A show about AI, labor, and power. Veteran tech journalist Brian Merchant chronicles the growing backlash and resistance to Silicon Valley and big tech.
The two words the AI industry hates more than anything
Friday, October 2, 2026 • Duration 50:02
“Stochastic parrots,” a term coined in an influential 2021 paper written by two AI researchers and two computational linguists, essentially holds that large language models are not “intelligent” but are rather systems that output material based on what’s in their training data. Now, while LLMs and technologies described as AI today have become more sophisticated and complex over the last 5 years, the term remains a fundamentally accurate way of describing LLMs.
Yet the industry hates it like the plague. AI industry folks absolutely lose their minds over the term, and often bring it up, unprompted, just to try to debunk it again and again.
Why does this particular term have a unique capacity to enrage AI industry folks? I had a theory that, even more so than other critiques about shaky financing or dubious functionality, stochastic parrots undercuts the very foundation that the AI boom is built on: That tech companies are building a machine god of infinite power and potential. These firms have relied on narrative to build up and to sell to the world a very expensive, resource-consumptive technology. If it turns out it’s an unthinking system that’s often fallible in predictable ways, well, that might not be worth quite as many trillions of dollars.
To discuss this theory in greater depth, I thought I’d go right to the source—to one of the paper’s co-authors, professor Emily M. Bender, of the University of Washington.
Quick PSA: I can only do this work thanks to readers, listeners, and viewers who chip in a few bucks each month to help me keep the lights on. If you’re one of them, you have my sincere thanks; you’re the reason I can cover subjects and do the sort of reporting so few others will. If you’re a fan of the “Great 2026 Blood in the Machine Podcast Experiment” or appreciate what we’re doing here, consider upgrading to a paid subscription; it takes no small amount of work to book great guests, research, write, and produce the thing, and we’re early days yet. Okay! That’s enough from me; thanks for tuning in, and hammers up.
While billionaires and world leaders expound the dangers of all-powerful, “super intelligent” AI, the reality on the ground is that it's a tech product that’s only made possible by countless human beings, laboring for sometimes dollars a day, doing work like “cleaning” data sets so they’re usable, scanning books and literature into databases, creating usable content, and moderating the output of AI systems to weed out offensive, gruesome, and racist content. These workers are underpaid, overtaxed, and fighting for better conditions—even as AI executives are declaring that the systems they prop up are so powerful they’re about to extinguish life on the planet.
In the meantime, workers like Joan Kinyua and Michael Geoffery Asia of the Data Labelers Association just want fair pay and basic protections. I recently spoke to both of them about what it’s like to perform crucial work day in and day out and yet be completely invisible to the company that relies on your labor—and to the world at large. To help feed Silicon Valley’s illusion that these systems are actually machine gods. And to fight for better conditions, rights and recognition in spite of it all.
In a year rife with backlash to big tech, data centers, and AI, one company still manages to stand out: Flock Safety. Over the past decade, Flock has been building out a massive network of surveillance cameras across the nation. The images those cameras record are uploaded to a central database, and can be accessed by law enforcement and federal authorities.
Abuse of the system has been rampant. Police have used Flock’s Automatic License Plate Readers (ALPRs) to stalk exes and to track women seeking abortions across state lines. ICE has accessed Flock data to assist in crackdown on immigrants. Innocent people have been wrongfully imprisoned after being flagged by the system.
With the anti-Flock movement growing, and more people looking into how to get Flock out of their communities, I sat down with Ed, an organizer with DeFlock, to explain, step by step, how you can start a movement to get the surveillance tech out of your town, too.
AI executives like Anthropic’s Dario Amodei and OpenAI’s Sam Altman have been forecasting an incipient AI jobs apocalypse for three years now. So where is it? Well it’s here, kind of. But instead of replacing tens of millions of workers, it’s mostly making them miserable.
The revolt against Meta’s ‘pervert glasses’ is part of a growing storm
Friday, August 28, 2026 • Duration 01:05:57
New fronts are opening up in the opposition to big tech. The anti-data center movement is rolling, AI is unpopular as ever, and now, a widespread backlash to Meta’s new AI glasses—termed Pervert Glasses by the masses—is unfolding at breakneck pace. Silicon Valley, of course, is in denial about all of it.
