If you grew up in a country where corruption was just the weather, you develop a sixth sense for it. You learn to smell the rot beneath the polished press releases. And right now, the air in Washington D.C. reeks of it.
When the government bans your product and then magically lifts the restriction weeks later, that’s not regulation. That’s a shakedown.
The Trump administration just lifted its restrictions on OpenAI’s GPT 5.6. A few weeks ago, they slammed the brakes on it. Today, the brakes are off. If you work in tech, you’re probably breathing a sigh of relief, thinking the system worked. You’re missing the point entirely.
The ban was never about AI safety. The lift was never about innovation. This is a predictable, dangerous cycle of restriction and reversal. It’s a deliberate strategy to signal that favorable treatment is available—to those willing to pay the toll.
We’ve seen this playbook in banana republics, not in the self-proclaimed capital of the free world. An executive inserts themselves into the decisions of private companies, creating artificial chaos. Then, once the company aligns, or pays the right political price, the chaos mysteriously vanishes. One commenter nailed it: it feels like a service where the government bans your product for a few weeks, then lifts it when you buy the right coin or kiss the right ring.
We thought we were building the future of intelligence. Instead, we’re just paying rent to the king.
This isn’t just an OpenAI problem. It’s a precedent. If the executive branch can arbitrarily freeze a model release and then unfreeze it based on undisclosed negotiations, then the entire AI market is now hostage to political whims. Market confidence doesn’t survive in an environment where the rules change depending on who is in power and who they had dinner with last night.
If you are an AI founder, your compute budget is no longer your biggest concern. Your lobbying budget is. If you are an investor, you aren’t evaluating technical moats anymore; you’re evaluating political allegiances. The government has effectively turned AI regulation into a rent-seeking mechanism.
We are trading a future of open, principled innovation for a system of patronage. The rule of law is being replaced by the rule of the deal.
The scariest part of the AI future isn’t that the machines will take over. It’s that they’ll be fully owned by whoever pays the political toll.
FAQ
Q: Isn't this just standard regulatory back-and-forth?
A: No. Standard regulation involves years of public comment, transparent safety standards, and independent review. A sudden executive ban followed by a quick, quiet lift reeks of a negotiated, behind-closed-doors deal.
Q: What does this mean for AI startups?
A: Expect to need a political lobbying budget as large as your compute budget. If your model threatens the wrong political interests, you should expect a sudden 'safety review' that only ends when you align.
Q: Could this actually be good for keeping AI in check?
A: Absolutely not. It means AI safety is now subjective to whoever holds executive power. It's no longer about technical alignment; it's about political allegiance.