If youre using a transformer model, why reinvent the stack? LLMs adapted to code, speech, proteins, all with custom vocab, modalities, and objectives. Why is b

If youre using a transformer model, why reinvent the stack? LLMs adapted to code, speech, proteins, all with custom vocab, modalities, and objectives. Why is behavior a domain that requires a whole new foundation, instead of fine-tuning? What breaks if you build on existing LLM infra?