This week on the podcast, to help make sense of this latest revolt against a Silicon Valley project, I spoke with Chris Gilliard, an author, thinker, and surveillance scholar who’s spent many years investigating big tech’s efforts to seduce us with sleek, aspirational products, track us, and harvest our data. Gilliard is the author of Luxury Surveillance, a forthcoming book about this very subject from MIT Press; you may better know him as Hypervisible, the handle he uses online and on Bluesky.
It was a great chat, and Gilliard’s work is absolutely vital. Hope you enjoy.
In case you missed it, I wrote about some of the topics we discuss above in this week’s column:
Other stories mentioned in the segment and the interview:
Going back decades, there’s been a faction in Silicon Valley whose chief interest was not coding, consumer technology or even business, but power. In some ways, the story of the last ten years of the tech industry has been the story of that faction, perhaps best embodied by Peter Thiel and Elon Musk, but also represented by figures like Marc Andreessen and figures like Alex Karp, Palmer Luckey, and David Sacks, winning the power they sought. The consequences of that victory have been disastrous.
Gil Duran, a veteran journalist and California political operative, is the author of the Nerd Reich: Silicon Valley Fascism and the War on Democracy, a book about how, exactly, this disaster has unfolded, and where it’s left us. That is, with a tech industry that’s overwhelmingly eager to abet, enable, and profit from authoritarian governments, by wielding power within them, and by building its tools of surveillance, automation, and extraction.
Duran calls the key bloc driving this push ‘tech fascists’, and many have openly called for an end to democracy, or for rule by monarch—or CEO. And it is hard to overstate the import of the ascendency of the tech right.
These figures now have a direct line to the White House, have one of their own in the White House in former VC JD Vance, and are often in the White House themselves. They hold massive government contracts and their unprecedented hoards of lobbying cash influence shapes public policy, in addition to controlling much of the digital architecture that guides our daily lives.
The results of all this are everywhere you look. This is, yes, how you get DOGE running wild in the federal government, laying off hundreds of thousands of civil servants and pulling USAID funding for the most vulnerable, leaving people dead as a result. It’s how you get the widespread cronyism and corruption in Washington, with government figures embracing crypto and the AI industry and enriching themselves in the process. It’s also how you get tech CEOs rising into a class that more resembles an oligarchy than the mere ultra-wealthy.
One reason Facebook and Meta have operated with such impunity over the years is that Peter Thiel helped Mark Zuckerberg structure his company in such a way that he would retain ‘supervoting shares’ and would be unaccountable to a board of directors or anyone else. Just this week, Anthropic CEO Dario Amodei ; he will essentially be impervious to boardroom dissent or activist shareholders. This means, for instance, that as Claude users revolt over fears that the company’s new watermarking process—which was an important win for everyone who opposes the internet being coated in slop, and the mass deskilling of creative workers—will , it will ultimately be up to Amodei himself to decide what to do, and no one else.
Get full access to Blood in the Machine at
The State of the AI Bubble with Ed Zitron
Friday, August 14, 2026 • Duration 01:02:59
There are fresh signs that the AI bubble is wobbling. Again.
In recent weeks, a much-vaunted AI-focused hedge fund, Situational Awareness, run by Valley wunderkind Leopold Aschenbrenner lost $35 billion; nearly its entire fund. South Korea saw its largest-ever stock market collapse as investors pulled bets on AI chips and tech stocks. The circularity of the AI economy, with hyperscalers investing in AI labs and AI labs purchasing cloud compute from the hyperscalers and hyperscalers reporting big profits, is reaching new heights of precarity. Due to a fresh series of moratoria and political opposition, the future of the massive data center buildout has never looked more uncertain. The US markets have dipped then boomed, dipped then boomed.
Where does all of this leave us? Just about everybody—Sam Altman, the Financial Times, AI critics, your local barista—seems to agree that we’re still, somehow, improbably, in some kind of a bubble. But how big? How volatile? When will it burst? How? This week, I dive into all these questions with the infamous AI skeptic, analyst, and industry scapegoat Ed Zitron, author of the Where’s Your Ed At newsletter and CEO of EZPR.
