The issue is that the risk of investing in a startup is extremely high even if you do limit investment to serial founders or post-product, post-traction busine

The issue is that the risk of investing in a startup is extremely high even if you do limit investment to serial founders or post-product, post-traction businesses. It's already very hard to generate a positive return even with those conditions, and impossible to tell with certainty which will work out. Investing in random unproven founders with no product or traction (apparently) makes it practically impossible. Even fewer will work out than in the pre-screened batch, and the hit rate is already close to zero. If it were at all possible, there would be no need for VCs at all. You'd just fill out some forms about your idea and then do an IPO.