I don't think I'll ever be up-to-date with all the progress in AI or trends in VC, but relying on his own expertise where he says the best VC's only identify 2 out of 10 of the really good opportunities, I would estimate that once somebody gets AI to do better than that, there may be some impact.
Some professions are going to be easier to replace than a Radiologist, others not so much.
That's why I've done science like art since way before 1980.
AI was comin' fast back then and I wanted to be resilient :)
It's especially cute hearing Marc "Invest in Crypto and the Police State" Andreessen pretend he's the augur of smart investments. Give a man an inch of credibility, and he'll take a mile.