The AI Reckoning Has Begun

The violence came first.

A Molotov cocktail landed at Sam Altman’s San Francisco home. Police arrested a suspect. The device burned an exterior gate. And within hours, the internet did what it does: debated whether this was terrorism, protest, or something else entirely.

But here’s what interests me more than the incendiary device itself: the timing. Altman runs OpenAI, which just became a symbol of unchecked technological power at precisely the moment when three separate fronts are opening fire on that power. This isn’t random. This is what happens when innovation outpaces society’s ability to regulate it, and society gets angry.

A captivating abstract image featuring dynamic and glowing light patterns creating an eye-catching effect. Photo by Robert Clark / Pexels

Three Crises That Aren’t About Each Other (But Are)

Start with the nuts and bolts. Amazon is ending support for older Kindles—devices released before 2013. Users are losing the ability to download new e-books to hardware they own. This sounds like a small corporate decision. It isn’t.

It’s a reminder that you don’t own anything anymore. You license it. And when companies decide to pull the plug, you lose access. This happened quietly, with minimal pushback, because it’s just books on an old device. But scale that up. What happens when it’s your car? Your medical implant? Your neural interface?

Volkswagen just announced it’s scaling back EV production in Tennessee. The company’s pivoting back toward gasoline. This is the second major automaker to signal that the electric transition isn’t happening as fast as promised. The bet on AI-driven automation saving the EV industry? Still unproven.

Meanwhile, China, the U.S., and Russia are in an explicit arms race over AI-backed weapons systems. The comparison being made seriously now: this is the nuclear moment. We’re living through the 1945-1950 window where everyone’s scrambling to build the bomb before someone else does. Except there’s no Manhattan Project clarity, no unified purpose. Just competing nations rushing forward without guardrails.

Businessman reading a financial newspaper at a desk, highlighting finance and commerce theme. Photo by nappy / Pexels

The Addiction Question Nobody Wants to Answer

Meta just lost a landmark social media addiction lawsuit in California. The company was pulling Facebook ads recruiting for these lawsuits—basically damage control in real time. Here’s what that tells you: the business model that’s funded Silicon Valley for fifteen years is cracking under legal scrutiny.

Social media platforms are built on addictive design. That’s not controversial. It’s their competitive advantage. But when a court finds liability, even in one state, investors get nervous. And when Meta’s own PR team decides ads recruiting plaintiffs are bad optics, you know they’re preparing for a cascade of lawsuits.

The question nobody’s asking clearly: if AI makes content recommendation more effective, more personalized, more eerily accurate at keeping you scrolling, does that make the addiction liability worse? I think yes. And I think Meta knows it.

Where the Anger Is Coming From

London’s mayor is warning about a “disinformation blizzard.” He’s not talking about traditional media or even social platforms in the classical sense. He’s talking about AI-generated content, deepfakes, synthetic media designed to portray his city as failing. This is happening now. Not theoretically. Right now.

White House staff were just told not to place bets on prediction markets. The markets themselves aren’t illegal, but the concern—that government employees are monetizing non-public information—signals how fast these platforms are moving ahead of policy. Prediction markets are basically just AI-powered betting on everything: conflicts, election outcomes, economic data. They’re accelerating the financialization of chaos.

Then there’s Breanna Olson. She’s a dancer with motor neuron disease who performed on stage again through a digital avatar controlled by brainwave technology. She described it as re-establishing “the expression and connection she felt had gone.”

This isn’t a crisis. It’s the opposite. It’s miraculous. And it’s also the clearest possible signal of where this is heading: technology that directly interfaces with human consciousness. We haven’t even started to think about the regulation for that.

Detailed close-up of a newspaper displaying global financial market statistics and country flags. Photo by Markus Spiske / Pexels

My Read

I think we’re watching the transition from “tech companies operate in a permissive regulatory environment” to “tech companies are now facing existential legal and political risk.” It won’t happen overnight. But it’s happening.

The violence at Altman’s house is a symptom, not the disease. The disease is that AI has become powerful enough that it’s no longer optional for geopolitics, capital, and daily life—but we’re still pretending we can figure out the rules later. We can’t. The rules need to exist before the weapons systems are fully operational. Before the deepfakes elect someone. Before the neural interfaces are standard.

Here’s what I genuinely don’t know: will the backlash come through policy (regulation, legislation) or through market rejection (people stop using these tools) or through something messier (escalating conflict between people who see AI as essential and people who see it as existential threat)?

My guess: all three, simultaneously, and messily.

The Kindle thing is a preview. Amazon’s not going to stop killing old devices because customers are upset. They’re going to scale back only when it becomes profitable to do so or when regulators make it mandatory. Same with everything else. Meta won’t change because addiction lawsuits are embarrassing. They’ll change when the liability cost exceeds the profit margin.

But—and this is crucial—the nuclear weapons comparison isn’t actually apt. Nuclear proliferation was managed by a handful of governments with shared interest in not destroying civilization. AI is being developed by hundreds of competing entities with competing interests. That’s exponentially harder to control.

Elon Musk posting on TikTok and Instagram while preparing to take SpaceX public is almost funny: the guy who’s been most vocal about AI risk is simultaneously building the literal infrastructure that will depend on AI. He’s also the guy who literally tweeted that his company was in jeopardy because of AI concerns. That contradiction isn’t a bug in his thinking—it’s the whole contradiction of this moment.

What I’m Watching

Prediction market regulation timeline. The White House warning to staff is early-stage signal. Watch for formal SEC guidance or legislation within Q1 2025. If prediction markets get throttled by regulation, it means governments are starting to see these platforms as destabilizing. If they don’t, it means finance won and policy is still asleep.
The next social media addiction lawsuit verdict. California opened the door. Watch for similar suits in New York, Illinois, or UK courts through 2025. The threshold I’m watching: if liability is found and damages exceed $100 million, expect a flood of regulation. If damages stay under $50 million, companies will just treat it as a cost of business.
One major AI weapons incident that reaches mainstream news. Not theoretical warnings. An actual military use of AI-backed systems that causes measurable harm or escalates a regional conflict. This will force the nuclear comparison from think tanks into living rooms. I’d estimate 40% chance this happens in the next 18 months.
Amazon or Meta walking back a consumer-hostile policy due to regulatory threat, not customer demand. This is the real tell. When they change behavior because of law, not because they care, it means the regulatory window is actually closing.

The age of tech doing whatever it wants is over. What replaces it is still being written. And yeah, sometimes people throw firebombs at the people writing it.