Marc Andreessen on The Future of VC: Will a16z Go Public & Why Introspection is Dangerous?

Summary

• Andreessen’s most tradeable claim: the AI labor-displacement narrative is “100% incorrect” — “it’s classic zero-sum economics, the lump of labor fallacy.” The layoff wave is rates going 0→5% “at record speed” plus COVID overhiring; every large company is overstaffed “at least 25%,” most by 50%, many by 75%, and AI is “the silver bullet excuse” — his evidence is that until “literally December” AI wasn’t good enough to do the jobs being cut.
• AI’s likely Schumpeterian gains go ~99% to users, not builders: like the internet and smartphones, “it might even be 99.9999% of the value of AI is going to accrue to the users.” The best AI in the world is a $20 (or free) consumer app heading toward 5 billion smartphone users — “the most hyper-democratic technology we’ve ever seen.” Investing in AI companies is a fight over the residual 1%.
• Yet the industry itself has whiplash-recentralized in Silicon Valley: after believing 2020–2023 had “cracked the code” on remote decentralization, Andreessen now says tech is more concentrated than ever — “something very close to 100% of the quality AI companies” sit within a 20-mile radius of him (11 Labs, Black Forest Labs the exceptions).
• On price: overfunding is “just as dangerous or more dangerous than underfunding” and no investor ever does a down round in someone else’s company — but at venture stage, “every time we passed on a promising venture company over price, I think it’s been a mistake.” And “don’t ever do diamonds in the rough, only do diamonds” — 9 times out of 10, or perhaps 95 or 99 times out of 100, if it’s investable, hungry VCs have found it.
• Learning from mistakes can be dangerous in venture because it creates scalded-stove aversion — AI itself “was a tremendously good way to lose a lot of money from 1945 to 2017.” Marc and Ben’s real job is keeping partners weighted toward the mistake of omission (not investing in Google) over the mistake of commission (losing $10M).
• Arthur Rock’s conclusion, which Marc adopts: he’d have been a better investor feeding every business plan “straight into the shredder” and spending 100% of time on the résumé — back special people “almost without consideration of other factors.”
• a16z going public: no catalyst — “there’s nothing that we’re missing today that we could solve by going public,” though he’d “never rule anything out.” The two products the firm has “kicked around for a long time” and not launched: public equity and credit.
• The likely Flow bet on Adam Neumann rested on one real-estate legend’s line: only two people in history built commercial real-estate brands people care about — a US president and Adam Neumann. The mirror-image regret: passing on a likely Anduril Series A because “we got scared off” by the politics — “100% we would not make that same mistake again.”

Deep dive

Not yet available upstream; scheduled sync will retry.