Experienced Engineers Converge On Strict Constraints For AI Coding Agents
Reactions from ranked influencers
5 posts
I love this because I see so many experienced engineers converging on the same approaches. Once you see the shape of the machine, the method to extract maximum value from your time is really obvious to people who have worked in large systems for years.
I’m significantly older than you. I started coding in the late 60s. My current strategy is to not read any of the code written by my agents. That’s the only way I can take advantage of their productivity. What I do instead is to surround the agents with extreme constraints. Unit tests, gherkin tests, QA procedures, quality metrics, mutation testing, test coverage, and a plethora of others. In the end, I have very high confidence in the code they produce because they’ve had to run the gauntlet of all of my constraints and tests.
I expect this kind of approach will be systematized in a manner similar to other development methodologies. It's not just code + agents, it's a system you need to operate well as a whole. There are better and worse ways to do this, and general principles one can extract.
Any mature software engineer feels pain when from most code in the AI ecosystem. It's obviously written by people who are not experts, and so it is full of terrible decisions with obviously bad downstream consequences. Things are going to get so much better soon.
Manhattan project but it's about forcing all the Old Unix Guys to work with agents until they fix the ecosystem.
The result will be named paradigms with manifestos and holy wars, OOP-vs-functional tier. What we have now is the era before anyone had said the word "object".