why doesn't openai, anthropic... the intelligence agencies too, for that matter... leverage their real moat, the user data, to find the rare ones? the non-consensus talent, technical and not. surprised no one's fired up their machine yet to a broken-mirror pop-up... "you've been selected." modern day willy wonka. the hard part's done. the state and big tech already built the prompt-psych-eval pipeline. now it's just filtering... and a killer cta: "accept... or your messages might find their way out." nothing could go wrong.
talked to a guy from anthropic for a long time last night. fascinating conversation.
& believe it or not, one of their biggest problems is still hiring.
there simply aren’t enough exceptional ppl, esp for long term high conviction bets where the stuff is not obvious yet. i.e. people who can independently generate good judgment under extreme ambiguity.
@signulll it's lopsided in the right direction though... silence is a vacuum, and the brain hates a vacuum, but only a brain with something worth sharing in it... happy for you, i guess. also, sorry that happened to you.
This is a super cool website.
You can scroll around and look at any country's population pyramid, fertility rate, and mortality rate by year.
Check this out:
most founders speak one capital dialect fluently and can't order coffee in the other. venture, sure. debt, structured finance, public comps, blank stares. private and public markets price the same risk in different languages, and the edge is being bilingual. almost no one bothers, which is exactly why it's an edge.
the cap table is where founders lie to themselves. a clean one says you protected your options. a messy one says you never learned what the next investor needs to see. years of discipline, or the lack of it, on one page. the deck is the story you tell. the cap table is the one you can't.
"The most lasting impact of Jobs’s brief convergence with national security is simply that it allowed Silicon Valley’s most iconic entrepreneur to make his way back to Apple." - WSJ ... IYKYK?
spent decades making physical things and software, and the lesson rhymes. your factory is your real co-founder. so is your codebase. you can rewrite the pitch overnight. you cannot rewrite a supply chain, or an engineering culture, that took years to earn. what you can't rewrite is the only thing that's really yours.
raising for a holding company taught me the tell: if you have to explain why the pieces belong together, they don't. a real thesis gets repeated back to you in a sentence you didn't write. everything else is a spreadsheet with a disembodied narrator.
building my son an eighteen-year education in how things hold value. lately i think i owe him the older curriculum too, how the norse, the egyptians, the irish held death. one teaches him not to get eaten. the other teaches him it won't matter as much as he thinks.
them: why do you like bessent?
me: we've always lived in the age of prestidigitation... you learn to watch the hand that isn't waving...
but the variance between right/wrong, good/bad, life/death, and the sheer scale of what rides on it... seems greater than any other time in recorded history. magnitudes upon magnitudes.
the tricks are the same as they ever were... the stakes are not. and it's the most exciting thing imaginable... the fuel for human ingenuity. the higher the stakes, the purer and grander the outcomes.
to wilder and weirder times.
waste is the only real sin in business and the worst thing you can waste isn't money. it's a gift you were handed and never used because using it was scarier than complaining about not using it.
my grandfather once taught me...before Crichton / Gell-Mann Amnesia, and with a twist in a bone dry martini:
he who defines the terms... who counts, what counts, why it counts ... it is he who has cast a spell upon the masses.
every time i end up on wikipedia, wrong precisely where he cast it, i am reminded to add the twist.