Somehow I missed the Anduril partnership announcement. I agree with you. National Security relationships in particular creates a moat that’s hard to replicate even with superior technology.
It seems possible OpenAI could maintain dominance in government/institutional markets while facing more competition in commercial segments, similar to how defense contractors operate.
Now we just need to find someone who disagrees with us and we can make a long bet.
It feels strange to say but I think that the product moat looks harder than the LLM moat for the top 5 teams right now. I'm surprised I think that, but I've assessed so many L and MLM models in the last 18 months, and they keep getting better, albeit more slowly, and they keep getting smaller while they lose less quality, and tooling keeps getting better on them.
At the same time, all the product infra around using, integrating, safety, API support, enterprise contracts, data security, threat analysis, all that is expensive and hard for startups in a way that spending $50mm with a cloud AI infra company is not hard.
Altman's new head of product is reputed to be excellent as well, so it will be super interesting to see where this all goes.
It seems possible OpenAI could maintain dominance in government/institutional markets while facing more competition in commercial segments, similar to how defense contractors operate.