I have good results with this prompt after every larger change: Now do a final code check. Is everything tidy and do the components adhere to the principle of separations-of-concerns. Is everything in an understandable and maintainable state? Do we make any assumptions that may not be true anymore? Is any code left over from previous edits or experiments that does not belong into the codebase? Is the documentation still representing the current state of code?
Write yourself a /review command. That is an empty markdown file at `.claude/commands/review.md`. In it, put a checklist of things the agent should look for. When you’re ready to have your agent review the code, type `/review`. The checklist will be examined and it’ll plan out some findings to ask you if you want them fixed.
Mine starts with “Enter plan mode. Examine the differences on this branch vs. main. Consider: ...” and proceeds to a bullet list of things.
Any time I notice something in code review and have to get the agent to fix it.. I throw it on the list!
My list is like 200 items now. Know what? Agents don’t care that they just got a wall of generic feedback, they happily look into all the bullet points.
I added “ensure the new things aren’t duplicating code that already exists elsewhere” and it gave me such a surprise - it really truly started planning cleanups!
We are just scratching the surface. We have to give tools to our tools so they can use them to be better tools for us.
I am curious what does it contain, for me a lot of times its a back and forth with agent until it "looks good to my eyes and taste", but haven't written any such list yet, because it is context dependant, in some projects I forgive minor issues, or allow magical numbers, but in other projects I force agent to use constants with meaningful names `SECONDS_IN_A_DAY = 24 * 60 * 60`
You should do an experiment of splitting that up to multiple reviews that are logically together. My hypothesis is that you may be losing signal due to the amount of text expected back.
Right now the comments that upset me the most are LLM TMI-style comments that break encapsulation by talking about the behavior of a function's callers
Sure, but the proportion of code that drives homicidal behavior is heavily weighted towards non-comments. You're a lot more likely to piss off whoever inherits your code with the code that actually does something being bad or a lack of documentation than with comments.
I'm quite fine with no comments but correctly named variables and functions. This can't become out of sync contrarily to "out of band" comments. I take this over commented code with poorly named stuff any day.
I've also seen a lot of comments that restate what the code already says and that's just noise, more work to keep in sync, an additional thing that can fail, and more cognitive load because you have to read twice the same thing (best case, if code and comment are still in sync). That's the result you risk when you think you must comment your code.
I appreciate the occasional comment that explains why something seems overly tricky or weird or not immediately intuitive. Once, I had left such a comment that saved myself years later from making a mistake. Of course, this should be kept at a minimal level. It leads to me liking clear code with few comments the most. (Some guidelines, even if it's not perfect, to limit complexity and spaghetti code help a lot).
Function, class, module documentation is also useful so you don't have to read the whole thing and you know what it's intended to provide (which is slightly different than simply what it provides, and this differences is important).
I worked with this guy. He'd write the code, and comment where needed, and then he would ask "How can I make this comment unnecessary?" The answer was usually to rename something, so that what he was doing was obvious.
I continually run codebases through different models to have them look for bad code smells like repeated code. That's been pretty effective. You do have to maintain over time or else you end up with a sloppy mess which I can only imagine compounds.
> continually run codebases through different models
When I have spare usage before a reset, I just throw a part (depending on how much usage left) of a non-critical codebase to refactor overnight and push to PR.
If it's marginally more maintainable/better after review it's good to merge.
That sounds like a good idea, but shipping 10x as many features and bugfixes sounds better.
I started using AI with the best intentions. Checking everything before committing. Improving output by hand if it didn't quite follow the existing code style guidelines or variables were not named as well as they should be. Or if it did something sloppy or hacky.
Now, AI GOES BURRRRRRRRRRRR! If the tests pass it's good to ship. AI can deal with the problems it may create. No problems so far.
How did you know you're not stuck at a local optimum where the AI could iterate even faster if you enforced higher quality on what it produced?
To make up some hypothetical numbers in order to illustrate with math: if you ship bugfixes 10x faster but then have 11x more bugs you need to fix, that's not a net improvement. Even if it's only 5x more bugs, maybe you could reduce that to 2x if you changed how you worked to only be 8x as fast in a way that produced higher quality code. Similarly, maybe you could cut the time it needed to produce a new feature by 50% if your code were higher quality by moving 20% slower.
