LLMS still can’t troubleshoot. I have been thinking about how to solve it and I think an entirely new architecture is needed. Quants got replaced by quant bots the same way human calculators got replaced by electronic adding machines.

I haven’t met an AI who can think clearly in a systems setting. I’m guessing it might be a decade or two. As an example, ask an LLM to do some 10th grade math. Inspect the thinking process. It can regurgitate the process and the rules but cannot perform them.

Same with troubleshooting.