‹ BackHN Continuity

Thread

The problem is not AI code, but not knowing about system architecture or intent

388 points · 240 comments · zazuke

  1. feverzsj · · focus · HN ↗
    The problem is always maintainability, which can only be achieved through human understanding of code.
    1. Supermancho · · focus · HN ↗
      Understanding code is trivial, with AI support. LLMs are better at summarizing and identifying side effects than writing code. Granted, not perfect, but about as good as a person.

      Looking at an architecture, module boundaries, and comparing the organization to the domain is the part LLMs don't do well and is unironically difficult for humans. The less you care about the domain (ie having a mental model), the harder it is.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.