Elon Musk is not the richest man in the world merely because he’s good at selling cars or rockets at a profitable margin. That might have sufficed for richest men in the world of decades past. But Musk became the world’s first trillionaire, however briefly, for a lot of reasons, and being good at moving a particular productis not chief among them. Musk sells a vision. Musk harnesses the state to implement his corporate goals, to secure his companies’ foundations beyond the vicissitudes of common capitalism. Musk mediates his message on social platforms, which he owns and controls, and uses to bend politics to his will. The list goes on.
This formula has been emulated by the most recent crop of tech titans; your Sam Altmans and Dario Amodeis, whose path to power lies in selling a vision over simple goods, who work the state to buffet their businesses, and who paint a bleak picture of the future through which their products will help you survive. To understand what’s going on here, not just with Tesla or SpaceX or OpenAI or Anthropic, but the modern American economy itself, I sat down with the historian and author Quinn Slobodian, who has put a name to this tendency: Muskism.
Slobodian is the co-author, with Ben Tarnoff, of the book by the same name, which I fully and heartily recommend. We dig into all of the above, and a lot more; including the under-examined role that growing up in apartheid South Africa played in influencing Musk’s present-day ideology and tastes.
I very much enjoyed this discussion, and think you will too. Understanding Muskism, after all, could not be a more urgent project.
Xbox workers speak out about layoffs, AI, and how big tech brought a crisis to gaming
Friday, July 17, 2026 • Duration 54:23
On July 6th, Microsoft announced that it was laying off 4,800 workers, largely from its Xbox gaming division. It’s now sadly routine in the tech industry; thousands of jobs gone in an instant, with little more than a corporate memo and some platitudes about AI.
To better understand what actually happens behind the scenes in a big tech layoff event, and to understand the particular crisis faced by the gaming industry, this week I sat down with three Microsoft workers for a special edition of Blood in the Machine. Two of these workers have been impacted by the layoffs, one saw longtime friends lose their jobs. All three are fighting back. This is how.
A brief note about how BITM is working these days: Subscribers get (at least) one written newsletter edition sent to their inboxes each week, typically, ideally, Tuesday or Wednesday, as well as one episode of Blood in the Machine: The Show, which runs every Friday. But I can only do all of this with the support of readers like you, so if you find value in independent, critical tech journalism, and you’re able, please consider a paid subscription. I want to keep as much of this un-paywalled as absolutely possible, and I can only do that when folks chip in. Many thanks to all who do so, it’s very much appreciated. Cheers, onward, and hammers up.
Members of what Duran calls the Nerd Reich then use their power to further concentrate wealth, and of course, more power. A report by the AFL-CIO’s executive pay watch found this week that Elon Musk’s Tesla pay in 2025—he took home $158 billion—was an astonishing 2.5 million times his average workers’ salary. As the lead researcher on the report told Fortune, it’s ‘unlike anything we have seen before.’
Tech CEOs operating at such a distance from ordinary people may also explain why they’re so willing to undertake practices and push products that large swaths of the public absolutely hates. We already well know that data center opposition is the single political issue most Americans are united against, and that Gen Z is vehemently anti-AI. But consider also Meta’s marquee new product—”AI glasses” that allow users to film people without their consent, and that have already been dubbed “pervert glasses” by the masses. Just this week, Futurism published a report about how the glasses are being used by bullies in schools to harass their peers, especially girls. The bullies are apparently aping tactics seen on viral videos, in which pickup artists film women they’re hitting on without consent; hence the pervert glasses name.
This is perhaps, providing fuel for another future lawsuit—this week Meta goes to court against twenty-nine US state attorneys general who accuse the company of knowingly designing addictive products that harm the mental health of children. They’re seeking $200 billion in damages.
It might also explain why tech CEOs are willing to scan and destroy books en masse to feed them into their AI systems, despite the horror most people experience when they hear about such a practice. In, a story that’s almost too on the nose for a segment discussing the Nerd Reich, 404 Media had a rare bookseller put an AirTag in a rare book bought by a mystery buyer, and it wound up in an Amazon Facility expressly designed to scan and mulch books to train the company’s large language models. At the very same time, meanwhile, AI-generated books are flooding Amazon’s online marketplace, and crowding out real authors, who are earning less and less for their own, human-authored books.
