Don't Answer the First Question

· ai · Source ↗

TLDR

  • Perfetto engineer argues the confusion behind a weird user question is a diagnostic signal worth exploring, not a puzzle to decode and discard.

Key Takeaways

  • XY problem framing stops short: the confusion itself is an opening that improves both the user’s mental model and your product radar.
  • Checklist: have you seen this before, does the ask fit the tool’s architecture, what context is missing?
  • Perfetto trace-splitting requests revealed users who didn’t know about periodic snapshots, solving a problem they shouldn’t have had.
  • Premature building is expensive: ad-hoc UI customization created a year of technical debt before a proper plugin API was designed.
  • Holding off on trace merging until the problem space was well understood led to a maintainable implementation built once.

Hacker News Comment Review

  • Commenters split on execution: the technique works when you answer the immediate question first, then probe deeper; gatekeeping the answer until questions are answered is the failure mode.
  • The SO comparison is pointed: “You shouldn’t do X so I won’t tell you how” is the anti-pattern this approach risks if the immediate answer is withheld rather than offered alongside the reframe.

Notable Comments

  • @dgellow: experienced a coworker who ran the quizzing process but still gave no answer, leaving the impression the process served the responder more than the asker.

Original | Discuss on HN