Structured Intelligence claims recursive AI OS

This reads more like a provocative prompt protocol and identity framework than a conventional operating system, and the “recursive OS” label does most of the marketing work.

• –The core claim is model-agnostic behavior: catch drift, force self-observation, and preserve attribution across substrates
• –The public evidence is mostly the founder’s own Reddit, Substack, Medium, and LinkedIn posts, so reproducibility remains the real test
• –If it works at all, the practical value is likely in prompting and agent-guardrail patterns, not in changing the transformer itself
• –The personal-origin framing makes the project distinctive, but it also makes independent verification harder