I’d argue there’s a _bit_ of secret sauce here, but the question is if there’s enough to justify valuations of the prop-AI firms, and that seems unlikely.