Thank you kindly Tom.
BTW, there are great resources for these topics:
https://www.knowledgegraph.tech/
https://www.connected-data.london/
In terms of why haven't graph technologies caught on more? A few key reasons, though there are more:
1. For the organizations on Wall St, or its London equiv, or the various 3-letter, 4-letter, 5-letter agencies which almost all make LOTS of use of graph technologies ... they use this tech although they typically don't talk _publicly_ about it. Too much risk of losing advantage. That said, to paraphrase recent quote from a friend in Pharma , if you aren't using these technologies (LLMs, GNNs, KGs., etc.) you aren't doing Pharma.
2. VCs love platforms, that way they can drive founder teams to follow a "known business model", aka copy Oracle if you don't have any better ideas. SQL platforms will tend to have wider audience, and most of the graph platforms focus on "graph queries" (again, "copy Oracle if you don't have better ideas") when graph analytics are the more powerful tech.
3. I used to work for a former IBM exec, who'd been in the room when Larry Ellison came down to license their "R" source code for launching Oracle. He remarked that IBM had sophisticated graph database platforms in the 1970s. For example, most people who've ever used parent/child links in IMS understand they are doing graph traversals. However, after user testing IBM decided they needed to "dumb down" the tech with guardrails if they wanted a broader _total addressable market_, so they made SQL instead. It's still true that simple graphs are easy, and complex graphs are mind-bogglingly complex to manage.
