It's true that VCs that invest in utter nonsense will eventually be eliminated, but I think most VCs who never invest in a success invest in enough doomed models that survive until IPO, (often with built-in failures like a model of exploitation that requires obscurity.)

The highs and lows of tech average together well enough to satisfy Wall Street and then have huge profits for some that VCs that find and hold good investments, and more VCs that feed off a constant influx of sucker investors.