My point in all of this isn't that you literally need to work the same way you did before you had these tools, but that framing it as either "move fast and ignore the code" and "use the same exact heuristics you would in the pre-LLM days for what code is acceptable" is a false dichotomy. If you aren't thinking about how effectively you're using these tools and whether there are changes you could make to move even faster because "AI go brrr", I think you've lost the plot in the same way you probably think that other people in this thread have.
Are you my colleague? It's fine if it's your own personal app, but please don't do this in a large complex codebase in a team. It's entirely depressing. You can use AI and still write good code. I think it's actually probably easier to write maintainable code with AI.
> That sounds like a good idea, but shipping 10x as many features and bugfixes sounds better.
This work great until you reach a certain size, then good (or even "not bad") code is required otherwise the model spins its wheel trying to ensure the change is correct.
The way I've measured how good/bad the code is (for AI) is to have one "baseline fixed change" that I measure how long time it takes to implement. Always in the beginning (less than 10K LOC, as just some measurement), this baseline change will take 2-3 minutes. As you add more code, the same change starts to take 5-6 minutes, and once you hit 1 million LOC, it can take as long as 10 minutes, even though the change is the same.
It's when this baseline task starts to take longer time, that you need to update the design/architecture/layout/whatever, to better fit the task/domain, and to actually make it easy to maintain and still possible to add changes without spending 10 minutes. So its at this point you refactor, and once done, the baseline task will again be easy for the model to do.
So yeah, if all you do is smaller projects, then "shipping 10x as many features" is easy and doable, for the lifetime of the projects. But once the projects start to accumulate technical debt, the model will have a harder time making sure the changes are correct, and suddenly "shipping 2x as many features" is maybe doable, but you could still have had 10x if you just spend slightly more time on the actual design and architecture of the program.
Yeah, the flip side of the article is that Fable level models can fix the majority of codebases created from the past 3 years and one shot it to a fixable state that is "human maintainable"
> shipping 10x as many features and bugfixes sounds better
I understand you're excited about the tool, but for the sake of earnest discussion here, maybe commenters like yourself can tone the hype down to plausibility?
Claims like this are just nonsense. It's not how product development works.
How do you even have so many bugs left to fix if the tool is so fast and productive? Surely, you didn't have a backlog of tens of thousands of bugs that you're still chewing through? And of course, the volume of new bugs much be minimal since the AI-composed additions introduce "no problems so far". If it works like you say, which we'll accept in good faith per HN guidelines, you must have exhausted your backlog long ago.
And if you've indeed exhausted your bug backlog long ago (incredible!), you're left to talking about shipping "10x as many features". Yet no product has a limitless capacity for features. Nobody would want to use software so bloated and churning that was gaining features at such a pace. And who is designing and specifying them so quickly anyway? If it works like you say, which we again accept in good faith, you must have stalled out on your feature list long ago.
If the AI indeed allows you to "[ship] 10x as many features and bugfixes", and we take what you say in good faith, then one of the following seems to be implied:
* you've fixed all your bugs and blew through your mature feature designs already, leaving your AI agents sitting idle for all but a few hours a week, while you're bottlenecked on feature design and your software product is bloated beyond imagination
* your coding productivity before AI was absolutely glacial by industry standards such that "10x" productivity for you is actually much closer to "0.5-2x" for others
Man, I bet Jia Tan is simultaneously kicking themselves and having a field day. All those years of wasted effort gaining trust and making good contributions to try to land a sophisticated backdoor into a tool via layers of indirection, and then not long after we have devs just going “I don’t need to read this code, or prioritise, or think about what makes sense, just prompt for fractals of kitchen sinks and ship it”.
Anthropic themselves have admitted you don’t need much to poison LLMs¹. I can’t wait for us to discover the backdoors that are being introduced. I hope it happens soon so people get to their senses. Bah, what am I saying, when (not if) that happens, the response will just be to throw more LLMs at it.
I find myself doing this but then I worry that the slop will just compound and 3, 6, 12 months from now as my services scale I'll have a harder time operating them. Maybe I'm wrong.
The key idea here is that your codebase is context that will be used for future changes. And context determines the model’s output, so it’s still worth having a well-designed codebase.
Easier said than done to be honest, especially if there are many people (and their agents) pushing code. It’s hard to keep up these days.
