Mario & Armin: A good engineer says no, a lot

Categories: Product, Startup

Summary

Good engineering requires saying 'no' frequently, but AI agents flip this dynamic—enabling quick 'yes' decisions that create problems. Senior engineers must actively push back against agent-generated solutions within 48 hours before junior team members treat them as gospel.

Key Takeaways

  1. Senior engineers' role is gatekeeping through strategic rejection. Track how many feature requests your team says 'no' to weekly—low numbers signal weak technical leadership or over-reliance on automation.
  2. AI agents remove friction from saying 'yes,' creating decision velocity without judgment. Set a policy: all agent-generated solutions require senior review before 48-hour implementation window closes.
  3. Junior engineers will defer to agent outputs as authoritative. The 'ChatGPT printout to doctor' analogy shows how tools can bypass critical thinking—establish a team norm that agent suggestions are starting points, not conclusions.
  4. AI-assisted development accelerates bad decisions at scale. Measure: review how many requirements generated by agents your team implements vs. rejects monthly. Healthy ratio should favor rejection.

Topics

Transcript Excerpt

A good engineer is an engineer that says [music] no a lot. And if you're using agents, the exact opposite happens. You say, "Yes, I want this and that and that and I want this because I don't have to type it myself." And that's where all the problems start. The senior person says no, knowing something, and then 48 hours later, the junior comes by [music] and said, "I talked to the agent. Now I have all the evidence of why we shouldn't be doing it this way." >> It's like people going to the doctor with a chat GPT printout and saying, "This is what the machine said. You better do that."…