Comment by googletron

It’s simply not understanding. In terms of latency, LLM processing for large fact-based decision-making was incredibly slow. The real unlock here was speed in decision-making.