CLAUDE.md (or AGENTS.md) is not good place to put all code change rules, because these rules would dillute the context and "distract" the agent e.g. during bugs triage or business analysis.
Instead modularize the knowledge with skills and specialized MD files. Agent should lazy load what is needed to do focused work.
Skills have usage description metadata, but with free files you can simply instruct agent with CLAUDE.md to load them, e.g.: "Before you attempt to change any frontend code first load and follow `docs/{JS|HTML|CSS}_coding_rules.md`".
linting tools, static analysis, CPD, etc. These are all old things you can continue to use and are much more robust than anything you can prompt. These should be standard when using LLMs. In fact, you can tighten the rules even more enforcing more restrictions so you ONLY get the output you want. put this behind a pre-commit hook and a CI job that runs on a PR, and it will work wonders.
You can have all the prompts you want on top of this, but if you don't have this automated stuff running behind the scenes, you aren't serious about these issues.
Looking through some of these comments here, I see lots of people rewriting concrete rules in markdown willing to spend tokens on the hope AI won't miss it where an actual program won't.
It's interesting that the author didn't mention considering updating their agentic code review prompt to keep an eye out for repetitive/duplicate code.
One thing that can help a lot here is to spend a lot of tokens on QA steps. It's easy to see agents writing code very quickly and then just dumping the code in a PR to make it somebody else's concern. But agents can also speed up a lot of back and forth on bad code.
I've tried various forms of workflows to run dedicated QA, code review (of various flavors) simplification and text simplification agents. Especially the simplification goes a long way to remove dumb padding, duplication and efficiency. Dedicated docs/comment simplification is also becoming more and more necessary on recent models. For things like feature development in my workflow, the majority of time the agents run and tokens spent is critiquing the code from various perspectives and it's not close.
Of course, this doesn't solve the overall issue that agents don't write code like you and still requires a lot of human attention in planning and code review out to clean up leftover issues, and e.g. challenge bad assumptions about architecture and real-world context. A human is still very much needed to cull the slop (or, more gratuitously: align the agent). But IME it does help avoid a lot of pitfalls and makes the code high quality a lot more quickly.
AI is so miserable for this. It's so focused on doing what you ask, it forgets that there's stuff worth doing that you didn't ask for, like defining reasonable abstractions.
Getting away from stuff like this is exactly why I want to use AI. When I say "implement this for idle but active users," I _want_it to define isUserActiveIdle() and stuff these 4 conditionals in it. Having to check the generated code for stuff like this undoes, like .... all the benefit of using AI.
AI makes all these little decisions for us. I can about some of these decisions. I just want to notice when it's doing this without having to make my eyes bleed reading 10k lines of generated code a day.
Just run weekly cron job to assess code quality and highlight candidates for refactoring. In addition to doing the same in each PR, of course, but things can get through.
Have a bit of a contrarian view on this tl;dr don’t write code for human consumption if you use AI; BUT you have to accept AI coding lock in and change how you work.
I absolutely will not write corporate code like humans are maintaining it anymore, because I don’t have any confidence actual humans will be maintaining it.
For personal projects, I can trust that I myself will be maintaining things so I still write things like it matters, but I do not extend the trust to others.
AI is trained on a various range of code quality, including very low one.
I'm paid to provide good quality code and not flood my company with more average code than it should.
In my previous job, I could regularly reduce a PR code down to 10-20%% of its size because someone overlooked something or was just "overengineer" a feature.
AI are such "bullshiters" that they produce more text than necessary.
Code bloat was already real, but from my personal experience it becomes realer with AI.
The outcome of this will likely be apparent when no one can dive into any code base because of the amount of fluff in it (and you will obviously need more AI to deal with this).
Just this morning I was trying to scrape nitter, for funsies. One hour and neither gemini nor kimi were able to write something working, despite trying selenium (or playwright), beautiful soup, and a specific library that can be used to scrape it.
I eventually read the library docs and managed to build a scraper for what I wanted in a few mins. Llms are great for a lot of things, but sometimes you stumble in something that's just outside of what they know/can do and you're sol. And of all the thinks, I didn't expect they would fail at this, to be honest the opposite
People are so desperate for this to be true. Maybe it comes from a subconscious recognition that their own self-imposed deskilling will inevitably catch up with them.
