets around the factory. That's how they treat the workers they're not afraid of. So there was this opportunity at one point for tech workers to use that power and consolidate it through a union. And they missed it, right? ([Time 0:26:44](https://share.snipd.com/snip/371fa157-5dd0-4caf-a53b-93f0c848de94)) - **AI Is A Hype-Driven Bubble** - The AI investment surge is a speculative bubble fueled by hype, not profits. - Major AI players burn billions and profitability remains elusive. Transcript: Cory Doctorow Buy long poles to dig through rubble with canned goods. Speaker 3 Let's try again, Ed. Ed Zitron But if you can't, know that gravity exists. Right now, the market is going absolutely batshit insane. I am not a stock analyst. I cannot give you financial advice. But right now, number going up. So maybe you could... I don't own stocks. I'm a psychopath. I live in cash. I invest in words. The problem we have right now is that NVIDIA is the largest stock on the stock market. There has never been a more problematic fact than that. 88% or more of their revenue is selling these fucking GPUs. You want to know what happens when you plug those GPUs in? They start losing you money. Nobody, not a single company other than Nvidia, is making any kind of profit on AI. In fact, they're burning billions. The reason I tell you this is you can't really navigate away from the bubble right now other than just selling, I imagine. So know what you're going into. Everything you're reading Oracle say right now about their relationship with OpenAI is bullshit. And to get specific is Oracle has a five-year $300 billion contract with OpenAI. Sounds amazing, right? They just need 4.5 gigawatts of data center capacity. Any guesses how much they have? 200 megawatts. Now, don't worry. That's less than 400 gigawatts, right? That's a lot less. That's a lot less. And I must be clear, I actually mean power. IT load, they've got about $1.30. So, not to worry, though, because OpenAI also doesn't have the money. Nevertheless, the stock is run. AMD now has a deal where they're going to sell chips to OpenAI. Who the fuck knows who's paying? NVIDIA, same deal. Whenever you see OpenAI involved, just don't. Just don't do it. There is no avoiding a bubble popping now. There just isn't. So, the smartest thing you can do is stick to the fundamentals. We're going to get into the more hysterical phase now. We're going to see some crazy shit. I had someone suggest the other day that OpenAI may do a SPAC IPO. I think it's insane, which is why it's possible. Do not let your money touch OpenAI under any circumstance. That company is cancer. I'd say the same thing about Anthropic. ([Time 0:31:02](https://share.snipd.com/snip/c89791aa-0875-4f2d-ada1-93c0be85da7f)) - **Protect Investments Against AI Hype** - If you're invested, prioritize fundamentals and beware AI stock hype and single-company concentration. - Expect market corrections; avoid speculative bets like OpenAI or Anthropic. Transcript: Ed Zitron But if you can't, know that gravity exists. Right now, the market is going absolutely batshit insane. I am not a stock analyst. I cannot give you financial advice. But right now, number going up. So maybe you could... I don't own stocks. I'm a psychopath. I live in cash. I invest in words. The problem we have right now is that NVIDIA is the largest stock on the stock market. There has never been a more problematic fact than that. 88% or more of their revenue is selling these fucking GPUs. You want to know what happens when you plug those GPUs in? They start losing you money. Nobody, not a single company other than Nvidia, is making any kind of profit on AI. In fact, they're burning billions. The reason I tell you this is you can't really navigate away from the bubble right now other than just selling, I imagine. So know what you're going into. Everything you're reading Oracle say right now about their relationship with OpenAI is bullshit. And to get specific is Oracle has a five-year $300 billion contract with OpenAI. Sounds amazing, right? They just need 4.5 gigawatts of data center capacity. Any guesses how much they have? 200 megawatts. Now, don't worry. That's less than 400 gigawatts, right? That's a lot less. That's a lot less. And I must be clear, I actually mean power. IT load, they've got about $1.30. So, not to worry, though, because OpenAI also doesn't have the money. Nevertheless, the stock is run. AMD now has a deal where they're going to sell chips to OpenAI. Who the fuck knows who's paying? NVIDIA, same deal. Whenever you see OpenAI involved, just don't. Just don't do it. There is no avoiding a bubble popping now. There just isn't. So, the smartest thing you can do is stick to the fundamentals. ([Time 0:31:12](https://share.snipd.com/snip/4499eff2-ec5a-4c23-82fe-76ee9e8d5162)) - **Bubbles Leave Varied Legacies** - Some bubbles leave useful infrastructure, others leave nothing but loss. - AI may produce useful tools and open models even if many firms fail. Transcript: Speaker 3 That would be nice. Cory Doctorow So some bubbles have productive residues and some don't, right? So Enron left nothing behind, right? Now WorldCom, which was a grotesque fraud, some of you will remember, right? They raised billions of dollars claiming that they had orders for fiber. They dug up the streets all over the world. They put fiber in the ground. They didn't have the orders for the fiber. They stole billions of dollars from everyday investors, people who just wanted to go through their old age without starving to death or not having a roof over their head. The CEO died in prison, and it was good riddance. But there was still all that fiber in the ground, right? So I've got two gigabits symmetrical fiber at home in Burbank because AT&T bought some old dark fiber from WorldCom because fiber lasts forever, right? It's just, it's glass. So once it's there, it is a productive residue, right? So what kind of