Former Twitter co-founder Jack Dorsey fired 4,000 people yesterday — 40% of Block’s workforce — citing a supposed AI revolution that is just around the corner. “we’re not making this decision because we’re in trouble. our business is strong,” wrote the CEO in an all-lowercase letter telling people he was taking away their livelihoods. “gross profit continues to grow, we continue to serve more and more customers, and profitability is improving. but something has changed.”
That something, I’m sure you’ll be shocked to hear from this longtime denizen of Silicon Valley, is AI. “we’re already seeing that the intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company. and that’s accelerating rapidly,” explained Dorsey.
Block is a publicly traded company (they operate Cash App as well as Square, its small business payment processing hardware and software suite), and when they announced this radical change to their business, the market rewarded them handsomely for it. Its stock was up over 20% at one point and it opened today about 13% higher than yesterday’s close. This is a pretty significant message from a market currently riddled with AI jitters and selling off in the face of good Nvidia earnings, and it’s not exactly a positive one for labor.
But are Jack Dorsey and the markets right? Or just mad? Or a little bit of both? He has spanned the full spectrum in his now two decades as one of the main public ambassadors from the Valley, so the answer is inevitably a little bit of both. The primary problem with discussing AI is that the public faces of the industry are a bunch of charlatans who will say any kind of nonsense in order to try to add millions or billions or even trillions to their company’s value (like asking for $7 trillion dollars with a straight face), all in the context of the larger arms race to see which company that can’t justify its valuation gets to cash out on the open markets first (and it looks like Elon Musk’s unholy accounting merger of SpaceX and xAI is in the lead). It is more than reasonable to want to respond to everything Sam Altman and his kind say with OHMYGOD SHUTTHEFUCKUP and reflexively oppose every syllable out of his mouth, but unfortunately, idiots like him are standing on the work of a machine learning industry that existed long before he was born. And innovation in that industry continues apace.
According to Lee Chong Ming in Business Insider, a former data analyst at Block who was part of that 40% let go over AI, he “could see in my own work very quickly how much of it was already being automated. So much of the data analyst world is finding the right dataset, writing something that will allow you to pull the data set that you want, and then generating output. Every single one of those steps is significantly faster and easier because of AI.” I can echo this in my own usage of Claude, as I have used it to help me build and analyze spreadsheets and run regressions for the purposes of assessing and managing my investment portfolio. This stuff is really powerful in some specific applications, a lot surrounding coding and data analysis and processing. It’s even capable of hacking nearly the entire Mexican government.
I also copy/pasted the draft of this article into Claude and asked it to suggest titles for it. I am not concerned whatsoever that it is coming for my job any time soon. LLMs are both impossibly stupid and shockingly smart at the same time. It really is a mindfuck to use a computer that’s really good at doing some computer things, then all of a sudden it royally fucks up something basic that computers could do in 1985, and you give back all that productivity you gained by trying to figure out why the computer won’t computer.
AI is innovating work and affecting the job market, just not all work the way these executive goobers promise it is doing. Last year was a very bad year for hiring in America, and while AI was not a main driver of the stagnation like the AI cultists freaking out over bad blogs that don’t understand Uber and DoorDash’s moats would want you to believe, it did play a part.
I detailed back in November how you can see AI-driven job loss in some entry level job data. Wired highlighted a Stanford study in August which revealed how jobs like customer service and software development have seen a 16% decline in employment for workers aged 22 to 25 in affected industries. Goldman Sachs’ chief economist Jan Hatzius wrote a note about this last August, affirming that “It is true that AI is starting to show up more clearly in the data.” The anti-AI brigade can claim this tech is totally useless all they want, but we have plenty of data demonstrating how that kind of rigid across the board thinking is more reflective of a black and white Sam Altman-style worldview than what the nuances of the data reveal.
Consulting firm Challenger, Gray & Christmas found that AI “was responsible for almost 55,000 layoffs in the U.S. in 2025,” while an MIT study published in November suggested that AI can already replace 11.7% of the labor market. This is the kind of number that is infuriating to both sides of the AI extremes, where it is far too small to say we are anywhere near the Silicon Valley executive class’ wet dream of replacing every American worker with a chatbot that hallucinates bullshit, but it’s also way too big to give credence to the “this is just NFTs and shitcoins all over again” extreme. Amazon, Google and Microsoft did not bet hundreds of billions of dollars in long-term capex on slurp juice, folks. Tell new college grads struggling to find work who are seeing chatbots pop up across the customer service landscape that AI is 100% hype. This is having an impact in growing pockets of the economy, and there is more than enough data to say so conclusively.
