Hearing the same thing from tech folks in SF. They see the current AI opportunity window. They have ideas. But they do not have the capital to iterate aggressively enough to uncover a breakout. Most angel investing already backs an idea, even if it is pre-pmf or pre-product. Random thought. What if it shifts even earlier? You are not betting on the idea. You are betting on the person you have vetted to find the idea. Help them fund the tinkering phase. Burn as many llm tokens as they want. Run all experiments. Kill most of them. Until they find the one. Discovery speed is the new moat.