the close of our six-episode arc.
“good, fast, cheap — pick two” was the rule for thirty years. ai looks like it broke the rule — you can build a working app in twenty minutes now, for nearly nothing. all three corners at once.
except go look at what’s actually shipping. the demos lie. the agents hang. the thing works on the laptop and dies in production. the triangle didn’t die. it moved. the old third corner was “good.” the new third corner is “works in production” — and that’s the one ai doesn’t hand you for free.
this is the constructive turn. five episodes on what’s broken; this one is what we actually do about it, under the same cheap-and-fast pressure that produced everything we spent the arc criticizing. how we build agents fast without shipping the broken thing. what you sell when building itself is nearly free. and the part of the arc nobody noticed — that the “slow” enterprise was the only buyer in the whole story who refused to let you skip the part where it has to work.
live on linkedin. friday, june 12, 12 pm et.
the article goes up here the week after.