The Brightest Gaslight
These notes connect directly to what I wrote last night about Power as understanding what’s happening with AI and jobs isn’t just an economic story. It’s a power story.
While tech CEOs warn about mass unemployment in one breath and promise productivity paradise in the next, the Trump administration has formalized a policy narrative that makes everything clearer. American AI leadership isn’t being framed as an innovation story or a prosperity story. It’s being framed explicitly as a matter of national security and economic dominance.
The language is unambiguous. “Win the AI race.” The “Genesis Mission” launched via executive orders to accelerate AI development at any cost. Deregulation designed to fast track power intensive AI data centers. Removing regulatory barriers that might slow deployment. The stated goal is to ensure the U.S. remains the leading economic and military power.
Notice what’s not in that framework. Jobs. Workers. The middle class. Social stability. Those are externalities, not objectives.
When you understand that power, not prosperity, is the actual priority, everything else starts making sense. The contradiction between public reassurance and private warnings disappears. The gap between “AI will create new jobs” and “humans won’t be needed for most things” resolves itself. They’re not confused. They’re not uncertain. They know exactly what they’re building and who it’s for.
What follows is about AI and jobs, but it’s really about who gets to make decisions that affect hundreds of millions of lives, and whether the rest of us will figure out what’s happening before it’s too late.
This morning at Planet Fitness I finished listening to Geoffrey Hinton’s interview with Ian Bremmer on the GZERO World podcast, and I have to tell you, my immediate reaction wasn’t awe or inspiration. It was more like watching someone finally admit to a crime they’ve been covering up for years.
For the past decade, we’ve watched Silicon Valley and Washington feed us a very specific, very comforting story. You know how it goes. Technology always creates more jobs than it destroys. They point to the printing press, the steam engine, the internet. They tell us AI is just a “tool” to make us more productive, a “copilot” that will free us from drudgery so we can all become creative directors and empathy specialists.
Just recently, former Google CEO Eric Schmidt was out there selling this exact line, claiming the idea of massive job loss is “almost certainly false” and that “for every job loss, there’s probably more than one job created.” The venture capitalists nod along. The think tank experts agree. The labor market will simply “adapt” as it always has, they assure us.
But then you hear Geoffrey Hinton, the man who literally won a Nobel Prize for building this technology, the so called “Godfather of AI,” and he doesn’t give you the PR approved version. He warns of “massive job loss” specifically targeting “mundane intellectual labor,” which, let me be clear, means the exact jobs that built the middle class. Paralegals. Call center workers. Entry level coders. When pressed on whether new jobs would replace the old ones, he didn’t dance around it. He called that assumption “dubious.”
Think about that for a moment. The man who built the foundation of this technology is openly questioning whether the optimistic story we’ve been sold will hold true this time.
His warning forces a critical question into the open. If these new jobs are really coming, where are the early signs?
Let me walk you through what’s actually happening on the ground. The overall U.S. unemployment rate remains relatively stable at around 4 to 4.5 percent, so we’re not looking at an economy wide collapse. But that aggregate number masks some deeply troubling patterns emerging in specific sectors and demographics.
In 2025 alone, U.S. employers announced over 1.1 million layoffs, the highest since the pandemic years. Technology and government cuts have led the way. Now here’s the part that should make you pay attention. For the first time, companies are explicitly naming AI and automation as major reasons behind tens of thousands of job cuts. This isn’t speculation about the future. This is happening right now in quarterly earnings calls and restructuring announcements.
Intel, Microsoft, Amazon, and Meta have all executed significant workforce reductions. The pattern is consistent. Companies announce they’re “investing heavily in AI infrastructure” in the same breath as they announce eliminating thousands of positions. The justification has shifted from post pandemic normalization to something more fundamental. They’re not just cutting inefficiency. They’re replacing human cognitive labor with systems that work 24/7 and don’t require health insurance.
But here’s where the story gets more complex than simple displacement. These layoffs aren’t happening in a vacuum. They’re driven by multiple factors working together. Rising interest rates. Cost pressures. Tariffs and trade tensions. Government fiscal tightening. AI is one force among several, but it’s increasingly the one that companies cite when they want to explain why certain categories of work are disappearing permanently rather than just being delayed.
The most revealing evidence isn’t in the aggregate numbers. It’s in what’s happening to the people who should be thriving in this supposed boom of new opportunities. Look at recent college graduates.
Recent data shows the unemployment rate for college grads aged 20 to 29 hit around 7.1 percent in late 2024, nearly double the overall unemployment rate. Compare that to 2019, when recent grads had an unemployment rate of just 3.25 percent. The advantage of having a degree, the edge that used to open doors, has eroded significantly.
And it’s not just about having a job. It’s about having a career that uses your education. Over 52 percent of recent grads are now “underemployed,” working jobs that don’t require a degree just to make ends meet. These are people who did everything right, who followed the rules, who invested in their education. The entry level positions that used to be their launching pad are simply not materializing at the same rate.
