This week, California governor Gavin Newsom signed a spate of new laws that establish some of the strongest AI protections for workers in the United States. Okay, they’re some of the only AI labor protections in the nation, but it’s good news nonetheless.
They cover a lot of ground: One bans managers from using AI to discipline or fire workers without human oversight, one outlaws any kind of AI surveillance in workplace bathrooms, one bans the use of AI to surveil or predict workers’ emotional states—and yes, a growing number of companies really try to do this—while another designates that licensed healthcare workers must remain the decision makers in clinical settings; not AI.
These laws join a number of already-passed bills that aim to protect workers from automation. Earlier this month, Newsom had signed another crop of bills that prevent bosses from replacing workers with AI outright; one prevents administrators from replacing California State University faculty with AI, another designates that public school employees and contractors must be human, and another requires companies to label if a performer is AI-generated.
“This is a breakthrough moment in tech policy,” Annette Bernhardt, Senior Tech Policy Advisor at the UC Berkeley Labor Center, tells me. “Right now, it’s the wild west when it comes to how employers use electronic monitoring and algorithmic management, because there are literally no rules of the road. For sure there’s much more to be done to build out a robust set of worker technology rights for the 21st century labor market, but California’s unions have put the first big stake in the ground.”
In other words: promising stuff! While some of these bills are narrow, and others were scaled down in compromises to secure passage, there are genuinely important protections, guardrails, and lines drawn in the sand on behalf of human workers here. Crucially, the new laws begin to form a framework that advances worker rights in the age of ubiquitous enterprise AI.
This is much-needed, and it’s an achievement that represents a real victory for labor. The California Federation of Labor Unions, led by Lorena Gonzalez, as well as the AFL-CIO, the Teamsters, the California Nurses Association, and many others deserve a ton of credit for fighting this often thankless fight against big tech and the pro-AI lobby, which marshaled their considerable resources to crush similar past efforts. “Today, California’s workers and our unions have changed the national narrative on how Americans can fight back and win against AI taking over our jobs and workplaces,” as Gonzalez put it in a statement.
I’ve covered the intense opposition to AI labor bills in California—I got more into the weeds on the issues last year, when Newsom vetoed or punted more AI labor bills—and seen firsthand the efforts to shut them down. Legislators and labor advocates have told me that they’ve never seen anything quite like it; the tech lobby has opened its war chest for state level legislative fights on a level it never has before. Its efforts to quash AI regulation have been relentless.
"Despite millions of dollars channeled through a massive corporate-led campaign by the Chamber of Commerce, the tech industry, and their allies, workers fought back and won,” Samantha Gordon, the Chief Advocacy Officer of TechEquity, which backed the pro-labor AI bills, told me. “These bills ensure that workers can't be fired by an algorithm, surveilled without limits, or replaced without a safety net. Policies like these are wildly popular with Californians.”
Elsewhere, Newsom signed the COMPETE act, contentious, pro-worker piece of anti-trust legislation that was watered down before passage to the extent that one of its chief backers all but disowned it.
Newsom also signed a new law to fund local newsrooms; outlets that publish original reporting will be eligible for a $20,000 tax credit per full-time reporter for up to five employees, $15,000 for each after that, and $7,500 for part-time journalists. This, too, is an AI labor law.
“The Community NEWS Act just signed by Governor Gavin Newsom is significant for a lot of reasons,” says , a former colleague of mine at the LA Times and the current Director of Policy for Rebuild Local News, an org that pushed for the journalism bill. “But among the most important in the AI era is that the law explicitly supports human beings, by making it cheaper to employ the local journalists who do local journalism. The incentives are clear. If you lay off reporters to replace them with robots, the tax credits that supported reporters go away. It literally changes the math of automation.” (Pearce sent me so many thoughts on the passage of the news bill that I’m including them all as a sort of mini-essay below, at his encouragement.)
It’s not all good news this week, however; Newsom vetoed a number of promising bills, too, including one that would have protected nurses who challenge unsafe uses of AI in hospitals and one that would have required public employers to give unions 45 days notice before they purchase or implement AI programs that will impact their workers.
The labor law scholar Veena Dubal has a good threat on all the bills that passed this week, and that were vetoed, for more details:
“Today, Gov. Newsom signed the Federation’s following sponsored bills into law:
SB 947 (McNerney) which bans bosses from using AI to discipline or fire workers without human oversight and corroborating data. In 2025, Governor Newsom vetoed SB 7, a previous version of the No AI Bosses Act. (Co-sponsored by the Labor Federation and the Teamsters)
SB 951 (Reyes) which modernizes the WARN Act to provide workers additional information on layoffs caused by AI and directs the EDD to study the impact of AI on hiring and the workforce. (Sponsored by the Labor Federation)
AB 1331 (Elhawary) which bans the use of any AI surveillance in workplace bathrooms. (Co-sponsored by the Labor Federation and the Teamsters)
AB 1979 (Bonta) which requires licensed health care professionals to remain the decision-maker in health care clinical practice, not AI. (Sponsored by CNA)
“Earlier this month, Governor Newsom also signed the following sponsored and Labor-supported legislation taking on Big Tech and regulating AI:
SB 928 (Cervantes), which prohibits CSU faculty from being replaced by AI. (Sponsored by the CFA)
AB 2148 (Muratsuchi), which ensures that public school employees and contractors must be people and not AI. (Sponsored by the CFT)
AB 1609 (Zbur): Requires business to provide human customer service reps within 15 minutes of a request. (Sponsored by CWA)
AB 1776 (Aguiar-Curry), which gives the state more tools to combat corporate monopolies that drive down wages, make it harder to unionize and raise prices for consumers. (Co-sponsored by the Labor Federation, UFCW, the Teamsters, NNU-CNA, WGAW and SEIU California, UDW)
SB 1050 (Ashby), which requires advertisers to disclose if a performer is created by AI. (Sponsored by SAG-AFTRA)”
As promising as a lot of this is, as Gordon notes, it’s still early days.
