The AI Boom Is Eating Its Own Tail

---

Sam Altman got a homemade bomb thrown at his gate. That’s not metaphorical. A 20-year-old Texan, charged with attempted murder and federal felonies, allegedly had a detailed manifesto about why AI executives needed to die. He wasn’t alone in his thinking—he had a list. Multiple names. This isn’t some fringe Reddit post. This is real radicalization happening in real time.

And somehow? This is the least crazy thing happening in tech right now.

Photo by Molnár Tamás Photography™ / Pexels

The Bomb That Reveals Everything

Let’s be clear about what just happened. The person who threw that explosive at Altman’s home isn’t a statistical outlier or a one-off disturbed individual. According to reporting, he’d written extensively about the threat of AI. He had documentation. He had targets. He had conviction. This is what happens when you spend five years telling the world that artificial superintelligence might end humanity, then move ahead at full sprint anyway—and make billions doing it.

I’m not saying Altman deserves sympathy for running into the logical endpoint of his own apocalyptic messaging. I’m saying the violence is a symptom of a much larger breakdown.

When a technology founder needs armed protection from people radicalized by the very problem he’s building faster, you’ve crossed into territory that doesn’t have good historical precedent. The closest analogy might be the early years of genetic engineering in the 1970s, when biologists genuinely worried their own research could create a plague. Except this time, the “plague” is potentially conscious software, and the guy building it is one of the most powerful people on Earth.

Everyone Wants In. Everyone’s Terrified.

Amazon just spent $10.8 billion to buy Globalstar, a satellite communications company. Why? To build out satellite internet and compete with Starlink. That’s not about giving rural America broadband. That’s about infrastructure for the next phase of the AI wars.

Meanwhile, China, the U.S., Russia—they’re all locked in what’s being compared to the nuclear arms race for AI-backed weapons and military systems. Not consumer AI. Military AI. The kind that makes decisions about firing. This is happening right now, and most people don’t even know it’s the actual competition.

Europe’s scrambling to figure out if it can catch up. Palantir’s getting NHS contracts in the UK. Quantum computing talent is consolidating. Every major player is consolidating everything—data, talent, compute, infrastructure—because they know that whoever locks down the supply chain first wins not just a market but possibly human civilization as we know it.

Photo by nappy / Pexels

This is why Amazon dropped $10.8 billion on a satellite company. Not for e-commerce logistics. For the infrastructure that’ll matter when every military on Earth is running AI war-planning on their networks. It’s why Rolls-Royce is building a limited-run electric car instead of focusing on engines—because they know the energy demands of the next decade are going to be insane, and they want in on the power game.

The weird part? It’s working. Everyone’s terrified, so everyone’s spending. The fear is the economy now.

The Backlash You Can’t Market Away

Here’s where it gets messy for the industry. Palantir’s facing parliamentary scrutiny in the UK because hospitals are being pushed to use its software. Not as a suggestion. As NHS guidance. That’s not adoption. That’s mandated adoption, and it’s triggering the exact kind of regulatory alarm that kills golden geese.

Google’s about to start punishing websites that use dark patterns to trap users. That’s June. Seems small until you realize it’s a sign that user patience with tech deception is officially done. The industry spent two decades getting away with manipulative design. The meter’s running out.

And then there’s Elon Musk posting on TikTok while preparing to take SpaceX public. That’s not a sentence I thought I’d write. It’s pure chaos cosplay—using the app he’s allegedly trying to ban in the U.S. while simultaneously building a rocket company that’ll depend on government contracts. The cognitive dissonance isn’t a bug for Musk; it’s the feature. It’s what keeps people paying attention.

The pattern here is obvious: the more powerful these companies become, the less they can hide their methods. And the more they hide, the more they provoke.

My Read on What’s Actually Happening

I think we’re watching the tech industry simultaneously solve a real problem and create a new one so massive it might dwarf everything that came before.

The AI capability gap is real. If you’re a nation-state or a major corporation, you have to invest now or you’re toast in ten years. That’s not paranoia. That’s arithmetic. So everyone’s running the same race. And because it’s a race with no finish line and unlimited stakes, nobody can afford to slow down for ethics or regulation or public consent.

That creates a pressure cooker. You get violence from people who think you’re building doomsday machines. You get regulatory backlash from governments realizing they’ve handed too much power to private companies. You get satellite internet companies being bought for $10.8 billion not because of the service but because of the infrastructure war underneath.

Here’s what I’d bet on: within 18 months, at least one major Western government will announce a genuine AI safety mandate with teeth. Not guidelines. Actual penalties. It’ll be framed as protection against China and Russia, but it’ll also be a way to assert control over these companies. They’ll fight it. They’ll lobby. They’ll say it stifles innovation. And then they’ll comply, because the alternative—losing government contracts, getting broken up—is worse.

The violence will keep happening. Ideologically motivated extremists don’t respond to market logic. But the regulatory pressure will become the real constraint on how fast this can all move.

Photo by Markus Spiske / Pexels

The Uncomfortable Truth

Here’s what I genuinely don’t know: whether any of this slows down the actual development of the tech.

History suggests no. The nuclear arms race didn’t slow down because people were scared of nuclear weapons. It accelerated. The genetic engineering field didn’t pause because scientists worried about accidental plagues. It kept going, just with slightly more oversight. And the AI race won’t pause because a 20-year-old threw a bomb or because the NHS is getting questioned or because Google’s tweaking search rankings.

It might slow down. Regulations might bite. Public opinion might shift. But the underlying incentive structure—the one that says “whoever builds the most capable AI wins”—that doesn’t change until someone actually loses because they didn’t.

And by then, it’s probably too late.

What keeps me up is that unlike nuclear weapons or genetic engineering, AI isn’t built in visible facilities with contracts and oversight boards. It’s built in data centers. It’s trained on scraped internet data. It’s deployed to billions of people before most of them even know it’s there. You can’t inspect it like a nuclear facility. You can’t regulate it like a lab.

The industry knows this. That’s why they’re spending so hard on infrastructure right now. Not because they’re scared of regulation. Because they’re building the machine that’ll make regulation irrelevant.

What I’m Watching

• NHS Palantir backlash timeline. If parliamentary scrutiny actually leads to contract restrictions by Q4 2024, that’s a signal that European regulation might actually slow adoption. Watch whether other public institutions follow or back down.

• Amazon’s satellite internet deployment schedule. Not the PR announcements. The actual number of ground stations they build in the next 12 months. That’s the real indicator of how serious the infrastructure war actually is.

• The next AI violence incident. Not because I want it to happen, but because the previous one had a manifesto and a list. When (not if) the next one happens, watch whether it triggers legislative action or just media cycles. That tells you how much political will actually exists to constrain this thing.

• Quantum computing company consolidation. Europe’s talking about being competitive. Watch which companies get acquired by U.S. giants in the next two years. That’ll tell you whether the quantum race is actually competitive or already decided.