You don't. This is demonstrated by
>Claude Code CLI ... is there anyone in the world that knows how that spaghetti machine works
Having a very active, and happy user base.
Now this a simplistic view, and I think even heavily AI first dev teams still glance at the diff to make sure things aren't going too awry or ask questions. But the concept "spaghetti code" and "optimized for human understanding" are assumptions being revisited entirely in 2026.
EDIT: also, from the software engineer perspective, it is an order of magnitude more tedious to maintain software projects that lack quality, consistency, or standards. If I could measure the effect of AI vibe-coding in terms of personal pain, I would say it makes software development more painful when code is not well thought-out. There is a way to use AI responsibly, but I see more often people taking shortcuts and throwing "engineering" out the window. It's possible to ship quality with AI, but that's not the path of least resistance.
That's the neat thing—you don't.
Frontier models, especially Opus 5.5, have vastly stepped up their game in terms of code quality. Furthermore, the sheer volume and speed of code they produce has meant that humans just don't have time to read and review the code even if it were the most well-written code in the world. The frontier is now structuring your loops, graphs, and evals in such a way that agents with a fresh context review the agent-produced code looking for bugs and errors. If humans review anything, it's an agent-produced summary of the code changes.
Business people now expect speed of delivery that humans can't match. Humans are now the bottleneck; agents will step up to replace them.