Ollama: All Aboard Open Models

Ollama just rolled into town with a huge $88 million funding round and a big, cinematic pitch: open artificial intelligence for everyone, on your own computer, with your own data, no giant company holding the keys. The founders are also playing the comeback story card hard — these are the same people behind Kitematic, which became part of Docker Desktop, now used by millions. On paper, it’s a classic tech glow-up: old startup heroes return, rally the open-source crowd, and declare that AI should be personal, private, and yours.

But in the comments? Oh, it got spicy fast. The loudest gripe wasn’t even about the money — it was about credit. One commenter came in swinging, basically asking whether Ollama will ever properly acknowledge llama.cpp, the open project many believe powers key parts of the magic. Others heard “VC funding” and immediately translated it to: enshittification incoming. That’s internet shorthand for “it starts great, then gets worse once investors want returns.” Ouch.

Then came the product snobs. Some users flat-out said don’t use Ollama, claiming it’s slower than alternatives and that its model files aren’t the best around. One person even dropped a dramatic anti-Ollama link like they were arriving with courtroom evidence. There was also a very practical complaint: if Ollama says it’s all-in on open AI, why no easy subscription option for embedding models? So while Ollama is celebrating its victory lap, the community is doing what the community does best: cheering, heckling, and nitpicking in equal measure.