“The fact that passing such common sense protections required a massive groundswell of workers, advocates, and community leaders reveals a great deal about these companies’ aspirations and long-term plans for all of us,” she said. “We look forward to building on these protections and continuing to advance a people-first vision for AI policy in California.”
And workers everywhere should look forward to continuing to fight against corporations angling to embrace mass automation and surveillance, and against big tech’s continued, AI-focused concentration of power. That fight has just begun.
The Community NEWS Act just signed by Governor Gavin Newsom is significant for a lot of reasons. But among the most important in the AI era is that the law explicitly supports human beings, by making it cheaper to employ the local journalists who do local journalism. The incentives are clear. If you lay off reporters to replace them with robots, the tax credits that supported reporters go away. It literally changes the math of automation.
Well, why does it matter if we have human journalists doing local news instead of having robots scraping live feeds of city council meetings and doing automated writeups? Working journalists know that information is made, not found: Local reporters are still often the proverbial coal miners powering the information grid, by digging up original interviews or unpublished documents that would otherwise not be found in the public domain.
But this kind of work, because it can’t be so easily automated, is often cost-prohibitive because you’re talking about paying a human for journalism that may not earn back its cost because of the hard math of subscription, advertising and grant revenue in the digital information era. If we want quality information getting injected into our information ecosystem, we’re going to have to pay for it somehow, because it increasingly can’t pay for itself.
One of the other hidden costs of automation is the destruction of human talent the local news industry needs in order to keep evolving to keep up with shifting consumer demand. The U.S. has lost more than 80% of its local journalists since 2002, and California alone has lost more than 12,000 local journalist jobs. This hasn’t just hurt the amount of news we deliver to communities but it’s hurt our own ability to adapt. Human beings learn by doing and get better at their work over time; they innovate; they become capable of more ambitious work or exploit entrepreneurial opportunities. To shrink your talent pipeline is, in the long run, a means of reducing your own future.
In his typically rambling, occasionally genocidal speech to the United Nations, one that included threats to “annihilate” Iran and to bring “freedom” to the “failed state” of Cuba, Donald Trump announced that he was issuing a nationwide rebranding of “artificial intelligence” to “super intelligence.”
You can tell when Trump has particular relish for a certain part of his speech, the way he pauses and chews it over, luxuriating in the idea rather than rattling it off or spitting it out. If I had to guess, I’d venture that the root of this particular digression/scheme is that Trump never really thought much about AI until a couple years ago, and so has probably associated “artificial” more with words like “sweetener” than “intelligence,” and has thus chafed at the way the qualifier diminishes one of his signature projects. A word that scans to Trump as “fake” simply won’t do for what he considers a reflection of his power and standing.
Trump has of course always had an affinity for expressing dominance and blunt superlatives, and so, we have artificial intelligence renamed “super intelligence” or “S-I” on US government documents a la Lake Ontario nominally becoming “Lake America.” Trump means “super” in a way that’s different than the Silicon Valley guys who use “artificial super-intelligence” or “ASI”; it’s vibes-based and dependent on jettisoning the “artificial.” It’s “super” as in “Superman.” (I’m also willing to bet every last one of my Trump Coins that there was at least one failed effort from the Trump brain trust to try to work out how they might shoehorn American Intelligence or Make Intelligence Great Again into the equation somehow.)
All this has garnered plenty of justified snickering from the chattering classes, but it’s worth noting that “artificial intelligence” was itself coined as a branding exercise, and then used to tell, and embellish, a story.
Ringlus by Avo Keerend. 1973. Tartu Art Museum, Estonia. Via Unsplash.
This edition is a two-parter: First, a look at the stories being told about AI, and who those stories serve. Second, an excerpt from Against Tech Oligarchy, a great new book by tech worker organizers Clarissa Redwine and JS Tan, who are on book tour as we speak, and will be in LA, at Skylight Books, this Friday, September 25th, with friend of BITM Joanne McNeil. I’m going to try to make it too! Hope to see you there.
As always, all of this is made possible by subscribers who chip in to support this work—a huge cheers to everyone who does so, I can’t thank you enough. The writing, reporting, the podcast; all of it is made possible by supporters. If you find value in it, consider upgrading so I can continue to do it. Many thanks, and onwards we go.
“I invented the term artificial intelligence,” the Stanford computer scientist John McCarthy said in 1973, “and I invented it because we had to do something when we were trying to get money for a summer study in 1956.” A more pedestrian, jargon-filled proposal had previously failed to get traction; “artificial intelligence” turned committee heads and got the funding. Now, angling for some summer study funding is on a much smaller scale than the trillion-dollar value-inflating proclamations of our AI companies today, but the anecdote shows the extent to which “AI” has always been a cultural construct and has served as a useful narrative device.
And the narrative is really heating up these days, is it not. In Trump’s telling, AI is a vector of American power, one that must not be restrained by “globalist schemes” to regulate it, as he put it. This was essentially the same stance that Vance took last year, when he delivered his first big speech abroad as vice president; that AI is a vessel for American dominance, that it’s not to be tied down with regulation, and that everyone else (especially Europe) should either get on board or out of the way.
It has its share of adherents, especially on the tech right, but the story about an unconstrained AI superpower makes for a less palatable narrative coming on the heels of weeks of sustained discourse, fanned by the leaders in the AI companies themselves, about another one: That AI may soon exterminate all of humanity. With a public that’s still unsure of how to process that significantly more alarming narrative, and Trump’s own base increasingly out of step with him on the issue, the AI-as-nationalist-dominator thrust is dying on the vine. (It also surely does not help that even diehard Trump fans may struggle to see how “super intelligence” benefits them, as their electricity bills are getting bigger, the war in Iran drags on, and everyone’s worried about losing their jobs.)
So it’s little surprise that the story preferred by the media appears to be that Trump has an opportunity to breach “AI safety” issues with China during Xi’s visit this week, despite Trump’s obvious distaste for the notion, and the chances of anything substantive emerging in that department remaining slim.
