Agentic Engineering

jnakano89

"Tests are how you turn an unreliable agent into a reliable system" <— 100%. Once you're running 5+ agents in parallel, review-as-trust stops being physically possible. There isn't time, and there are too many concurrent diffs. The trust mechanism has to move into the gates: precise specs, deterministic test suites, exit codes as ground truth. It's roughly the same shift teams went through when they moved from "senior dev reads every PR" to "CI is the source of truth." Mechanical, unromantic, and the only thing that truly scales. On the skill atrophy point, I think the analysis is directionally right. The senior engineering skill that agents reward most won’t be review. It's having the system design fundamentals and writing the quality gates to assure functional needs are met and non-functional fundamentals are in tact.

rvz

"Agentic Engineering" (not vibe-coding, or vibe-engineering) sounds like a far better fit on what the future is going to be (or already is). You are still responsible for the code that the agent is producing. You still need to understand the outputs and the tests that make sense to create otherwise you are creating tests for the sake of testing, which is useless. The best comparison of all of this is autopilot and advanced driver-assistance systems (ADAS) in cars. You still need to watch the road at all times and keep your hands on the wheel just in case you need to intervene. The "vibe-coding" equivalent of this in cars means you are not looking at the road, nor you have hands on the wheel and allowing the car to drive itself without any intervention and no teleoperation. Great for short journeys but problematic for longer journeys and absolutely unacceptable when the AI goes down or is offline and you are stuck in the middle of the highway. So "Agentic Engineering" is much better without losing the responsibility.

k3liutZu

I could not read the whole piece. The phrasing smells like AI too much.