the frontier is not where the money is landing

dario amodei published an essay last week arguing the ai industry should slow itself down.

within days sam altman agreed and matched the commitment, elon said dario is right, and hassabis and nadella lined up behind it. the labs agreed with each other in public for once.

one man in jakarta couldn't have cared less.

i spent thirty minutes this morning with an operator who runs a corporate venture studio in southeast asia. he wanted to talk about a marketplace his co-founder built that charges sellers zero percent, in a market where everyone else takes twenty or thirty.

it's profitable. it runs on open models. nobody near the top of any leaderboard had a hand in it.

pace the frontier all you like. that business lives three layers below the frontier.

the value keeps landing away from the place the capability gets made. it shows up in the hands of people running models that were state of the art eighteen months ago and cost close to nothing today.

his second point stuck harder. the tension between american and chinese models reads to him as an opening, because every country that would rather not run its economy on somebody else's politics needs a local version of everything. he sees a market where the west sees a safety debate.

and he was blunt about his own region in a way people avoid. the power law doesn't work there. venture math built for silicon valley outcomes dies on contact with indonesia. so they take real stakes in boring companies, roll up an accounting firm and make it ai native, and chase profit while everyone else chases a unicorn.

i'm fine with pacing. i'd rather five labs argue this in the open than settle it privately, and permanent access for outside evaluators is a real commitment.

pacing the frontier does nothing to the man in jakarta. it does nothing to the accounting roll up, the zero percent marketplace, or the thousand small companies being rebuilt on last year's weights. that engine left the building and no lab holds its throttle.

headlines live at the frontier. money lives in the diffusion. those two have never shared an address, and the gap widens every quarter.

so while everyone watches five companies negotiate the speed limit.... who's watching the road?