It's unlikely that AI will get to the point where it makes handwritten coders redundant, and then not immediately be at the point where vibe coders are redundant too. So if you earnestly take the position that handwriting code is a "ngmi" type activity, you also need to take the position that the vibe coder (or agent- assisted-developer/loop-architect, or whatever its nom de guerre is this week) is "ngmi".
> I can do projects in 3 days what would take 6 months.
The hyperbole on this keeps growing every time I see it. Soon we’ll be having people claiming they can do in 12 seconds what used to take them 17 years. What is never presented is proof. People (and programmers are no exception) are notoriously bad at estimating. We already did studies where people thought they were being faster with LLMs when they were in fact being slower.
As companies begin to rehire to fix the mess made by LLMs, it’s clear that just getting something out the door isn’t enough. It never was. Maintenance is an important part of any long-standing system.
Why stop there? If you _use_ handwritten products you’re ngmi. I only use vibe coded operating systems, JavaScript sandboxes, compilers, TLS libraries, databases, rendering engines..
It’s bizarre to me that so many people feel the need to keep parroting this corporate talking point. What do you care? If you think people who eschew LLMs for coding “are not going to make it” or “are going to get left behind”¹, then let them. More opportunities for you, right? Go do your own thing.
¹ As if “moving forward” or “progress” were always a positive. It’s not. Just look at how many regulations we have to forbid or curtail uses of stuff we found to be harmful.
AI isnt taking my job. my company is supporting local AI for development. who ever comes after me will have the same hardware and models or better. unless a MBA is put in charge, my boss and predecessors can maintain and build out as needed.
bottom up AI use seems a godsend compared to the corporate AI rat race.
i setup some slop reporting systems and ensured my boss knows theyre great starting points but serious use requires real time investment.
I’m pretty sure many people who use AI to write emails or blog posts add "make it sound like a human wrote it" to their prompts. We all know what the result usually looks like.
If AI is writing my code, I'd rather have it focus purely on correctness and efficiency than on making the code easy to read.
heck! I might even ask it to imitate Arthur Whitney’s style.
I am following similar steps from this article https://www.lucasfcosta.com/blog/backpressure-is-all-you-nee...
Mine starts with “Enter plan mode. Examine the differences on this branch vs. main. Consider: ...” and proceeds to a bullet list of things.
Any time I notice something in code review and have to get the agent to fix it.. I throw it on the list!
My list is like 200 items now. Know what? Agents don’t care that they just got a wall of generic feedback, they happily look into all the bullet points.
I added “ensure the new things aren’t duplicating code that already exists elsewhere” and it gave me such a surprise - it really truly started planning cleanups!
We are just scratching the surface. We have to give tools to our tools so they can use them to be better tools for us.
I am curious what does it contain, for me a lot of times its a back and forth with agent until it "looks good to my eyes and taste", but haven't written any such list yet, because it is context dependant, in some projects I forgive minor issues, or allow magical numbers, but in other projects I force agent to use constants with meaningful names `SECONDS_IN_A_DAY = 24 * 60 * 60`
Will try your approach to distill the code to bullet points.
"Add comments to your code under the assumption that the next person to maintain it is a homicidal maniac who knows where you live"
I've also seen a lot of comments that restate what the code already says and that's just noise, more work to keep in sync, an additional thing that can fail, and more cognitive load because you have to read twice the same thing (best case, if code and comment are still in sync). That's the result you risk when you think you must comment your code.
I appreciate the occasional comment that explains why something seems overly tricky or weird or not immediately intuitive. Once, I had left such a comment that saved myself years later from making a mistake. Of course, this should be kept at a minimal level. It leads to me liking clear code with few comments the most. (Some guidelines, even if it's not perfect, to limit complexity and spaghetti code help a lot).
Function, class, module documentation is also useful so you don't have to read the whole thing and you know what it's intended to provide (which is slightly different than simply what it provides, and this differences is important).
Or is it more about the review process and a context reset?
When I have spare usage before a reset, I just throw a part (depending on how much usage left) of a non-critical codebase to refactor overnight and push to PR.
If it's marginally more maintainable/better after review it's good to merge.
I started using AI with the best intentions. Checking everything before committing. Improving output by hand if it didn't quite follow the existing code style guidelines or variables were not named as well as they should be. Or if it did something sloppy or hacky.
