It's totally infeasible for a small startup without significant funding to build a competitive LLM from scratch. What is possible is fine tuning an existing LL

It's totally infeasible for a small startup without significant funding to build a competitive LLM from scratch. What is possible is fine tuning an existing LLM. I suspect that this will become popular. One can imagine open source platforms with built-in plugin registries for easily installable agents. Each agent may come with it's own knowledgebase and vectors or keyword for selecting them. It may be a quantized fine tuned LoRA that can be loaded on the fly.