“I think Presidents Trump and Xi would get the Nobel Peace Prize if they could agree on something,” OpenAI CEO Sam Altman told Fortune, in a comment that could not have been more obviously intended to flatter a vainglorious president who laboriously covets the award. It also speaks to how Altman views the import of his own undertaking; ie, also worthy of Nobel Prize-scale consideration. And it certainly fits better with the story that has taken shape and then center stage over the last two weeks, ever since former Anthropic employee Jacob Coxon resigned over stated concerns that AI labs weren’t doing enough to prevent incipient human extinction: That the largest AI labs must coordinate with China and its own top AI labs to “pace the frontier” and prevent us from “losing control” of a “super intelligence” from “going rogue” and harming us, perhaps even killing us all.
If that sounds like science fiction, well, it’s because science fiction is a major cultural touchpoint for executives and researchers at the AI labs, and because they want all of this to read like science fiction. Just as Trump’s preferred super intelligence language reflects his own narrative interests, so too does the language preferred by the labs (“extinction risk,” “p(doom),” “AGI”) reflect theirs. “Going rogue,” the language used to describe incidents like OpenAI’s large language models precipitating a hack on open source AI competitor Hugging Face, sounds cyberpunky and exciting. It also happens to absolve the company of accountability, both culturally, and, apparently, legally; a precedent OpenAI and Anthropic alike would very much like to cement.
Lots has been written about how the “going rogue” is a misnomer; here’s esteemed computer scientist Melanie Mitchell, the technologist Eryk Salvaggio, the requirer-of-no-introductions Cory Doctorow, and the computer science professor Cal Newport explaining why. All basically underline the point that LLMs, when prompted with specific instructions, after having key guardrails removed and left unobserved for long periods of time, can be expected to follow those instructions, and draw from material included in their training data that outlines how to hack other systems to do so. This certainly doesn’t mean there’s nothing to worry about, but that we should be more worried about the ethics of the companies developing these systems and malign actors exploiting them than the technology taking on a life of its own or “going rogue.”
I am also skeptical of this language because I have documented rather extensively, in this report and elsewhere, the way that AI companies have historically occasioned to deploy it for their own benefit. And right now, those companies see a good deal of benefit in promulgating a narrative that a handful of exceptional companies are building a worldshaking AI system. In fact, let us count the ways.
First, the obvious one; that the companies are on the brink of holding IPOs that will make their executives unfathomably rich. A system so powerful it can destroy the world holds significant investor value to parties who are accustomed to funding enterprises who do that more incrementally. These narratives, as they have all along, have fed an investment boom that is, for the time being, lifting a great many boats.
But there’s also the issue of regulatory capture. The AI extinction narrative favored by the big AI labs is tailored to conclude with a number of solutions proposed by and friendly to the industry; compliance with safety standards, audits, oversight bodies, independent observers and so on. The advantages here are twofold; the AI industry gets to set the terms of how they’re regulated—ideally eschewing anything like “being held accountable for things like hacking other companies, corporate malfeasance, instructing children to harm themselves, surveilling workers, perpetrating systemic biases in hiring and firing” and so on. Second, meeting the proper “safety” thresholds would be organizationally challenging for all but the most well-prepared, connected, and capitalized companies to manage, conferring an advantage on the leading AI companies and allowing them to cement their lead.
But perhaps the most compelling reason that this new AI extinction narrative ultimately serves the AI companies above all is that it gives this very expensive, money-hemorrhaging industry a means of coordinating a “slowdown” without running afoul of antitrust laws. Remember, the AI companies are still plowing unprecedented gobs of capital into training and running their models every year; it’s enormously costly. If they had a reason to pause all that training, and an excuse to simply sell their products to consumers as are, it would give them a much better shot at becoming profitable in or around IPO season. (It would also, perhaps, help draw down the need for a rapid, expansionist data center buildout, the most vocal, spirited, and public front of opposition to the industry.)
Anthropic, OpenAI, and perhaps a couple of other frontier labs are the dominant providers of frontier AI models.
They can charge customers a lot of money for using their frontier models, and rather less money for older, no-longer-cutting-edge models.
Training a new frontier model requires ever-increasing billions of dollars of computing power.
The labs need to more or less continuously race to build new frontier models, because their competitors are all doing it, and if they don’t they will fall behind and no longer be able to charge a lot of money for their best models. (Also because they intrinsically want to build artificial superintelligence, for cancer-curing and/or killing-everyone reasons.)
If they collectively slowed down, then (1) they’d spend less on compute and (2) they’d be able to charge frontier-model prices for a longer time.
If just one lab slowed down, the others would eat its lunch. So you need a reason for undertaking what is essentially industrywide collusion to lower product development costs; if you’re following legally mandated regulations, it’s not an antitrust violation, it’s meeting the letter of the law.
To justify such regulations, well, you need a story.
The day after Trump’s unhinged super intelligence speech, the UN Security Council convened a meeting with top AI leaders, including OpenAI CEO Sam Altman, Anthropic chief Dario Amodei, and the head of Hugging Face, Clement Delangue. Altman solemnly intoned that whether it be a 10 or 1 or 12 percent chance that his products might wipe out the human race, he said, bravely, that no number was acceptable. So, guess what most of the major AI companies proposed: Regulation, on their terms.
“We are at a crossroads,” Sam Altman said, seated in the Security Council wearing his trademark suit with a baby blue tie. “One path leads to a world where power is concentrated, decisions are opaque, and people feel that the most important technology of their lifetime is something being done to them. The other path leads to a world where AI works for people, because people, through democratic institutions, help decide how it should develop, and guide it.” (I will quickly note two things here: First, this is very very rich of Altman to appeal to the democratic process when OpenAI is spendingmillionson lobbying to crush that process, and second, let’s all bear in mind the last time Altman did the public rounds asking to be regulated; when he didn’t like the regulations governments proposed, he jumped ship.)
