Sam Altman rizzes up Congress on the first date

What's going on here?

The ChatGPT guy got summoned to Congress today to explain this whole AI thing.

Makes sense. ChatGPT hit 100 million users in two months, and most senators who read “A.I.” still think it's a guy named Al.

Normally these tech CEO hearings are a beatdown: TikTok's CEO got roasted for five straight hours (although 80% of that was about him being “Chinese” because Singapore “isn't a country” lmao), Congress dragged Bezos out of his rocket lab in 2020 just to yell at him, and Zuckerberg still has PTSD from 2018.

But instead of arguing, Sam agreed with every concern, admitted his own tech scares him, and straight up begged Congress to regulate him.

He had the senators swooning.

Sen. Durbin said he can't recall corporations ever coming in to “plead with us to regulate them.” And Sen. Kennedy basically offered Sam the job of running the agency that would regulate him.

The room was his. So Sam pulled out his wishlist:

• A new federal agency that licenses any AI model above a certain capability level (and can revoke that license)
• Safety standards, including tests for whether a model “can self-replicate and self-exfiltrate into the wild” (his own words)
• Independent audits by outside experts

Not bad Sam, not bad at all…

What does this mean?

Congress finally met a tech CEO willing to play ball (suspicious, but okay).

To be fair to Sam, the safety concerns are real:

• Chatbots inventing facts
• Jobs getting automated away
• Voice clones and deepfakes scamming people
• Disinformation campaigns targeting voters before the upcoming election

However, “please regulate me” also happens to be the oldest play in the regulatory capture handbook.

Regulatory capture is when the industry being regulated ends up writing the rules for its own benefit. The big player helps design requirements it can easily afford (licenses, audits, compliance lawyers) while smaller rivals drown in the paperwork and fees. The regulation meant to police the giant ends up protecting it. AT&T ran this play. Big Tobacco ran this play. Countless others have too. And it works disgustingly well.

Maybe, just maybe… Sam is begging for regulation because it would lock in OpenAI's lead.

The fine print of Sam's ask is also suspicious: licenses only apply to models “above a certain capability level.” Today, that's basically three companies: OpenAI, Google, and Anthropic.

Sam may mean every word. But the plan that protects humanity and the plan that protects OpenAI are the same plan.

Convenient.

Why should I care?

1) AI law just became “when,” not “if.”

Section 230 (the legal shield that protects social platforms) doesn't cover AI. Congress admits it fumbled social media and refuses to run it back. So new rules are coming. The question is who writes them, and today Congress handed the pen to the guy being regulated.

2) The strongest moat is one written into law.

If Sam's licensing wall goes up, the next AI startup needs a government permission slip just to train a model, open source gets walled out, and three companies set the price of intelligence. So watch the threshold: if compliance turns out to be easy for OpenAI, Google, and Anthropic, but brutal for everyone else… Then Sam's rizz worked.