Product Launch
anonymous
132 points
61 comments
Posted12 days agoActive6d ago
AI Code Review
oldmanrahul.comlauncheddeveloper_tooldevelopers
Key Features
AI-powered code reviewinstant feedbackcode analysis
Tech Stack
AI
Key Features
Tech Stack
(That's if you don't want to go full Codex and have an agent play around with the PR. Personally I find that GPT-5.2 xhigh is incredibly good at analysing diffs-plus-context without tools.)
Here are the commits, the tasks were not trivial
https://github.com/hofstadter-io/hof/commits/_next/
Social posts and pretty pictures as I work on my custom copilot replacement
"Diff to master and review the changes. Branch designed to address <problem statement>. Write output to d:\claudeOut in typst (.typ) format."
It'll do the diffs and search both branch and master versions of files.
I prefer reading PDFs than markdown, but it'll default to markdown unprompted if you prefer.
I have almost all my workspaces configured with /add-dir to add d:/claudeOut and d:/claudeIn as general scratch folders for temporary in/out file permissions so it can read/write outside the context of the workspace for things like this.
You might get better results using a better crafted prompt (or code review skill?). In general I find claude code reviews are:
- Overly fussy about null checking everything
- Completely miss on whether the PR has properly distilled the problem down to its essence
- Are good at catching spelling mistakes
- Like to pretend they know if something is well architectured, but doesn't
It will absolutely assume too much from naming, so it's kind of a good spot if it's making wrong kind of assumptions about how parts work, to think how to name things more clearly.
e.g. If you write a class called "AddingFactory", it'll go around assuming that's what it does, even if the core of it returns (a, b) -> a*b.
You have to then work hard to get it to properly examine the file and convince itself that it is actually a multiplier.
Obviously real-world examples are more subtle than that, but if you're finding yourself arguing with it, it's worth sometimes considering whether you should rename things.
The objective of this initial review is to catch those low-hanging fruit that your colleagues would waste cycles on.
LLMs can catch syntax and basic semantics. Peers can spend time on more interesting things like design and relevant biz context.
As for PR reviews, assuming you've got linting and static analysis out the way, you'd need to enter a sufficiently reasonable prompt to truly catch problems or surface reviews that match your standard and not generic AI comments.
My company uses some automatic AI PR review bots, and they annoy me more than they help. Lots of useless comments
> My company uses some automatic AI PR review bots, and they annoy me more than they help. Lots of useless comments
One way to make them more useful is to ask to list the topN problems found in the change set.
You can also append ".patch" and get a more useful output
My experience is you can trust any code that is well tested, human or AI generated. And you cannot trust any code that is not well tested (what I call "vibe tested"). But some constraints need to be in natural language, and for that you need a LLM to review the PRs.
We are sooo gonna get replaced soon...
And those are the worst models we will have used from now on.
Now, I use it as well to review, just like you mention it pulls it via gh, has all the source to reference and then tells me what it thinks. But it can't be left alone.
Similarly people have been trying to pass root cause analyses off as true and they sound confident but have holes like a good Swiss cheese.
Colleague's feedback:
Claude> Address comments on PR #123
I didn't see this mentioned, but we've been running bugbot for a while now and it's very good. It catches so many subtle bugs.
Sure, they will leave comments about common made errors (your editor should already warn about this before you even commit it) etc. But to notify about this weird thing that was done to make sure something a lot of customers wanted is made reality.
also, PR's are created to share knowledge. Questions and answers on them are to spread knowledge in the team. AI does not do that.
[edit] Added the part about knowledge sharing
But they're fantastic at spotting dumb mistakes or low-hanging fruit for improvements!
And having the AI spot those for you first means you don't waste your team's valuable reviewing time on the simple stuff that you could have caught early.
A PR is also a decent UI for getting the feedback but especially so for documenting/discussing the AI review suggestions with the team, just like human review.
AI review is also not equivalent to linter and static checks. It can suggest practices appropriate for the language and appropriate for your code base. Like a lot of my AI experiences it's pretty hit or miss and it's non-deterministic but it doesn't have much cost to disregard the misses and I appreciate the hits.
The ergonomics of doing this in pre-commit make no sense.
Spin up a PR in GitHub and get Cursor and/or Claude to do a code review — it’s amazing.
It’ll often spot bugs (not only obvious ones), it’ll utilise your agent.md to spot mismatched coding style, missing documentation, it’ll check sentry to see if this part of the code touches a hotspot or a LOC that’s been throwing off errors … it’s an amazing first pass.
Once all the issues are resolved you can mark the PR as ready for review and get a human to look big picture.
It’s unquestionably a huge time saver for reviewers.
And having the AI and human review take place with the same UX (comments attached to lines of code, being able to chat to the AI to explain decisions, having the AI resolve the comment when satisfied) just makes sense and is an obvious time saver for the submitter.
