Reviewing AI-generated code is the new bottleneck

I find myself reviewing AI-generated code less and less because the code quality is very good -- and getting better every month. Soon reviewing AI-generated code will be like manually verifying the results of a calculator--it is a waste of time.

flowerlad · Oct 17, 2026 · comment in “Microsoft and CMU team up to train early career software engineers”

HN users can try it out, give you feedback, and ask questions in the thread" (https://news.ycombinator.com/showhn.html) That isn't possible with many AI generated projects because OP didn't write the code, probably doesn't understand the code and can't discuss the code or their process in creating it (because said process doesn't exist.) There is little if anything of substance to discuss or play with to satisfy int…

krapp · Sep 22, 2026 · comment in “Ask HN: Is AI use not allowed in projects which is shared here to show?”

I'm perfectly happy to put time into reviewing AI code as long as the people behind that PR are, well, people. Tools like AgentScan are helping reduce the burden of the spam that doesn't follow the contributing guidelines, and giving us time back to review real PRs.

onlyspaceghost · Aug 17, 2026 · comment in “Humanity in Open Source”

Now ops can point claude to the repo and it can most likely [but not always] find the issue. So this ops person can do most of the work of reproducing the bug and finding the problem/solution but doesn't understand the code.

mlrtime · Aug 3, 2026 · comment in “Don't be a meat proxy”

Show HN: Vdiff – CLI to help you review AI-generated code Hey, you probably already saw that reviewing AI-generated code is a nightmare and quickly becomes a bottleneck.

fforbeck · May 2, 2026 · story in “Show HN: Vdiff – CLI to help you review AI-generated code”

To me, the important thing isn't that I can run it, it's that I can pay someone else to run it. I'm finding Opus 4.7 seems to be weirdly broken compared to 4.6, it just doesn't understand my code, breaks it whenever I ask it to do anything.

CJefferson · Apr 24, 2026 · comment in “DeepSeek v4”

The security angle is one of the most underappreciated risks in vibe coding. When developers don't understand the code that gets generated, they can't assess whether it's introducing vulnerabilities — hallucinated packages, insecure patterns, broken auth — and your finding that a "crazy amount" of AI-generated repos have severe structural flaws matches what a lot of teams are discovering in production.

nachocoll · Mar 10, 2026 · comment in “Security with Vibe Coding Platforms”

> My biggest unexpected finding: we’re hitting limits of traditional tools for sharing code. The pull request model on GitHub doesn’t carry enough information to review AI generated code properly — I wish I could see the prompts that led to changes.

bigfishrunning · Dec 23, 2025 · comment in “A year of vibes”

> The pull request model on GitHub doesn’t carry enough information to review AI generated code properly — I wish I could see the prompts that led to changes.

anshulbhide · Dec 22, 2025 · comment in “A year of vibes”

> My biggest unexpected finding: we’re hitting limits of traditional tools for sharing code. The pull request model on GitHub doesn’t carry enough information to review AI generated code properly — I wish I could see the prompts that led to changes.

divbzero · Dec 22, 2025 · comment in “A year of vibes”