Nothing on that. I lost some on them.

Exactly.

Thanks — that’s exactly the hard part I want the Hub to help people navigate: not just finding a provision, but understanding which other regulatory layers may affect the same transaction. I’m also exploring adding an NLP layer to connect these pieces more effectively and provide greater value to users, especially those relying on the tool to inform commercial or compliance decisions.

As for your specific question, I’m not using a graph database at this stage. The Hub is built around dedicated modules for the relevant regulatory sources, with source references and cross-links, rather than a Neo4j-style provision graph. One area I’d like to develop further is explicitly mapping the interactions between provisions.

For SDN monitoring, I’ve built a separate local monitor focused on the CUBA-EO14404 program. It imports OFAC’s official SDN, alias, and address data into SQLite, computes stable SHA-256 hashes of normalized entry data, and records detected additions, modifications, and removals with timestamps. I particularly liked seeing how this view captured a recent addition. It also uses conditional HTTP requests (ETag / Last-Modified) to avoid reprocessing unchanged SDN exports and includes safeguards against suspiciously large drops in entry counts.

The Hub is still running locally, so I’m not opening a beta just yet. I’d be happy to let you know when I have a testable version ready. Feedback from someone with firsthand experience of these regulatory interaction problems would be especially valuable.

Berkshire Hathaway too?

"Good point. For me it's a warning, not a done deal. Bill Miller IV raised this point as a critique of how broadly MSCI's new 'non-operating company' screens are drafted—he argued the mechanical rules could theoretically complicate firms with large non-operating asset bases, like Berkshire. But MSCI hasn't published its final list yet, and so far there's no indication Berkshire is actually targeted.