But whether a drastic move like firing 40% of your workforce so you can throw it all away for AI agents is justifiable is also very much in doubt. The consulting firm Forrester found that 55% of employers regret laying off staff because of AI. “Too often, company management lays off employees based on the future promise of AI,” they wrote summarizing their analysis. They did say that HR functions are at great risk of being replaced by AI, but the notion that higher-level jobs in the org chart are presently threatened is questionable at best. “We predict that much of this work will be placed on low-paid workers, either offshore or at lower wages,” Forrester said.
Harvard Business Review surveyed 1,006 global executives in December 2025 about AI, and what they found “is that AI is behind at least some layoffs, but that these are almost completely in anticipation of AI’s impact.” AI is a genuine technological advancement, but it is being weaponized by the executive class and sold as a so-called revolution in the hopes that they can finish the job the last generation of capital started in reaction to the New Deal, and tear down a century of labor progress. Silicon Valley’s executive titans are very explicit about their goals here: They want to put you out of work forever, all while they support Donald Trump’s fascist project. The mask is all the way off now.
And most of the AI spending is coming from companies who back the current fascist project in America. Meta is building a SuperPAC to flood political races with money to support pro-AI candidates, all while they build a chatbot that will flirt with your children but will not tell you where to get an abortion. Jeff Bezos is lighting a miniscule part of his net worth on fire to destroy a journalistic institution in patronage to dear leader. Google and Microsoft gave a million dollars to Trump’s inauguration slush fund. It would be one thing if we lived in a normal time in America where the interests of our largest and most powerful companies were based in some kind of shared vision of society, but Silicon Valley’s elite lobbying to overturn the 10th Amendment and convince Trump to impose a nationwide ban on any kind of AI regulation shows how they view themselves as a bunch of feudal lords who see this country as a land where they can exploit 330 million digital serfs. America cannot be a democratic country so long as this iteration of Silicon Valley exists.
This convergence of unhinged Silicon Valley C-suite fascism from a bunch of raging hypocrites who are reportedly high on drugs all day, and the useful tools their far more talented employees built is where AI becomes most difficult to discuss. It’s also where the rabid Bluesky “everything AI is useless, fascist, and just like crypto” worldview runs into a crossroads. Is AI useful or not? If not, why are you worried about it being used for fascism?
But it is being used for fascism beyond the truly useless image-generating AI being the aesthetic of it. Pete Hegseth is currently doing his usual “incoherent” bark but no bite routine, trying to bend Anthropic to the Department of Defense’s will, and Anthropic so far is resisting this edict to let a drunken child use Claude to spy on all of us. Sam Altman is also trying to get good press in a fight not featuring him, saying that OpenAI is on Anthropic’s side of supposedly not letting the fascists in the front door.
But as Edward Snowden taught us, Silicon Valley is more than happy to install a secret illegal backdoor for the surveillance state, so trusting tech companies to respect our Constitution is not something I would suggest people do. It’s on them to tell us why this time is supposedly different with their self-replicating dragnet that can run autonomously and improve upon itself.
People’s fears of AI replacing our jobs or being used to further our fascist moment are real and justified. It is surrounded by a hurricane of crypto-style moonboi bullshit, but unlike crypto, there is real meat on the bone once you get past the hypebeasts. AI has created a construction boom where data center construction spending is now roughly equal to all US office construction spending. There is even growing backlash to this expansion due to the increased energy costs and massive environmental harms it brings with it, and it seems we are at a moment where Act I of the AI arc has completed. We are entering a new unknown period of a more honest moment around what AI can and more importantly, cannot do, with stock market shivers looming over everything as $64 billion in data center projects have been blocked or delayed by local opposition.
The ChatGPT so called revolution is over, and people’s skepticism over chatbots that encourage teenagers to commit suicide is growing, and they can see the impact that this orgy of data center construction is having on their bills and communities. People don’t like a partially broken product being shoved down their throats by a management class eager to replace them with it, especially when the AI productivity numbers don’t justify the hysteria whatsoever. That the Silicon Valley executive class and their lemmings are sniffing their own farts to the degree that they read one bad blog and became convinced that we’re imminently about to be replaced by a teenager suicide machine just shows you what a bubble they live in.
The most charitable data from the best sources is that somewhere in the neighborhood of 10-20% of specific and low-wage work is at immediate risk of being replaced by AI, which honestly tracks with past technological advancements automating away rote tasks. But the idea that 40% of a workforce can be replaced with AI agents is still to be seen, and I suspect that like all present advancements in AI, will contain specific use cases where it does work and many where it doesn’t (as someone who spent time in the payment processing industry, I would bet on a company like Block being able to pull something like this off more than like, Microsoft). I would also bet that this next stage will see a lot of companies fail or falter thanks to their executives literally buying the hype that capital is about to finally win the centuries long war they have waged against labor. So far, the data suggests that they usually come to regret that belief.
