Thats true if we already had a way of knowing they would fail but we don't. Startups are how you try new ideas to find the good ones. Do having them work there and fail creates informational value.
This is exactly the point. It's very hard to know which ideas will work, and which teams are best suited for the task. The beauty of the VC model is that it starts companies on the cheap. Many companies fail, and those people can go chase either the companies that won, or new startups. It's much better than a world where you have a couple monopolies with grand strategic plans that don't actually get anything done.
The East Coast Monopoly mindset was why Xerox PARC had money to create things, but never got anything done with them.