Now, AI GOES BURRRRRRRRRRRR! If the tests pass it's good to ship. AI can deal with the problems it may create. No problems so far.
To make up some hypothetical numbers in order to illustrate with math: if you ship bugfixes 10x faster but then have 11x more bugs you need to fix, that's not a net improvement. Even if it's only 5x more bugs, maybe you could reduce that to 2x if you changed how you worked to only be 8x as fast in a way that produced higher quality code. Similarly, maybe you could cut the time it needed to produce a new feature by 50% if your code were higher quality by moving 20% slower.
My point in all of this isn't that you literally need to work the same way you did before you had these tools, but that framing it as either "move fast and ignore the code" and "use the same exact heuristics you would in the pre-LLM days for what code is acceptable" is a false dichotomy. If you aren't thinking about how effectively you're using these tools and whether there are changes you could make to move even faster because "AI go brrr", I think you've lost the plot in the same way you probably think that other people in this thread have.
This work great until you reach a certain size, then good (or even "not bad") code is required otherwise the model spins its wheel trying to ensure the change is correct.
The way I've measured how good/bad the code is (for AI) is to have one "baseline fixed change" that I measure how long time it takes to implement. Always in the beginning (less than 10K LOC, as just some measurement), this baseline change will take 2-3 minutes. As you add more code, the same change starts to take 5-6 minutes, and once you hit 1 million LOC, it can take as long as 10 minutes, even though the change is the same.
It's when this baseline task starts to take longer time, that you need to update the design/architecture/layout/whatever, to better fit the task/domain, and to actually make it easy to maintain and still possible to add changes without spending 10 minutes. So its at this point you refactor, and once done, the baseline task will again be easy for the model to do.
So yeah, if all you do is smaller projects, then "shipping 10x as many features" is easy and doable, for the lifetime of the projects. But once the projects start to accumulate technical debt, the model will have a harder time making sure the changes are correct, and suddenly "shipping 2x as many features" is maybe doable, but you could still have had 10x if you just spend slightly more time on the actual design and architecture of the program.
The solution, as you say, is probably to break it down into isolated sub-components that are only aware of each other's APIs and nothing more.
I understand you're excited about the tool, but for the sake of earnest discussion here, maybe commenters like yourself can tone the hype down to plausibility?
Claims like this are just nonsense. It's not how product development works.
How do you even have so many bugs left to fix if the tool is so fast and productive? Surely, you didn't have a backlog of tens of thousands of bugs that you're still chewing through? And of course, the volume of new bugs much be minimal since the AI-composed additions introduce "no problems so far". If it works like you say, which we'll accept in good faith per HN guidelines, you must have exhausted your backlog long ago.
And if you've indeed exhausted your bug backlog long ago (incredible!), you're left to talking about shipping "10x as many features". Yet no product has a limitless capacity for features. Nobody would want to use software so bloated and churning that was gaining features at such a pace. And who is designing and specifying them so quickly anyway? If it works like you say, which we again accept in good faith, you must have stalled out on your feature list long ago.
If the AI indeed allows you to "[ship] 10x as many features and bugfixes", and we take what you say in good faith, then one of the following seems to be implied:
* you've fixed all your bugs and blew through your mature feature designs already, leaving your AI agents sitting idle for all but a few hours a week, while you're bottlenecked on feature design and your software product is bloated beyond imagination
* your coding productivity before AI was absolutely glacial by industry standards such that "10x" productivity for you is actually much closer to "0.5-2x" for others
Any insight into which of those it might be?
Anthropic themselves have admitted you don’t need much to poison LLMs¹. I can’t wait for us to discover the backdoors that are being introduced. I hope it happens soon so people get to their senses. Bah, what am I saying, when (not if) that happens, the response will just be to throw more LLMs at it.
¹ https://www.anthropic.com/research/small-samples-poison
* define the software layers, their function, and the max depth allowed
* establish a corp code formatter for each language, along with a process to PR it
* establish a business vocabulary and what the terms mean
* establish a data dictionary, make it part of the database schema/table/col comments
Are far more successful with LLMs. You _should_ have been doing this years ago, but with LLMs its a super power.
Taken from: https://github.com/zakirullin/cognitive-load/blob/main/READM...
Easier said than done to be honest, especially if there are many people (and their agents) pushing code. It’s hard to keep up these days.
