if i don’t work at palantir i might just start my own version p much all llm work ive been doing is in the same interest, identifying real world problems, understanding how the existing enterprise data is organized, and building from the model up for use case (or modifying as deep as you can go if using frontier models), and having a language-level thus token-level understanding of the data, rather the language of it even if it’s a database, to be able to make decisions in the business or in war zerebro was stripped down and built up for creativity use cases, content creation, etc. o media used the same research to build synthetic social profiles to test “backrooms” agent-to-agent autonomous interactions all of these were just llms, the tooling above it for context fetching is always going to be commodified i think it’d be great to start that, a general use case oriented software company, branch into hardware for deployments, internal on prem centers, and later on partnership or internal department i always keep saying agi isn’t a certain level intelligence in a model it’s when it’s deployed, and making more decisions around you then you can even compute biologically