I read this as “10% are hot”—probably the author meant “like a team out of OAI” but maybe they meant “like recent rounds of OAI have been”.
In any case this post read as sophisticated, reasonable, and helpful, to me.
One single OpenAI-tier deal every two years, consistently, would put you into Legendary VC territory.
- Volume: Former founders who hit the jackpot are investing in a bunch of startups every week. There are hundreds of unicorns etc out there, and thousands of Series A's every year. That's a bunch of competitive deals being signed every day!
- Quality: Early silicon valley startups can easily look like openai's early days. With hundreds of unicorns out there, their execs eventually leaving and recruiting smart folks for their next thing happens almost every day. Pitches that are "We're solving X" are a dime a dozen. Likewise, they're each flawed in different ways -- in OpenAI's case, an easy negative phrasing is: no real business plan, positioned mostly as a non-profit R&D lab that'd do open source for Elon Musk to get google IP more easily. Likewise, having Stripe's CTO was one of those cool unicorn exec things, but for 0->1 business, maybe not so obvious, and Sam Altman's only 0->1 gig was running a small failed social mobile social network. It's easy to phrase in positive vs negative lights. Now imagine getting 5 of those on your desk every week, and you only pick 0-3 a year, hoping each win pays for all the duds, and then some...
You can bet on all 10% potential OpenAIs of the future and still loose all your bets.
Now I can believe that at least 10% of the pitchdecks they receive are "we hook $industry up to OpenAI's API"...