GUgus_massa·Feb 20, 2026
In The Car Wash Problem: A variable isolation study on prompt architecturejo-heejin
interview_mate
HISTORICAL RECORD2026–2026
jo-heejin / interview_mate · Not publicly available
interview_mate
Discover this repository through its Hacker News conversations.
2Distinct people
2Mentions
0Explicit recommenders
What to know before exploring
What it is
A public description has not been verified.
Where it is used
No documented product use has been verified yet.
What people are saying
No recommendation phrase was found in the indexed comments.
No classified warning is attached to this page. Read the discussions before treating silence as approval.
Also in these discussions
No other repository is repeatedly linked in the same discussions.
This week on Hacker News
Nov 7–Nov 14 UTC · Direct repository links and separately labeled related project stories. Indexed through 11/14/2026, 1:49:46 PM UTC.
No direct repository links were found in the indexed HN discussions over the past seven days. Older discussions are available below.
Mentions over time
HN mentions by hour2 dated mentions
Observed mentions: 2026-02-20 13:23:56 UTC to 2026-02-20 19:25:20 UTC.
Counts come from the repository’s dated source evidence. Index cutoff: 2026-11-02 08:16:59 UTC.
View mention counts and date ranges
| Period | Start (UTC) | End (exclusive, UTC) | Mentions | Submissions | Comments |
|---|---|---|---|---|---|
| Feb 20, 13:00 | 2026-02-20 13:00:00 UTC | 2026-02-20 14:00:00 UTC | 1 | 1 | 0 |
| Feb 20, 14:00 | 2026-02-20 14:00:00 UTC | 2026-02-20 15:00:00 UTC | 0 | 0 | 0 |
| Feb 20, 15:00 | 2026-02-20 15:00:00 UTC | 2026-02-20 16:00:00 UTC | 0 | 0 | 0 |
| Feb 20, 16:00 | 2026-02-20 16:00:00 UTC | 2026-02-20 17:00:00 UTC | 0 | 0 | 0 |
| Feb 20, 17:00 | 2026-02-20 17:00:00 UTC | 2026-02-20 18:00:00 UTC | 0 | 0 | 0 |
| Feb 20, 18:00 | 2026-02-20 18:00:00 UTC | 2026-02-20 19:00:00 UTC | 0 | 0 | 0 |
| Feb 20, 19:00 | 2026-02-20 19:00:00 UTC | 2026-02-20 20:00:00 UTC | 1 | 0 | 1 |
What Hacker News says
2 sourcesThe Car Wash Problem: A variable isolation study on prompt architecture
Most AI products inject facts and hope reasoning follows. But intelligence is not measured by how much a model holds in its context window.
It is measured by knowing to pick up the keys before leaving the house.
Last week, the "Car Wash problem" (50m away, walk or drive?) went viral here on HN. Every major LLM failed because they missed the implicit physical constraint: the car must be there.
While testing InterviewMate's prompt architecture, I posed the same question. It answered drive immediately. Every other LLM had failed. But I didn't actually know why it worked — so I ran a variable isolation study to find out.
100 API calls, Claude Sonnet 4.5, 5 conditions:
Baseline (no prompt): 0%
Role only: 0%
Context injection (user profile, car location): 30%
Structured reasoning (STAR framework): 85%
Full stack (both combined): 100%
Throwing f
Read the rest
acts at the model doesn't work unless the architecture forces it to explicitly evaluate the task goal first. Without structure, the model jumps straight to the distance heuristic: "100m is short, walk."
I'm writing a paper on this. Wanted to share the raw data with HN first.
Code and raw eval data: https://github.com/JO-HEEJIN/interview_mate/tree/main/car_wash