I don't think that's true. Latency is critical when the limiting factor is that you know what you want to do but in order to do it, you need to issue many requests all at once. With GPT4 I can issue requests at a high enough level that the limiting factor is actually me thinking of what the next request will be.
Huh? I understand that latency is a limiting factor for Cursorless, I'm not arguing with that. I'm saying that if you're using GPT, latency stops being the limiting factor most of the time.