bubbles are we living through? Well, crypto is not going to leave behind anything. Crypto is going to leave behind shitty Austrian economics and worse JPEGs, right? AI is actually going to leave behind some stuff. So if you want to think about like a post AI bubble world, and I just got edits from my editor. I wrote a book over the summer called The Reverse Centaur's Guide to Life After AI. And if you want to think about a post-AI world, imagine what you would do if GPUs were 10 cents on the dollar, if there were a lot of skilled applied statisticians looking for work, and If you had a bunch of open source models that had barely been optimized and had a lot of room at the bottom. ([Time 0:35:21](https://share.snipd.com/snip/1ddc4247-1fbf-42d3-87e9-d0a434f4880f)) - **Open Source Could Salvage Value** - Open-source models and skilled applied statisticians can repurpose AI after a market crash. - Community-driven tooling may create public-benefit uses beyond corporate monopolies. Transcript: Ed Zitron Ai gpus are mostly owned by private equity firms and big tech the majority of the gpus are not owned by people who will let them enter the market and there are really not a ton of like i if The idea is that gpus someone will work something out sure but you don't think all the king's horses and all the king's men would have come up with something else? Because that's the thing. The thing that I worry about, the thing that terrifies me about this bubble is, this is not useful infrastructure at all. Everyone loves to say, oh, it's just like the dot-com bubble. It's nothing like that. It's not even like the fiber. At least the fiber was somewhat useful. These GPUs, the amount of power alone. And one of the reasons that OpenAI and Anthropic have been able to have their monopolies is because they have the capital, and no one's going to have the capital to run these things. I mean, they'll be selling A100s. They'll be paying you 50 cents an hour. It's just what terrifies me is the, and I'm not saying you're doing this, that there are people already trying to rationalize this and say, well, and you were the one that actually told Me the stories of the dot-com where there were useful service space because it was before AWS. I don't see that usefulness here. So I take your point. Cory Doctorow I think that there's going to be a lot of firms in receivership, right? I don't think that private equity bosses' preferences are going to enter into it. I think that there's going to be a lot of firms in receivership. And I do think that when you contemplate the intersection of optimization of existing open source models, think about what happened when DeepSeq entered the market. You've got Chinese firms that are prohibited from using the more advanced GPUs. So rather than doing this sort of display of how serious they are about AI by spending as much money as they can, they went and they said, okay, well, what can we juice out of, you know, previous Generations of GPUs? And then they got some pretty impressive results. Yeah, but what result? Ed Zitron So they made it cheaper, sure. But what actually happened as a result of DeepSeek? Other than everyone, all of Silicon Valley went, didn't happen. Sure. They can't do it cheaper. Cory Doctorow They're Chinese. So my point is not about the market reality. It's the material reality. So I'm talking about what happens after the market pops, right? So I have seen people do interesting and useful things with AI. I think you've probably seen people do some useful and interesting. ([Time 0:36:59](https://share.snipd.com/snip/b532fe13-60b4-47a0-8d54-4e26ea9b5381)) - **Local Transcription With Whisper** - Cory used Whisper locally to transcribe 30 hours of podcasts in an hour on a laptop. - He completed the task without needing a GPU or heavy hardware. Transcript: Cory Doctorow I think you've probably seen people do some useful and interesting. I'll give you an example. I was writing an essay. And I couldn't remember where I'd heard a quote I'd heard in a podcast. I couldn't remember which quote it was. So I downloaded Whisper, which is an open source model from open AI to my laptop, which doesn't have a GPU, right? A little commodity laptop. I, through 30 hours of podcasts that I'd recently listened to at it, I got a full transcription in an hour. My fan didn't even turn on. That's awesome. Yeah. So I know tons of people who use this. ([Time 0:39:12](https://share.snipd.com/snip/bcb3ea12-ffa7-430a-a9b0-2753b1e55ae9)) - **National Champion Argument Backfires** - 'Protecting national champions' often shields domestic monopolists, not innovation. - Using foreign competition as justification preserves entrenched corporate power. Transcript: Cory Doctorow And protect large American firms to defend America against foreign firms. So for many, many years, the argument against breaking up AT&T was that they were our national champion and that they were keeping us safe from foreign aggressors. In the mid-1950s, AT&T was almost broken up. The DOD intervened to say that if we lost AT&T, if it wasn't intact, America might lose the Korean War. So AT&T won the Korean War because they got another 30 years, right? The only people arguably who won the Korean War were AT&T, including both sides in the Korean War. So when the 80s rolled around, right, there was this long, the 70s and the 80s rolled around, and we were once again thinking about breaking up AT&T, there was this long argument about How there was this belligerent foreign power in the Pacific Rim that they weren't original, they stole our IP and cloned our technology, and they would destroy our high-tech industries If we did not have a giant company to defend us against them. That country was Japan, right? Now, it turns out that the major project of AT&T was not defending America from Japan. It was preventing Americans from getting modems because AT&T really did not want you and anyone else in