I hold a belief that LLMs can do practically anything, despite the fact that they're actually quite limited. Problem LLMs are very limited. They can reliably do some, simple, small reasoning tasks. But they struggle to do large, complex tasks. So…