In most of the repos I work with, it tends to make a large number of false positive or inappropriate suggestions that are just plain wrong for the code base in question. Sometimes these might be ok in some settings, but are generally just wrong. About 1 in every 10~20 comments is actually useful or something novel that hasn't been caught elsewhere etc. The net effect is that the AI reviewer we're effectively forced to use is just noise that get's ignored because it's so wrong so often.
He'd make giant, 100+ file changes, 1000+ worded PRs. Impossible to review. eventually he just modified the permissions to require a single approval, approves his changes and merges. This is still going on, but he's isolated to repos he made himself
He'd copy/paste the output from AI on other people's reviews. Often they were false positives or open ended questions. So he automated his side, but doubled or tripled the work of the person requesting the review. not to mention the ai's comments were 100-300 words with formatting and emojis.
The contractors refused to address any comments made by him. Some felt it was massively disrespectful as they put tons of time and effort into their changes and he can't even bother to read it himself.
It got to the CTO. And AI reviews have been banned.
But it HAS helped the one Jr guy on the team prepare for reviews and understand review comments better.
AI code review does not replace human review. But AI reviewers will often notice little things that a human may miss. Sometimes the things they flag are false positives, but it's still worth checking in on them. If even one logical error or edge case gets caught by an AI reviewer that would've otherwise made it to production with just human review, it's a win.
Some AI reviewers will also factor in context of related files not visible in the diff. Humans can do this, but it's time consuming, and many don't.
AI reviews are also a great place to put "lint" like rules that would be complicated to express in standard linting tools like Eslint.
We currently run 3-4 AI reviewers on our PRs. The biggest problem I run into is outdated knowledge. We've had AI reviewers leave comments based on limitations of DynamoDB or whatever that haven't been true for the last year or two. And of course it feels tedious when 3 bots all leave similar comments on the same line, but even that is useful as reinforcement of a signal.
AI is a mixed bag. I'm the type of person who is compelled to have a deep understanding the code they write. Writing my own code vs fixing AI-generated code is a wash timewise, and the AI generated code is so limited (assuming you pared down the uselessly elaborate code and fixed all the critical runtime bugs) as to restrict further iterations. And I'm talking about uploading an architectural blueprint with every function a documented but otherwise empty stub.
AI is a great bellwether. I bounce ideas off AI for a consensus. The closest equivalent is reading StackOverflow comments. I once offhandedly complained that python had no equivalent for setattr at class scope (as __class__ is not defined until after __new__), but AI showed me how to provide a closure in __prepare__ over the class namespace, which was introduced in 3.3 (?) and to which I paid little attention. What a gem.
AI is great for learning. If you follow a textbook or blog or paper and don't understand, AI can clarify. But be careful with less structured learning - it is important to build a full mental model accounting for every possible outcome and explanation, otherwise you're susceptible to hallucinations. I remember my first derivative in which the end result could be obtained via two separate proofs, one of which would imply an incorrect calculus. You've got to play with it until you're satisfied your mental model accounts for all the facts.
Because AI facilitates learning so easily I feel the best skills for a future generation are those pertaining to memory and retention. Ya know, assuming we don't develop individualized and personalized AI that can model your next word and act as a personal memex.
AI is great as a search assistant. I have much better recall when rereading content. Thus, I prefer to ask AI to search for links to content I vaguely recall, rather than ask AI for it's own summary or recollection.
Despite being terrible at writing decent code, AI provides fantastic code review. It catches everything from subtle high-level errors to api mismatches to syntax errors. I actually wish I was as fluent at namespaces and cgroups as AI, and I'm well versed.
AI is interesting at comments. It can be hard to provide germane comments while wading through the weeds. I feel the best comments are written a few days after the code is complete. AI provides that fresh perspective instantly. And if AI can't understand your code, you had better improve your comments.
I'm about to try AI for unit tests. I prefer hypothesis tests. I took a quick glance at the generated code, and it seemed overly complicated. So I'm not optimistic.
Outside of software, AI is great for things you don't really care about. Yeah it might hallucinate, but my back-and-forth with AI is more about holding up a mirror to myself, revealing inner biases. Especially useful for interior decoration / remodeling.
Of course, take everything I said with a grain of salt as security at my company actively discourages AI. So, everything I said applies to free/cheap plans only. And I haven't tried skills yet.
In order to be time efficient, human review should focus on the 'what' rather than the 'how' in most cases.
This is only one purpose of a code review. Another is to catch bugs, potential performance issues, etc and AI is pretty good at that. It shouldn't be the only reviewer of course.
edit: I saw your below comment, I agree that it's best to have AI review your code before you even put it up as a PR. But either way, I think it's valuable.
i use https://www.coderabbit.ai/ and it tends to be aware of files that aren't in the diff, and definitely can see the rest of the file your are editing (not just the lines in the diff)
also works if you have the GitHub cli installed. Would setup an AGENTS.md or SKILL.md to instruct an agent on how to use gh too.
Not affiliated with Hacker News or Y Combinator. We simply enrich the public API with analytics.