AI for operators

Topic1 pieces

What language models are genuinely good for in an operating business, and where they quietly cost more than they save.

Written for people who run something, not for people who build models. The test is always the same: does this hold up when it is wrong, and who notices when it is?

AI security

Benchmarks you can actually run

Most AI security numbers are unfalsifiable. Here's the standard I hold my own project to, and why it made the numbers worse.

May 28, 20263 min read