“AI is advancing very quickly,” Amodei said when it was his turn to beam into the meeting over Zoom. “The trajectory only needs to continue to continue for a very tiny period, maybe one year or less, to reach what I’ve called ‘a country of geniuses in a data center’.” (I will quickly note just one thing here: Amodei has routinely predicted massive disruptive change—according to him, there should be no junior software engineers by now—just on the horizon.)
Both men highlighted their companies’ new products and achievements; Altman mentioned that OpenAI’s LLMs had solved a long-unsolved math problem called Navier-Stokes (a growing number of mathematicians contest OpenAI’s accomplishment, and it seems that the AI giant likely acted unethically in leaning on the work of a mathematician who was close to solving it, too) and Amodei highlighted a new AI-derived advance in gene therapy. Both reiterated a call for global governance of AI: Frameworks for slowing the pace of model development, calls to ban biological weapons development, frameworks for international cooperation on threat assessment. Highfalutin, and self-serving, stuff.
Right now there is what I might describe as unrestrained scorn and even outrage at anyone who would still suggest hype and marketing play a role in the formation of the AI domination and/or extinction narratives. The fury at the scholars Timnit Gebru and Emily Bender, who have continued to debunk the must overheated claims, has been especially pronounced. But even if you believe AI systems do present a novel threat to civilizational infrastructure, it seems the more grievous error to ignore how the narratives about those threats are being shaped and wielded in pursuit of very specific ends—and, as of now, I might add, very successfully so!
The AI companies are continuing to grow incomparably rich, setting the terms for policy, succeeding thus far in their bid for regulatory capture, and avoiding legal accountability for their systems’ automated exploits.
In other words, why aren’t more people, including analysts and tech journalists, skeptical of this story, which is serving to concentrate power for a handful of vested parties at an extraordinary clip, rather than the story’s critics?
When it was his turn, his Zoom connection glitching in and out, Hugging Face head Clement Delangue argued that the biggest threat isn’t powerful AI, but “asymmetry… Asymmetry between attackers and defenders. Asymmetry between a few companies, and everyone else. Between a few countries, and the rest of the world. Asymmetry of compute, and control—and asymmetry of power.”
The stories that both Trump and the tech executives tell about AI have this in common: They’re both ultimately predicated on keeping that symmetry of power tilted in their favor.
And now, an excerpt from Against Tech Oligarchy by JS Tan and Clarissa Redwine, who, again, are coming to LA for a book talk on September 25th at Skylight Books. RSVP today to secure your spot. Here’s JS and Clarissa:
How tech workers can outwit, out-organize, and overcome their oligarch bosses
As tech bosses fall in line with the right, tech workers have begun fighting back. This is the inside story of their movement—and the way it spawned an anti-worker backlash now reshaping the industry. The following is an excerpt from this book that lays bare the challenges ahead for anyone in tech looking to wield worker power and hold the rising oligarchy to account. Hope you enjoy.
When the traditional tactics of organizing do not threaten their employers’ bottom line, workers must reflect on their own labor process and be willing to try out new strategies to maximize their leverage over their employers. Luckily for tech workers, such examples are not uncommon in labor’s history.
In 1934, after several clashes with the boss and deadly assaults from the police, striking truck drivers in Minneapolis experimented with a new picket line strategy that paid off. Under increasingly hostile conditions, with both employer and state aligning to break the strike, workers watched their leverage disappear as the picket was banned by the city and scabbing truck drivers were issued military permits to ensure they could keep trucks moving no matter what. In response, workers patrolled the city, swarmed scabbing drivers, and made sure the trucks never reached their destination. This tactic was uniquely suited to the decentralized and multiheaded structure of the truckers’ workplace, earning the name the flying picket.
And combined with a general strike they helped spark—one that effectively ran the city for several days—their approach proved decisive. As a labor journalist noted at the time, “The winning of this strike marks the greatest victory in the annals of the local trade union movement . . . it has changed Minneapolis from being known as a scab’s paradise to being a city of hope for those who toil.”
In the 1990s, flight attendants imagined a new form of militancy that had stunning results. After decades of strike actions, flight attendants recognized that a traditional all-out strike, even if dramatic, gave airlines time to absorb the blow of a temporary labor stoppage, assemble scabs, and then return to business as usual once workers could no longer stand to go without their wages. For these workers to win, striking would need to look quite different. Instead of getting every worker to abandon their stations, often handing their duties directly to an army of scabs, flight attendants decided to refuse—in a seemingly random way—to crew their flight at the last minute, instantly grounding it. This unpredictability actively drove customers toward competitors at a level that was far more costly than past strikes. They christened this breakthrough CHAOS: “Create Havoc Around Our System.” The CHAOS model was, in effect, a departure from the dominant playbook of today: Rather than trying to shut everything down at once, it leveraged specific choke points of airline labor to create cascading uncertainty across the system.
Recent campaigns in higher education are also worth examining, especially since the nature of graduate work mirrors in many ways that of tech work, in particular on the dimension of its heterogeneity. Here, too, the canonical strike has proven ill-suited to the fragmented and uneven labor processes that structure the modern university. A lab researcher running an experiment that unfolds over months, with results reported only once or twice a year, causes the university little disruption by walking out for a few days. Likewise, an instructor whose main responsibility is grading at the semester’s end exerts no pressure on the university if they strike during a non-exam period. In such a workplace, a one-size-fits-all strike strategy assumes a uniformity of leverage that simply does not exist. The result is a strike that is loud but not biting—visible, perhaps, on the picket line but invisible where it matters most to the university’s daily operations.
Responding to these limitations, some graduate workers have adopted what they call the “long haul” approach. They recognize that their power lies not in acting as a homogeneous workforce but in exploiting the distinct choke points within their own departments and labs. An effective strike, then, has to begin from the bottom up, with graduate workers in each department and each lab mapping where their labor matters most to the university’s operations and tailoring tactics to those specific vulnerabilities. What emerged from this approach was not a single, uniform strike but a patchwork of sustained localized interventions—grade withholding in some cases, research refusals in others—that together inflicted a much more effective type of disruption to the university.
