The engineer as the correction sensor
Model a system built with AI the way you'd model a thermostat: a closed control loop.
The agent produces the output. We read that output and correct it: tighten a skill, sharpen the context, adjust the routing. The value was never in typing the code. It's in judging whether what came out is true, and worth shipping.
That reframes the job. You stop guarding every line and start owning the result. Most of the time we barely touch the code. We shape the rules precisely enough that the agent writes something you can read like a sentence, then we check it against the business. The senior isn't removed from the loop. The senior is the loop.
This is how we'd run your backlog.
Bring a real one to a Sprint Zero and watch the loop close on your code.
More field notes