Is it really a race? It feels more like a slog. I continue to try to use AI (google, openai, and anthropic), and it continues to be a pain in the ass. Their consumer interfaces are garbage, both being buggy/bloated and clunky to work over multiple threads, with its "memory" being nearly nonexistent outside a single thread. They randomly fail to do the thing they did successfully 5 minutes ago. I struggle to get them to do basic things while other things they do effortlessly. They're bad at logic, spatial reasoning/engineering, and I have to constantly correct them. Often they'll do things in agents that I never asked them to do, and I have to then undo it... The time I used to spend doing things manually, I now spend in fixing the thing that's supposed to be automating the manual work... and no matter how I try to fix it, it finds a new way to randomly fail. I am much happier just doing things by hand.

It sounds like you have found an approach that works for you, and that's great. In my experience I've had to devote a lot of time to learning to use AI tools. Most of this learning is understanding how to create the necessary context for success and getting an intuition for what questions to ask.