@EndicottInvests AI infra run rate makes 0 sense because no one really knows what rev will be in the next 5 years. Going to continue to go on an insane run and print $$ for anyone paying attention
Dumbest thing ever is when food influencers used Worcestershire in their recipe and completely over exaggerate how hard it is to pronounce #ItNotDatHard#WhyTheyDoDat
@VChokwa50038@unusual_whales Yeah realistically if you want super low latency api calls you’re going to have to spend $$. The best alternative would probably be to run an agent locally. Good Mac book mini use case right here
@david_nix@forgebitz Nah not an infra company and never will be, they’d be competing against the very own cloud footprints that they need to consume. It’s a agent based deployment, completely SaaS but does rely on infra scaling to grow arr
@shiri_shh Nebius Group. People are looking at their GPU infra but the real thing here is in ClickHouse which is a monster db rn. The amount of data running through them rn is outrageous and there’s really isn’t a better alternative since Pinecone died out