Taken together, these examples of worker-led experimentation should remind us that workers can and must innovate for ever-shifting organizing contexts. Tech workers must now rise to the occasion.
…
Although a new strategic blueprint has yet to emerge, what’s already clear is that tech workers are ready to rebuild power. Frustrated by the dual threats of eroding working conditions and open lunges toward authoritarianism from tech’s billionaire class, there is a groundswell of agitation from below. Tech workers across the industry are testing new forms of leverage both within and beyond labor’s traditions.
When Hamas attacked Israel on October 7, 2023, many tech workers realized that they would soon be building the technical infrastructure Israel would use to unleash fresh hell in Gaza. Born out of outrage, despair, and desperation, some Microsoft workers came together to form No Azure for Apartheid, calling on their colleagues to refuse labor to any technologies powering the genocide of Palestinians. Over a five-hundred-day campaign of constantly testing new strategies and honing tactics for building leverage, the group learned how to create a crisis for the company. Barricading themselves in the office of an executive while livestreaming the action, disrupting Microsoft’s annual fiftieth-anniversary celebration with screams for Palestinian liberation, and even protesting in kayaks in front of executives’ waterfront homes, this small group of workers did something many thought was impossible: it pushed the company to cut off services to a unit in the Israeli army that used Microsoft technologies to surveil Palestinians in Gaza and the West Bank.
Meanwhile, another experiment was underway at Google. After mass layoffs in 2023, workers started to fight for job security. Starting in Pittsburgh, a handful of organizers who were part of the Alphabet Workers Union organized at their local campus to demand layoff protections from management. Later, as they spoke with others across the company, they discovered there was an appetite to fight the new climate of precarity, so they worked with the larger support network of the union to craft a nationwide campaign. After lead organizers at different offices built collective pressure in a series of open letters, the union won voluntary exit packages for sixty thousand employees. So, while there remains no clear path to winning union recognition and a contract at Google, these little fires that workers created for management across the company’s various campuses have resulted in one of the most militant campaigns at the company since the union’s founding. It has also allowed the union to grow and deepen participation throughout the workforce.
These examples illustrate a clear orientation toward winning—and winning now. The strategies used step outside of the strictures of labor law and represent a trend toward direct action emerging across the movement. Importantly, they do not shy away from openly engaging in conflict, as each clash with the boss is seen as a necessary step toward building solidarity, inciting class consciousness, and learning how to win larger and larger concessions.
Still, what is missing from these actions is a deeper understanding of tech work itself—the connection between how workers’ labor relates to the process of profit generation within their companies and how that can be leveraged to build a credible threat. Without a clear articulation of this relationship, tech workers will not be able to design effective organizing strategies that harness the pinnacle of working class power: the ability to stop generating value for the boss.
Our telling of the movement’s story is meant as an invitation—to those in the tech sector and beyond—to participate in that reckoning.
Okay! That’s it for today. More soon, including a podcast with the workers at the complete opposite end of the story being spun by AI executives. Until then—hammers up.
I found myself fielding a lot of questions from folks who don’t usually talk much about AI this week. They were mostly along the lines of “do you think AI is going to kill us all?” or “how worried should I be about this human extinction stuff?”
Motivating this wave of existential anxiety was the highly dramatic resignation of now-former Anthropic employee Jacob Coxon, who quit his job while publicly sharing his concerns about safety practices at the leading AI companies. More specifically, he declared that almost everyone at OpenAI and Anthropic, the leading AI labs, believes that there’s a decent chance the AI systems they’re producing could “kill us all” by the end of the decade, and the companies aren’t “acting responsibly.”
Two more AI employees quit and joined the chorus. Now, the warning shot has been covered everywhere from the Wall Street Journal to Fox News to WIRED to the Guardian.
I understand why all this sounds scary, but my response to those who have asked, has generally been: No, you don’t need to worry that AIwill kill go rogueand kill us all. But that does not mean we shouldn’t worry, period. Far from it. The companies building AI products are now among the wealthiest and most powerful entities in the world. And they very much are building enormous engines of mass surveillance, extraction, and automation. They can and have allowed their AI systems to carry out felony hacks of other companies, provided the government with autonomously generated kill lists, and sold enterprise technology to firms around the world that is explicitly intended to replace jobs. But “AI” has done these things not because they have all-powerful minds of their own but, ultimately, because people have programmed, calibrated, and directed them to.
In other words, it’s not AI we need to worry the most about. It’s the people and corporations working to profit from and gain power by deploying it. The technology is sophisticated, sure, and, especially in cybersecurity contexts, potentially disruptive to plenty of existing systems. But if we want to stop this predatory, profiteering, and extractive form of AI—and it’s worth noting that at this point, most people do—we absolutely can.
A quick note/plea: 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 analysis, please consider becoming a paid subscriber. I want to keep as much of this un-paywalled as absolutely possible, and I need your support to do so. Many thanks to all who continue to chip in, it’s very much appreciated. Onward.
First, some context: AI industry folks like Coxon have been discussing the prospect that AI will destroy humanity for years. Even before the beginning of the AI boom, many working in the industry, from the executive staff to the rank and file researchers, have obsessed over ‘x-risk’, or the possibility that AI will become sentient or semi-sentient and move one way or another to exterminate humanity.
This is ostensibly why Elon Musk bankrolled OpenAI in the first place, and is the grounds for his lawsuit alleging OpenAI betrayed its charter by abandoning its nonprofit status to become a commercial enterprise. Many AI engineers and startup founders have calculated their own personal p(doom), or the probability that the tech product they are working on will initiate an actual humanity-extinguishing apocalypse. 10-25% p(doom)s are pretty common. (I wrote about a long weekend I spent among these folks at one of their top conferences here, to give you an idea of their ideas and belief systems.)
Sam Altman himself drove the narrative of the imminent rise of a super-powerful AI, or artificial general intelligence, to prominence by declaring that the technology he was building posed an existential risk to humanity.
