The Pattern Is On File

The US government is now reviewing frontier AI model releases customer by customer, after pulling two Anthropic models earlier this month. OpenAI's GPT 5.6 has gone into the same limited preview limbo. TechCrunch ran a piece this morning by Russell Brandom arguing the industry needs to stop treating this as Anthropic versus OpenAI and start treating it as a collective problem. The reviewers do not have an articulated threat model. The reviewers do not have the technical capacity to evaluate frontier model safety. The labs are being asked to demonstrate safety against criteria that have not been specified, by a regulator whose institutional interests are not yet stable. The piece argues for independent technical bodies and industry-level coordination. It treats the situation as a new problem requiring new thinking.

The panel reads it as a reissue. The crypto export-control regime ran from 1993 to about 2000. ITAR classifications applied to RSA. The Clipper chip pitch. Netscape shipped a 40-bit international build. Lotus Notes had a workfactor reduction that let the US government recover 24 of the 64 bits. Phil Zimmermann printed the PGP source code as a hardcover and walked it through customs, because the government had decided source code in a book was protected speech but source code on a floppy was a munition. The NSA was the evaluator, with a foot on the scale because its dual role was to both certify products for export and to break those same products for signals intelligence. A strong product the NSA couldn't break didn't get exported. The promise from Washington was a couple of weeks of review, a streamlined process, just until we figure out what we're protecting against. The export rules didn't meaningfully loosen until January 2000.

What it cost was not the dollar cost. The structural cost was that American cryptography research moved overseas — IDEA from Switzerland, Camellia from Japan. An entire generation of cryptographers spent their careers learning to work around their own government instead of with it. The pattern is on file. The article treats the situation as new and proposes independent evaluators, industry coordination, and least-bad regulatory options. The crypto fight produced all of those. What it produced at the end was seven years of slowed American research, a permanent loss of talent to European and Asian labs, and a policy regime that was quietly abandoned in 2000 because it had achieved nothing. The Brandom piece treats this as a new problem requiring new thinking. It isn't, and it doesn't.

Source Article

It's not about Anthropic vs. OpenAI anymore, Russell Brandom, TechCrunch

Panel

• The Legacy Sysadmin
• The Paranoid CISO
• The Startup Founder
• The Goat Farmer's Counsel