America to be able to provide a service to one another without them being able To veto it or charge rent on it, right? If you think about like the rollout of caller ID, $6.99 per month when it rolled out, right? This was what it cost you to find out who was calling you before you picked up the phone. You can't do that once people have modems. There is no caller ID for email, right? If your email provider says, until you click the message in the list pane, you don't get to find out who the from address is, you would just change email providers. So by controlling the network, by centralizing the network, they were able to just basically kneel on the throat of the American tech industry. And so it is always the case that when we defend monopolists in order to prevent foreign firms from destroying our high-tech sectors, what we're actually doing is we are defending the Firms that are structuring and controlling the market domestically for our own innovation. So these very large firms that are doing AI now, AI is only like one of the things they do. Obviously, we have Anthropic and OpenAI, but Microsoft, Google, Apple, Oracle, these are firms that are effectively market structures. They get to decide what products exist and what products don't exist, how much they're going to cost, and who can see them and who can use them. And when we defend them in the name of preventing China from pulling ahead in AI, what we're effectively doing is we're saying that we should maintain this shadow FTC that has more power Than the FTC ever managed to exercise, but that never exercises it in the interest of the American public, but only in the interest of their shareholders. ([Time 0:48:50](https://share.snipd.com/snip/a355ed0f-6b65-4be4-802e-c4271021a096)) - **Learn Finance To Counter Power** - Learn finance and corporate structures to pierce corporate obfuscation and hold power to account. - You don't need elite math skills; anyone can parse earnings and influence outcomes. Transcript: Ed Zitron Finance. It's interesting. I want to tell you a little story. So I taught myself economics in this past year and a half. I'm a fucking idiot. I just want to be really clear. I don't consider myself particularly smart. However you feel, it's up to you. Don't know shit from fuck. Those are technical terms. In all seriousness, finance is not as difficult or complex as it sounds. And indeed, the world runs on money. But also, these corporate structures are run fairly simply. But they dress them up in these confusing and annoying terms that actually, when you break them down, are pretty simple. There are people to blame. They've got houses. They're very flammable. But in all seriousness, these people, they're empowered through ignorance. They're not powered through ability, intelligence, or even making good products. They're powerful because we have, and I'm not trying to sound paranoid, this is just the truth, media that in many cases does not understand the things they're writing or indeed have The capability of reading an earnings statement. This stuff seems very scary. You may say, I'm not a numbers person. I couldn't possibly. You are a numbers person. It's not that. It really isn't. If I can do it, I failed mathematics years and years in a row. It really is fairly straightforward. They literally format things to make them boring. They word them to make them seem confusing. You can pick apart corporate structure. Any one of you can do it. That is where anyone can have an effect. I think a high schooler could do it. You want to change the world. You pull away their power. You take away their ability to obfuscate their wealth and the way they accumulate power by learning how these things work. I've written a lot about it. Everything you read, it's so long because I'm learning as I go. It's 10,000 words because I'm teaching myself like, what the fuck is this? Anyone can do this. You want to teach a young person anything? Honestly, teach them who the fuck Milton Friedman was. Teach them how the world got the way it got, but also learn about how companies are structured. It is not that complex. And the reason that you get knocked off course is because the powerful people go, no, it's not that simple. It's actually very mystical. Sam Altman is not just a con artist. He's not just a guy that's really good at convincing rich people to give him money. He has secret brilliance. No, he doesn't. None of these people do. They're McKinsey eggplants. These fucking people. Long story short, learn finance, seriously, and you don't have to go to school. ([Time 0:52:10](https://share.snipd.com/snip/7908df25-5177-4479-9ce1-de5f4b3bee42)) - **Choose Trades Over Uncertainty** - Consider vocational trades like electricians or HVAC instead of immediate university. - Skilled trades pay well, provide apprenticeships, union options, and durable demand. Transcript: Cory Doctorow Absolutely. There is so much work for electricians and we are going to be solarizing for the next 40 years. There's going to be infinity work for electricians. It's like being a plumber, but you don't have to touch poo. You can be, you can be an electrician who just does emergency call outs when money's getting low and you charge 500 bucks an hour. You can be an electrician on a cruise ship. You can be a theater kid. You can be an electrician on a job site. You can be an electrician for the government. You can be an electrician on a battleship. You can go abroad and be an electrician because electricity is the same, right? There's so, and it's, and it's interesting work and you get paid on the job, right? You get paid for your apprenticeship. And you can be in a union. And if you decide later you want to learn more and you like it, you can become an ENG. And if you don't, you can put yourself through college by being an electrician and learn finance. ([Time 0:55:15](https://share.snipd.com/snip/46fd096b-30eb-46ba-882c-56f6b428f5f0))