Ask ten practitioners what changed and you will get ten answers, but they rally around the same rough shape: the thing that used to require a team now takes one person and an afternoon.
That compression is easy to celebrate and easy to underestimate. The work does not disappear; it moves up the stack, toward judgment, taste, and knowing which questions are worth asking in the first place.
Critics point out, correctly, that the flashy results are cherry-picked and the failure modes are quietly filed away. The honest read sits between the boosters and the skeptics, closer to the ground than either.
“We stopped measuring success by the demo and started measuring it by the boring Tuesday,” one engineering manager said. “Does it hold up when nobody is watching? That is the only benchmark that matters now.”
The economics, as always, decide who wins. Cheaper is never free — the cost reappears somewhere else, and the teams that map that shift early get to set the terms for everyone downstream.
What is certain is the direction, not the timeline. The trend is not going to reverse. The open question is who adapts in time to shape it, and who spends the next two years explaining the delay.