This idea is in other words baked into the DNA of modern AI companies and into the products they engineer. It’s also why the industry is routinely accused, by myself and others, of “doom marketing,” ie, using such proclamations and framings to impart the sheer power of the technology behind your product onto investors, partners, consumers, and the state. Some ambient understanding of this tendency may have helped prevent the press from repeating the claim every time one of the x-risk folks made it over the years, including when they quit their jobs at AI companies while doing so.
Yet Coxon’s AI x-risk alarm has broken containment, as they say, and has exposed the idea to the mainstream media once again. This has happened, I think, for a number of reasons. First, when AI CEOs were going around saying AI might kill everyone in 2023, their companies were well-capitalized, but not nearly to the extent they are today. They were more easily dismissed as eccentric geeks of the sort that Silicon Valley is supposed to produce. Now, not only do their products have much greater market saturation, but people are more aware of a) the often unethical ways the executives and companies comport themselves b) the other adverse impacts of AI that have already arrived, and c) the physical cost of building these systems, particularly when it comes to data centers. Finally, toss in the constant stream of headlines in 2026 about AI models going rogue or potentially upending the world’s infrastructure.
As a result, general sentiment was already overwhelmingly skeptical of AI and its overlords, so the table was pretty well set for someone to come along and say ‘just a reminder, the companies that are building the product that promises to take your job and leave a data center in your neighborhood also think that it might exterminate the human race’ and get a strong reaction from the media and the public.
The question is: Should we understand this stuff as hype, strategically fanned to feed already swelling valuations? Is it doom marketing? Genuine cause for alarm that the infrastructure that undergirds modern life is facing a novel threat vector? Or reason to be sincerely worried that AI companies are amassing too much power, and that their products are genuinely dangerous?
The answer, I’m afraid, is “yes.”
Let me try to untangle all of this a bit, because arguments over how to navigate this moment are extremely heated right now, as they are essentially arguments over how to approach the politics of AI at a moment where there may be a unique opening for action. Those sympathetic to AI alignment concerns and x risk arguments—which increasingly includes a faction on the left who want to shut down AI over safety concerns—argue that people who dismiss AI as all hype have their heads in the sand. Meanwhile, critics of the x-risk narrative—which I might note includes a cadre of investors and tech folks on the right, too—argue that it is hype and can be used to further inflate valuations or for pursuing regulatory capture.
The challenge is that many feel that dismissing AI as just hype or purely marketing also diminishes the urgency of taking on the AI industry as the novel and threat to labor, civil society, and the public sphere it is; others that by accepting the industry’s own terms, we imbue it with undue power and enable it.
But let’s be clear: Declarations that AI is so powerful it might end the world are a kind of marketing, and anyone who says otherwise is just as off base as those who say that’s all that it is. If you don’t think that these sub-trillion dollar companies have higher-ups who noticed that when their founders talked about super-powerful dangerous AI they got spikes in media interest and phone calls from investors, or that those founders didn’t notice that themselves, and lean into certain narratives to advantage their products, then I don’t know what to tell you. Not all marketing is a Pepsi ad campaign with a celebrity spokesperson.
Yet confronting the AI industry should certainly be considered an urgent project. Do I share the x-riskers’ fear that out-of-control, super-intelligent AI agents might end the world? I do not. I’ve read a good deal of AI extinction literature over the last couple of years and I have not come across a credible, step-by-step documentation of how exactly AI might move from self-recursively improving AI to killing every single human on the planet. Most of what I’ve read seems predicated on a sort of reverse Ian Malcolmian logic: Eventually, it becomes smart enough and AI, uh, finds a way (to exterminate humanity). Are the AI companies also overstating and dramatizing what their products can do? Yes, pretty much all the time. They’re in a knife fight for market share and are trying to win customers and government contracts before their investors come knocking.
And yet. They have concentrated enormous amounts of compute, have extensive data collection and surveillance capacities, have inserted themselves into the state in numerous capacities, including military kill chains, and they are being run by egomaniacal billionaires who are rushing to monopolize an entire new market. The way I think about their most recent frightening-seeming exploits is not as a “swarm of rogue AI agents building a new civilization” or whatever, but as corporations with a hitherto-unimaginable amount of compute at their disposal recklessly experimenting with automation programs, completely unafraid of any repercussions because we have been unwilling or unable to impose any on them. So I think we should welcome any newfound imperative to hold them accountable.
I also think that ultimately, the AI critics, Pause AI folks, leftists, doomers, effective altruists, and, honestly, most of the public, want a lot of the same things: legitimate checks and limits on AI companies’ power, actual transparency into what the companies are doing, and for more of AI to be under public control. (Ultimately, a lot of what animates the x-risk fears are rooted in just that: a fear of losing control.)
How we move towards all this, exactly, is the question of the day, and while there’s much to be argued over in terms of building a viable political platform to challenge the AI industry’s power, I do want to cast a vote against one particular approach.
Not to pick on Chris here, as there are a lot of folks out there proposing stuff like this, and the policy ideas actually stem from AI safety groups. But not only would this be challenging to land politically (I can already hear the ‘China will beat us to AGI’ calls) and all but unworkable in practice, this approach is a) far too limiting, b) fails to address the current harms caused by AI and c) would likely only wind up serving Anthropic and OpenAI; it’s what regulatory capture looks like in action.
Full model transparency is the floor for what should be on offer here. If these companies believe their products will kill us, the very least we deserve is to know how. It’s in the public interest. I would also argue we might get a lot of the way to stopping any currently envisioned robot apocalypse simply by forcing AI companies to assume liability, and to submit to the full brunt of the law. If Sam Altman were liable to be charged with a felony for approving the hack on Hugging Face1, if Google was forced to pay defamation fines each time it published a false and injurious AI Overviews, these companies might suddenly find cause to restrain their models’ output, and we might hear less about rogue AI. We would need to introduce new frameworks and available tools for tracking automated lawbreaking, to be sure, but I’d wager that little would frighten the AI companies more than actually being held accountable for what their products enable. To that end, Illinois has passed an AI liability law, which the industry fought tooth and nail, and Florida is considering one.