Maybe someone has the perfect claude.md that solves this problem but I have not seen it.
Instead modularize the knowledge with skills and specialized MD files. Agent should lazy load what is needed to do focused work.
Skills have usage description metadata, but with free files you can simply instruct agent with CLAUDE.md to load them, e.g.: "Before you attempt to change any frontend code first load and follow `docs/{JS|HTML|CSS}_coding_rules.md`".
You can have all the prompts you want on top of this, but if you don't have this automated stuff running behind the scenes, you aren't serious about these issues.
Looking through some of these comments here, I see lots of people rewriting concrete rules in markdown willing to spend tokens on the hope AI won't miss it where an actual program won't.
I've tried various forms of workflows to run dedicated QA, code review (of various flavors) simplification and text simplification agents. Especially the simplification goes a long way to remove dumb padding, duplication and efficiency. Dedicated docs/comment simplification is also becoming more and more necessary on recent models. For things like feature development in my workflow, the majority of time the agents run and tokens spent is critiquing the code from various perspectives and it's not close.
Of course, this doesn't solve the overall issue that agents don't write code like you and still requires a lot of human attention in planning and code review out to clean up leftover issues, and e.g. challenge bad assumptions about architecture and real-world context. A human is still very much needed to cull the slop (or, more gratuitously: align the agent). But IME it does help avoid a lot of pitfalls and makes the code high quality a lot more quickly.
Getting away from stuff like this is exactly why I want to use AI. When I say "implement this for idle but active users," I _want_it to define isUserActiveIdle() and stuff these 4 conditionals in it. Having to check the generated code for stuff like this undoes, like .... all the benefit of using AI.
AI makes all these little decisions for us. I can about some of these decisions. I just want to notice when it's doing this without having to make my eyes bleed reading 10k lines of generated code a day.
Funny enough, discussed this yesterday
Stop Optimizing Code for Humans https://youtube.com/live/eLn4-XA-KdQ?feature=share
For personal projects, I can trust that I myself will be maintaining things so I still write things like it matters, but I do not extend the trust to others.
I'm paid to provide good quality code and not flood my company with more average code than it should.
In my previous job, I could regularly reduce a PR code down to 10-20%% of its size because someone overlooked something or was just "overengineer" a feature.
AI are such "bullshiters" that they produce more text than necessary.
Code bloat was already real, but from my personal experience it becomes realer with AI. The outcome of this will likely be apparent when no one can dive into any code base because of the amount of fluff in it (and you will obviously need more AI to deal with this).
I eventually read the library docs and managed to build a scraper for what I wanted in a few mins. Llms are great for a lot of things, but sometimes you stumble in something that's just outside of what they know/can do and you're sol. And of all the thinks, I didn't expect they would fail at this, to be honest the opposite
I've been doing this for 15 years, I love coding manually.
However, with AI-assistance I can do projects in 3 days what would take 6 months.
It's not vibe coding, everything is controlled, reviewed, understood, refined by me in the end.
But still the dev time is magnitudes faster. I would not hire anyone that is adverse to AI.
I'm actually happier. With age and a family I was getting a bit slower.
Now I have more time to spend with them AND I'm getting more done. Including personal projects I never had the bandwidth for.
The hyperbole on this keeps growing every time I see it. Soon we’ll be having people claiming they can do in 12 seconds what used to take them 17 years. What is never presented is proof. People (and programmers are no exception) are notoriously bad at estimating. We already did studies where people thought they were being faster with LLMs when they were in fact being slower.
As companies begin to rehire to fix the mess made by LLMs, it’s clear that just getting something out the door isn’t enough. It never was. Maintenance is an important part of any long-standing system.
¹ As if “moving forward” or “progress” were always a positive. It’s not. Just look at how many regulations we have to forbid or curtail uses of stuff we found to be harmful.
Gotta touch grass.
— John F. Woods (1991)
bottom up AI use seems a godsend compared to the corporate AI rat race.
i setup some slop reporting systems and ensured my boss knows theyre great starting points but serious use requires real time investment.
I’m pretty sure many people who use AI to write emails or blog posts add "make it sound like a human wrote it" to their prompts. We all know what the result usually looks like.
If AI is writing my code, I'd rather have it focus purely on correctness and efficiency than on making the code easy to read.
heck! I might even ask it to imitate Arthur Whitney’s style.
/s