As a result of all of the above, there is a powerful, enraged backlash growing against the nascent tech oligarchs, multibillionaire AI CEOS, and the Nerd Reich; whatever you’d like to call them. I’ve been covering the tech sector for nearly two decades now, and I’ve never seen anything like this. People are shouting down big tech reps and data center proponents at public meetings. Flock, the private provider of police surveillance equipment, has been under siege for months; a national week of action against their product, ALPRs, is underway right now, as I record this. A survey conducted by CNBC found that young people aged 18-34 found tech CEOs overwhelmingly untrustworthy. “Young People Hate AI CEOs So Passionately That It’s Almost Hard to Believe,” as Futurism put it.
Even Elon Musk’s own AI bot, Grok, has appeared to turn against him, accusing him of watching child porn and calling for his assassination on X.
The tech elite may have the wealth, the influence, and the ear of the president, but not even that can insulate them from a growing mass movement, increasingly intent on taking them on. And we have every reason to think that what we’re seeing now is only the beginning.
So that’s episode number 7? 8? depending on how you count? of the podcast in the bag there. I tried something a little different this time, trading the headline roundup/AI resistance digest in the beginning for more of a short essay/composed intro. Not sure which I prefer, tbh, though I did enjoy the chat with Duran, who has laid out the key strands of modern Silicon Valley’s dominating ideology in blunt and easily comprehensible terms, regardless.
In fact, let me know what you think in general here. I’m happy with, and proud of, the work that I’ve put together here with my excellent producer Ryan Hodes. We’ve covered a lot of ground over the last couple of months, from checking in on the New Luddite movement and discussing their wildly successful Summer of Ludd in NYC,
tried to answer the question of ‘What is an Elon Musk’ with author and historian Quinn Slobodian
sat down with laid-off Xbox workers who are still organizing to fight back against their employer, Microsoft (this was, at least to me, such a powerful and impactful conversation)
discussed the state of the AI bubble with the world’s leading (and loudest) industry bear, Ed Zitron
dug into how AI and crypto companies are pouring hundreds of millions of dollars into the current election cycle to buy favor and influence with the great Molly White,
and talked about how to de-Google our lives with tech critic extraordinaire Paris Marx.
Anyway, every one of the guests have been great, in my humble and unbiased opinion, and I’ve enjoyed all these conversations. HOWEVER, getting the thing up and running has been a lot more legwork than anticipated. I feel like I’ve already written some variation of that sentence or another one just like it a time or two already, but it’s true. I never meant for the podcast to replace or crowd out the newsletter; it’s supposed to be the opposite. The show is supposed to augment the writing and reporting I do here, and amplify it in a new medium or two.
But I admit I’ve gotten sucked into the show a bit, planning out future episodes, booking great guests, writing and editing scripts and intro copy, wrangling YouTube algorithms, trying to better understand podcast distribution, and so on. The good news is that attention seems to be paying off. After starting from scratch, the YouTube channel got over 100k views in its last period, which my YouTube-native producer tells me is good. The podcast is outpacing expectations as well, and is finding an audience even outside these pages, which was always the idea. And it’s all growing, with some big algorithmic bursts here and there and steady growth underneath.
Yet I don’t want it to eat into my writing and reporting time so much; I don’t want to be a YouTube guy, ultimately. I want to be a guy whose writing people find on YouTube. I imagined a model of a) writing the newsletter, b) doing an abridged audio version of that plus updates on the general tech/AI resistance, criticism and backlash, plus c) having a conversation with a person or org in the thick of all of the above. And I do think after a couple months of ironing things out, we’ve hit on a format and workflow that will make that model more of a reality. So yes, I wrote a lot less over the last month or two (travel and dad stuff has played a minor role, too), but next week we’ll get the intended model rolling in earnest, with a piece planned for Tuesday publication, and we’ll see if we can’t have our cake and eat it too.
Do let me know what you think of this whole experiment. I would love to hear everyone’s thoughts, what you think is working and what isn’t, and what you’d like more or less of. The comments are open as always, or I can be reached at brian@bloodinthemachine.com.
Thanks as always for reading and listening. Hammers up.
Discover shows related to Blood in the Machine, based on actual content similarities. Explore podcasts with similar topics, themes, and formats, backed by real data.