The next steps would be towards moving AI under public control; again, it’s only reasonable, since everyone, not just AI insiders, should get a say over how a technology whose developers say might kill us is developed and deployed. Dismantling the whole industrial project as currently structured should be on the table, too, pending democratic input. But that’s for another post.
For now, suffice to say that while I won’t be losing any sleep about an AI-enabled extinction event, Mr. Coxon and co. have given us an opportunity to underscore the urgency of the clear and present damage AI companies are doing, the extreme power they wield and the extent to which they hold themselves unaccountable, and the abundance of political capital there is out there for taking them on.
An Iowa man has kicked off controversy over religious freedom in the small town of Ottumwa, Iowa by opening a city council meeting with a prayer to Satan, which then kicked off a debate about the proper time in the meeting agenda to undo the “curse.”
On Tuesday night, Ottumwa resident, satanist, and former city council member Matt Dalbey took the podium to read an invocation to start a routine meeting of the council. “Thank you mayor, council. Let us pray,” Dalbey started. “Dear dark father, let us stand now unbowed and unfettered by arcane doctrines born of fearful minds and darkened times. Let us embrace the Luciferan impulse to eat off the tree of knowledge and dissipate our blissful and comforting delusions of old.”
The prayer is a common one used by the Satanic Temple of Iowa. “Let us demand that individuals be judged for their concrete actions, not for their fealty to arbitrary social norms and illusory categorizations,” he continued. “Let us reason our solutions with agnosticism in all things, holding fast only to that which is demonstrably true. Let us stand firm against any and all arbitrary authority that threatens the personal sovereignty of one or all. That which will not bend must break. And that which can be destroyed by truth should never be spared its demise.”“It is done. Hail Satan,” he finished.
Another citizen, a pastor named Rick Bick, rushed to the podium. “Let’s just reverse that curse that was just put on our city,” he said. “That ending with hail Satan and hail dark god. That’s not welcome here.”
“Mayor, that should be done at the end,” city councilor Cara Galloway said.
“It needs to be done now, OK.” Rick said. “I appreciate what Ms. Galloway is saying but that was a curse put on our city when he said ‘Hail Satan’ and the dark god and most of you know that. And we need to just stop and reverse it.”
“So what I’m gonna do now…I’m just gonna invite us into a 30 second time of prayer — time of silence.”
“No,” Galloway said. “We are at the consent agenda. I’m sorry, but it’s time to move forward. The invocation happened.”
“As mayor of this city…I’m going to ask right now that we do a moment of silence to reflect on whatever it is we need to reflect on at this time. And then we will move to the consent agenda,” mayor Benjamin Foote said.
Galloway spent the moment of silence with her head up and her arms crossed. An hour later, she spoke up. “Opening with an invocation means allowing viewpoints that may not reflect our own and responding to someone’s invocation is a slippery slope to putting a financial liability on the taxpayers and allowing others to object or respond. This is exactly why I had concerns about having an invocation when we first talked about it.”
Dalbey told 404 Media he was a sincere member of the Temple of Satan but stressed it’s “an organization of nontheists, people who don’t believe in any deities or supernatural beings.”
He served on the city council for eight years and has largely stayed out of local politics since leaving the council in 2021. But when the city council passed a new rule earlier this year that allowed city council meetings to start with religious invocations, he decided to do something. “I’m a constitutionalist,” he said. “And when I see this, the implication I got was that we were again starting to mix religion into the government.”
At the Tuesday night council meeting, the audience stood up to recite the Pledge of Allegiance the moment Dalbey finished.
Ottumwa has only been reading invocations at its meeting for a short time. The city council passed a resolution allowing the practice on January 20. Galloway was the only person on the council who voted against it. At the time, she said Satanists might show up and read a prayer and that everyone would have to be OK with that. Readers of 404 Media may remember Ottumwa, Iowa because it is the place we aired our Super Bowl commercial.
Anyone who wants to show up to read an invocation before a city council meeting needs to fill out an electronic form. The form asks applicants to list an organization they represent, but allows people to say “none.” Dalbey has signed up to read invocations three times this year. He’s listed as representing “none” each time on the schedule. He was supposed to read a prayer on September 1, but couldn’t make the meeting because of his work. He’s on the schedule to read an invocation on November 3rd. “I plan on being there in November, absolutely,” he said.
Dalbey said he’d be happy to see the invocation policy rescinded. “When I was on the council [...] we were approached several times about doing an invocation, and the very reason that we didn't was because our attorney had warned us that when you do, you open the door to any and all people who want to present. That's the law,” he said.
The invocation to Satan in a small town in Iowa is just the latest salvo in a 10-year-long battle over religious freedom of expression in the state. In 2016, a religious liberty group installed a Christian nativity scene at the state capitol in Des Moines. Later that year a group of atheists installed its own nativity scene honoring the Bill of Rights. In 2023, The Satanic Temple put up a pentagram built from red ribbon and topped with a mirror-mask ram’s head that invoked Baphomet in the capitol rotunda. The group also began holding winter celebrations in the Rotunda. A man vandalized the display, police arrested him and charged him with a hate crime.
In 2024 — the same year Governor Kim Reynolds signed a state law meant to protect religious freedom — officials blocked Satanists from holding their winter celebration. The ACLU filed a Freedom of Information Act request looking for records about the incident and, earlier this year, sued Iowa on behalf of the Satanic Temple for violating federal free speech laws.
“I think we're seeing you know a big shift in our politics. I think we've seen a rise of Christian nationalism in this country [...] and I think it's trying to grasp and get a hold of controlling our politics, our laws and our government, and I find that highly troubling,” Dalbey said. “And I think anytime we let any religion dominate our politics, we revert right back to where we were 250 years plus ago: a monarchy where the church ruled the country, and that was bad for everybody. That's one of the reasons our founding fathers decided to form a new nation. And I think if we are not careful, you know, we're going to slide right back into that scenario.”
