Narrating Your Thought Process
cross-cutting · 4 min · Email: 'Focus on explaining your thought process as you build'
Why Narration Wins
The interviewer is answering one question: 'Would I want to pair-program with this person?' They're also asking: 'Could they ship this if I left the room?' Narration answers both. Too scripted can come off as canned. Extended silence reads as opaque. Narrating the obvious risks condescension. The sweet spot: talk like you're thinking alongside someone you respect — irregular, genuine, unperformed.
The Fallback Mantra
"Done with X. Next: Y, because Z." — This is your safety net, not your script. Use it when you're stuck, tired, or your brain is full. It makes your train of thought traceable even if the interviewer stepped out. But if every update has the same shape, you sound like you're reciting a prep guide. Real thinking is irregular — let your narration be irregular too.
Recoverable Reasoning
Goal: a colleague who stepped away could follow the reasoning behind each major choice. Name rejected alternatives when there's a real fork: 'I went with X over Y because…' Chaining 'next I'll do X, next I'll do Y' without ever naming a rejected alternative is a missed opportunity to demonstrate depth — not a fail, just room to improve. If you genuinely only know one path, say so plainly. Fabricated alternatives are worse than none.
When Silence Is Correct
Typing boilerplate. Fixing a typo. Executing a single obvious path. Tracing a complex data flow through multiple services — genuine thinking takes time and silence is fine. But if you've been quiet through two distinct actions (e.g., reading output then typing), surface what you're doing. The screen shows what; you're saying why.
Overload Hierarchy
When your brain is full: Layer 1 (always possible): 'Done with X. Next: Y, because Z.' Layer 2 (if you have capacity): name one rejected alternative. Layer 3 (if you have more): self-test — 'Most important decision I made — did I say why?' Layer 1 alone carries most of the signal the interviewer needs.
Anti-Patterns
Hoarding reasoning — the interviewer asks 'why did you do X?' and you have an answer but didn't volunteer it. Fixed-shape narration — every update the same chose/rejected/because triplet reads as scripted grammar, not thinking. Meta-apologizing — 'I should have narrated that' can sound like you're reciting a prep guide rather than self-correcting naturally. If you catch a miss, recover by narrating the next decision — don't narrate the miss.
The interviewer is answering one question: 'Would I want to pair-program with this person?' Every behavior should make the answer yes. If a rule tells you to do something annoying, ignore the rule.
Full Guide → · AI Toolchain & Pre-Flight Setup → · Sketching Architecture Live → · Prototype Strategy — Thin Slice to Working Demo →
Flashcards for this topic