Now, the situation varies considerably by field and timing. Some majors and some cohorts are doing better than others. But the overall trend shows young workers bearing a disproportionate burden of whatever transition is underway. When the economy “adapts,” it’s not adapting equally for everyone.
The most striking part of this whole situation is what happens when the people building and deploying this technology speak candidly about what they see coming. The contrast between the public narrative and these warnings is remarkable.
Dario Amodei, CEO of Anthropic, is in the business of selling AI tools that enterprises are rapidly adopting. Yet he’s warned that AI could eliminate up to half of entry level white collar jobs within roughly one to five years. He’s not predicting a smooth transition where new jobs seamlessly replace old ones. He’s warning that unemployment could spike into the 10 to 20 percent range and explicitly urging policymakers to prepare for that scenario now, not after it happens.
Mustafa Suleyman, CEO of Microsoft AI, has publicly stated that AI will create a “serious number of losers” in the labor market. He’s talking about fundamental changes to white collar work over the next five to ten years and suggesting that policy tools like Universal Basic Income may be needed to cushion the economic shock. His point is stark. AI doesn’t need to be super intelligent to disrupt labor markets. It just needs to be cheaper and more reliable than human workers.
Sam Altman, the face of the AI revolution himself, has acknowledged that AI will “totally take some jobs away.” His fascination with Universal Basic Income isn’t academic curiosity. It’s rooted in his expectation of large scale labor market disruption from AI and automation. He talks about new jobs being created too, but he’s clearly anticipating a transition painful enough to require fundamental changes to the social contract.
And then there’s Bill Gates. For years, Gates was one of the technology optimists, talking about how innovation would lift humanity, create opportunities, solve our biggest problems. But something has shifted in his recent public statements. In March 2025, he told CNBC and repeated on The Tonight Show that within about a decade, AI progress could mean humans “won’t be needed for most things.”
That phrasing is worth sitting with. Not “humans will do different things.” Not “humans will focus on higher value work.” Humans won’t be needed for most things. He’s suggesting that expertise in many professions will no longer require a human at the center. He’s predicted that the work week could shrink to two or three days because machines will handle most economic labor.
Gates still emphasizes potential upsides. Medical breakthroughs. Climate solutions. Dramatic productivity gains. But he couples all of that with explicit warnings about what he calls “mass job disruption” and the need for society to make deliberate choices about how to distribute work, wealth, and leisure in this new economy. His message has become less “technology will automatically create prosperity” and more “we need to actively shape this transition or it could turn dystopian.”
These aren’t Luddites predicting doom. These are the architects and beneficiaries of the AI revolution itself. And they’re not declaring certainties. They’re warning about high probability risks that policymakers and institutions seem unprepared to handle.
What strikes me most is the gap between the confident “more jobs will come” narrative still dominating think tanks and business press, and the conditional, uncertain warnings from people actually building the systems. Hinton, Amodei, Suleyman, Gates, Altman are all emphasizing the same things. Enormous uncertainty about pace and distribution. Non linear risk of social upheaval if institutions can’t adapt fast enough. The possibility that productivity gains and severe displacement could happen simultaneously.
The “tech always creates more jobs than it destroys” narrative isn’t a law of physics. It’s an empirical pattern from previous industrial revolutions, and past performance doesn’t guarantee future results. The steam engine replaced muscle power. AI is replacing cognitive labor, the very thing that made human workers valuable after machines took over physical tasks. That’s a fundamentally different kind of transformation.
I’ve spent years in the technology world, and I can tell you that what we’re witnessing is a disconnect between public messaging designed to maintain confidence and internal conversations that are far more guarded. The people with the deepest technical knowledge aren’t declaring that human labor is obsolete. But they’re also not dismissing the possibility. They’re saying we need to take the risk seriously while we still have time to shape the outcome.
The question isn’t whether to believe the optimists or the pessimists. The question is whether we’re willing to prepare for scenarios that the people building this technology consider plausible enough to warn us about publicly. We’re not in free fall yet. The labor market hasn’t collapsed. But the warning signs are real, the pressures are building, and the window for proactive policy responses may be narrower than the comfortable narratives suggest.
What do we do if the value of routine cognitive labor drops dramatically? What happens to a society built on the assumption that people work for their living when a significant fraction of that work becomes economically unnecessary? These aren’t hypothetical philosophy questions anymore. They’re scenarios that serious people with serious credentials are telling us to prepare for.
The confession is out there. Einstein once said that “unthinking respect for authority is the greatest enemy of truth.” The authorities in this case are the very people whose fortunes depend on mass AI adoption. They’re telling you everything will work out fine while simultaneously warning their investors and policymakers to brace for mass unemployment. They’re selling you productivity tools while cutting headcount. They’re promising a future of creative fulfillment while twentysomethings with college degrees compete for jobs that don’t require one.
You’re watching it happen in real time. The question isn’t whether to believe the optimists or listen to the warnings. The question is whether you’re going to keep trusting the people who profit from your compliance, or whether you’re going to look at what they’re actually doing versus what they’re saying. The window for shaping this transition is closing.
They warned you.
…but as I said in the preface, the atoms of power have made their bet…