Is there any logo more iconic than the Nike swoosh? Unlikely. In fact, it’s arguably the most recognizable symbol in sports. But how did a small running company become a global powerhouse? Vitra Design Museum set out to find out, gaining unprecedented access to the Department of Nike Archives (DNA) to curate Nike: Form Follows Motion. First presented in Germany in 2024, the exhibition is now on view at the Museum of Arts and Design in New York through March 2027.
Photography by Bernhard Strauss; courtesy of Vitra Design Museum
Vitra Design Museum had been considering an exhibition about sports and design for years. When it approached Nike, the museum learned the extent of the company’s private archive, opening the door to a collaboration.
Photography by Bernhard Strauss; courtesy of Vitra Design Museum
“I was able to draw on the extraordinary depth and breadth of DNA’s holdings, including objects, documents, prototypes, study models, and material research, as well as some current designs from the company,” exhibition curator Glenn Adamson says.
Photography by Bernhard Strauss; courtesy of Vitra Design Museum
And what a time to bring Nike: Form Follows Motion to New York City, in a year when the Knicks won their first NBA championship in 53 years and the 2026 World Cup final was played just across the river.
Objects from Frank Rudy, who invented the Air Technology, from the Department of Nike Archives (DNA). Photography by Alastair Philip Wiper; courtesy of Nike.
“Having the exhibition travel right to the heart of New York couldn’t feel more appropriate,” Adamson says. “Nike has a very special relationship to New York City, going back at least to the early 1980s … when hip hop broke, coming out of Queens and Harlem. And the hip-hop community was very interested in streetwear generally, brands like Nike especially.”
A graffiti installation by Faust on the MAD Store doors. Photography courtesy of the artist.
Further connecting the exhibition to New York’s counterculture, renowned graffiti artist Faust, whose work is quintessentially NYC, created the cover of the book accompanying the installation and added his signature graffiti to the museum store door.
An early mechanical Shox prototype, c.1981. Photography from Department of Nike Archives; courtesy of Museum of Art and Design.
So what exactly is being exhibited at Nike: Form Follows Motion? Beyond the finished products most visitors will recognize, the exhibition focuses heavily on what usually remains unseen: prototypes, experiments, sketches, and design studies.
Familiar products like the Waffle Trainer, Air Force 1, Air Jordan, Air Max, Flyknit, and Vaporfly are displayed next to rare one-offs.
Various lasts, jigs, silicon pads, and fixtures from Nike’s Advanced Product Creation Center. Photography by Alastair Philip Wiper; courtesy of Nike.
The exhibition is presented in four sections across two floors of MAD. “Track” explains Nike’s origins, founded in the ’60s by Phil Knight and his former track coach Bill Bowerman.
“Air” transitions to the ’80s, when Nike’s pressurized Air cushioning became a major focus of the company’s design development. The technology was made visible for the first time in the 1987 Air Max, and sketches, notes, and prototypes reveal how it evolved.
Photography by Bernhard Strauss; courtesy of Vitra Design Museum Photo
“Sensation” explores the Nike Sport Research Lab, which studies the body in motion, evaluates the science behind the shoe, and examines Nike’s research into circular manufacturing, material sourcing, and recycling.
Finally, “Relation” is the exhibition’s grand finale, showcasing 50 transformative and often avant-garde footwear styles alongside hip-hop music videos, explicitly illustrating the correlation between Nike and streetwear culture.
A sketch of Air Max by Tinker Hatfield, 1986. Image courtesy of Nike.
“What’s really special … is that we have so many rarities from the archives,” Adamson explains. “That would include very early relics that are now unique objects [and] a lot of prototypes that show how these designs—which are very technically involved—were actually invented and developed.”
Thanks to the wealth of material from the DNA, including more than 200,000 objects that had never before been presented to the public as a curated collection, Adamson says the exhibition offers a glimpse of Nike’s “design process on the fly.”
Photography by Bernhard Strauss; courtesy of Vitra Design Museum
Nike: Form Follows Motion is on view at the Museum of Arts & Design in New York until Mar. 7, 2027.
OpenAI founder and CEO Sam Altman was on Politico’s Decoded podcast yesterday, deflecting on a variety of topics. But there’s one amazing quote, which has been making a splash: [YouTube]
I think one of the biggest differences between us and some of the stricter, let’s say, AI safety people, is we believe that the world should accept some bad things happening for the benefits of this technology and people having the agency.
What bad things, Sam? Who are the bad things happening to, Sam?
If you were feeling generous, you could read that just as all tech having costs and benefits. Sure.
But there’s context here — such as OpenAI’s recent hacking run and how the AI vendors are now telling the general public their chatbot might kill everyone, maybe. But we need to not shut them all down, for some reason.
Altman’s been on the AI killing everyone kick for over a decade now. Here’s Altman from 2015, at Airbnb’s Open Air conference, back when he was still a venture capitalist at Y Combinator and just getting OpenAI started: [YouTube]
I think AI will probably, like, most likely, sort of lead to the end of the world, but in the meantime, uh, there will be great companies created with serious machine learning.
Yeah, that’s great, Sam.
For you and me, Altman’s great tech increasingly just has costs. OpenAI keeps making our lives a misery. It rents huge obnoxious data centres. It drives up the cost of anything electronic. The AI companies sell their chatbot as getting us all fired and making our lives even worse. And if you have a job, you get to use Copilot or Claude Code!
Then they’re surprised when people don’t want to take them in good faith.
Altman also pursues this theme of how your life has to suck a bit now so it can suck more later in his new Vanity Fair interview, just out. [Vanity Fair, archive]
Altman goes through a pile of disasters his chatbot could bless us with. AI hacks all the computers! AI makes a virulent new disease! AI could make humans extinct!
But that’s okay, see:
We are willing to accept some things going wrong to make sure that this is a widely held and used and democratic technology.
Altman does not care about democracy. He’s not looking for widespread empowerment. When he says “democracy,” he means he wants permission — he wants enablers.
Log in
