Original Summary

I’ve been thinking about a different way to invest in very early founders, and I genuinely want people to tell me where this breaks. One thing that bothers me about finance is how much effort goes into trying to predict risk before anything has actually happened. We look at credit scores, income, history, pitch decks, founder pedigree, and a bunch of other signals, then try to decide how much someone is worth betting on. But if someone comes to me asking for $10,000, why do I need to decide today whether they are worth $10,000? Why not spend $200 finding out? Say John has an idea for a business. Instead of approving or denying him for the full amount, we figure out the smallest, useful experiment that would tell us something important. Maybe all he really needs at first is a domain, a simple landing page, and a waitlist. Give him $200 and see what happens. Can he get 100 people to sign up? If he can’t, we learned something for $200 instead of discovering the same problem after putting $10,000 into the idea. If he can, maybe the next experiment gets $500. Then $2,000. Then eventually $10,000. The basic idea is that capital grows with evidence. Now imagine doing this across a thousand people. Instead of putting $500,000 into one founder because they look great on paper, maybe you run small experiments across a large number of people and only start committing serious money once you have real evidence about what they can actually do. Obviously $500 is not enough to test every kind of business, and the exact number is not really the point. The principle is to commit the smallest amount of capital needed to answer the next important question. What makes me think this might be more practical now is AI. I don’t mean AI picking stocks or a chatbot telling people what to invest in. I mean using AI to make the analysis and follow-up cheap enough that you could actually pay close attention to thousands of small experiments. Each person could have an ongoing case that tracks what they are trying to accomplish, what assumptions are being tested, what happened, why it happened, what was learned, and what should happen next. More capital, a different experiment, a pivot, or stopping entirely. Humans would still make the important investment decisions. The AI would mainly make it possible to understand far more people and far more experiments than a team of analysts could realistically follow today. The institution would also learn from everyone at once. If five people try roughly the same strategy and all fail for the same reason, the sixth person should not have to waste money discovering the exact same thing again. I also don’t think failure would automatically mean someone was a bad investment. Maybe John’s first idea completely fails, but during the process we discover that he moves quickly, reports bad news honestly, learns fast, and is careful with money. That is useful information too. His original idea may have been wrong, but the person might still be w

  • 情报分类：商业与市场研究
  • 分类依据：内容涉及商业、投资或市场动态
  • 信息来源：Reddit · SideProject
  • 发布时间：2026/10/5 03:49:47