skip to content

Using AI as Software Engineer

The main point of using AI is to stay flexible. Things change fast, but some basics remain — here’s more about them.

Good things to seek as a software craftsman in AI era

  • Spot high-impact changes quickly
    • This keeps AI from silently degrading your architecture. You’re the most important guardrail.
  • Know what you want before you engage with AI
    • You still need product vision. You can’t define inputs and outputs for a system you haven’t thought through. AI without a clear goal produces a mess that takes more human effort to untangle.
  • Capture knowledge about the problem domain
    • Split findings into small chunks, write them down, and let the agent synthesize everything in the next session.
  • Automate aggressively
    • A lot is automated out of the box today. Leverage tools like Claude Skills or Hooks to streamline your workflow even further.

Things to stop doing as a software craftsman in AI era

  • Losing control when the speed increases
    • You’re the guardrail — don’t let AI silently accumulate debt or introduce security vulnerabilities.
  • Focusing on code cosmetics
  • Doing any low-cognitive-effort task yourself
  • Diving into solution without gathering context and planning possible options first