Ask HN: How to deal with shipping AI-written code you don't understand?
- ID
- 37230
- Status
- summarized
- Published
- 19 Oct 2026, 7:24 PM
- Fetched
- 20 Oct 2026, 8:59 AM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://news.ycombinator.com/item?id=50151036
- Source URL
- https://hnrss.org/best
Summary
- Score
- 7.5
- Created
- 20 Oct 2026, 8:59 AM
- Tags
- Audience
- developersvibe_codersai_ml_learnerssaas_founders
What happened
An Ask HN thread (245 points, 388 comments) by user rzk asks how engineers handle shipping AI-written code they don't fully understand, quoting a LinkedIn post from someone who describes themselves as an engineer at Meta saying their code now feels like a black box and that reviewing every line 'is going to take forever.' The replies split sharply: one commenter with roughly 20 years of experience says they prompt AI and don't review the code at all, comparing it to not understanding how their car or residence was built, while others argue about how much understanding is actually required before shipping.
Why it matters
The thread gives you no consensus to copy, so you have to pick a review threshold yourself — and the specific pressure point is the quoted claim that line-by-line review doesn't scale. Decide per-repo which files must be understood before merge (auth, payments, data migrations, anything touching production state) versus which you'll accept at the high-level-only level, and write that down before the next AI-assisted PR lands, because the commenters who skipped review entirely and the ones reviewing everything are both operating without a stated rule.
Discussion angle
Ask everyone in the room to name the last AI-generated piece of code they shipped without reading line by line, and what made that file safe to skip — then compare answers to see if the group actually has a shared threshold or is just improvising.