I think this is where the future of coding is. It is still useful to be a coder, the more experienced the better. But you will not write or edit a lot of lines anymore. You will organize the codebase in a way AI can handle it, make architectural decisions and organize the workflow around AI doing the actual coding.
The way I currently do this is that I wrote a small python file that I can start with
llmcode.py /path/to/repo
It then creates a prompt like this:
Look at the code files below and do the following:
{task_description}
Output all files that you need to change in full again,
including your changes. In the same format as I provide
the files below, that means each file starts with
filename: and ends with :filename
Under no circumstances output any other text, no additional
infos, no code formatting chars. Only the code in the
given format.
Here are the files:
somefile.py:
...code of somefile.py...
:somefile.py
someotherfile.py:
...code of someotherfile.py...
:someotherfile.py
assets/css/somestyles.css:
...code of somestyles.css...
:assets/css/somestyles.css
etc
I then look at the changes via "git diff".
It's quite fascinating. I often only make minor changes before accepting the "pull request" the llm made. Sometimes I have to make no changes at all.
> You will organize the codebase in a way AI can handle it, make architectural decisions and organize the workflow around AI doing the actual coding.
This might sound silly, but I feel like it has the potential of resulting in more readable code.
There have been times where I split up a 300 line function just so it’s easier to feed into an LLM. Same for extracting things into smaller files and classes that individually do more limited things, so they’re easier to change.
There have been times where I pay attention to the grouping of code blocks more or even leave a few comments along the way explaining the intent so LLM autocomplete would work better.
I also pay more attention to naming (which does sometimes end up more Java-like but is clear, even if verbose) and try to make the code simple enough to generate tests with less manual input.
Somehow when you understand the code yourself and so can your colleagues (for the most part) a lot of people won’t care that much. But when the AI tools stumble and actually start slowing you down instead of speeding you up and the readability of your code results in a more positive experience (subjectively) then suddenly it’s a no brainer.
You could have done all that for your peers instead.
I already do when it makes sense… except if you look at messy code and nobody else seems to care, there might be better things to spend your time on (some of which might involve finding an environment where people care about all of that by default).
But now, to actually improve my own productivity a lot? I’ll dig in more often, even in messy legacy code. Of course, if some convoluted LoginView breaks due to refactoring gone wrong, that is still my responsibility.
I disagree that you won’t edit lines, but I think you’re right.
At work this week I was investigating how we could auto scale our CI. I know enough Jenkins, AWS, perforce, power shell, packer, terraform, c++ to be able to do this, but having the time to implement and flesh everything out is a struggle. I asked Claude to create an AMI with our workspace preloaded on it, and a user data script that set up a perforce workspace without syncing it, all on windows, with the tools I mentioned. I had to make some small edits to the code to get it to match what I wanted but for the most part it took 2-3 days screwing around with a pretty clear concept in my head, and I had a prototype running in 30 minutes. Turns out it’s quicker to sync from perforce than it is to boot the custom AMI , but I learned that with an hour in total rather than building out more and we got to look at alternatives. That’s the future to me.
I don't understand what you disagreed with. The OP said
> edit a lot of lines
"a lot" being the keywords here.
Even just "organizing" the code requires great amounts of knowledge and intuition from prior experiences.
I am personally torn between the future of LLMs in this regard. Right now, even with Copilot, the benefit they give fundamentally depends on the coder that directs them - as you have noted.
What if that's no longer true in a couple years? How would that even be different from e.g. no code tools or website builders today? In different words will handwritten code stay valuable?
I personally enjoy coding so I can always keep doing it for entertainment, even if I am vastly surpassed by the machine eventually.
> Even just "organizing" the code requires great amounts of knowledge and intuition from prior experiences.
> I personally enjoy coding so I can always keep doing it for entertainment, even if I am vastly surpassed by the machine eventually.
I agree with both these takes, and I think they’re far more important than wondering if hand written code is valuable.
I do some DIY around the house. I can make a moderately straight cut (within tolerances for joinery use cases). A jig or circular saw makes that skill moot, but knowing I need a straight clean cut is a transferable skill. There’s also two separate skills - being able to break down and understand the problem and being able to implement the details of the problem. In trade skills we don’t expect any one person to design, analyze, estimate, build, install and decorate anything larger than a small piece of furniture and I think the same can be said of programming.
It’s similar to using libraries/framesorks - there will always be people who will write shitty apps with shitty unmaintainable code - we’ve been complaining about that since I’ve been programming. Those people are going to move on from not understanding their wix websites to not understanding their AI generated code. But it’s another tool in the belt of a professional programmer
> "But you will not write or edit a lot of lines anymore"
> "I wrote a small python file that I can start with"
Which one is it, chief?
Then you aren’t a coder, you are an organizer or manager
Many people identify themselves with being "a coder". Surely there are jobs for "coders" and will be in the future too. But not everyone writing programs today would qualify to doing the work defined as what "a coder" does.
I like to be a "builder of systems" , "solver of problems". "Organizer or manager" would also fit in that description. And then what tool you use to get stuff done is not relevant.
I'm sure a few decades ago people would say that for not fiddling with actual binary to make things work.
I disagree. If I put hook a library into a framework (e.g. laminar into rails) that doesn’t make me an organizer
But this does
> You will organize the codebase in a way AI can handle it, make architectural decisions and organize the workflow around AI doing the actual coding.
“Anyone” can implement a class that does something - the mark of a good engineer is someone who understands the context it’s going to be used in and modified , be that a one shot method, a core library function or a user facing api call that needs to be resilient against malicious inputs.
Would you be kind to share your script? Thanks!
Added some benchmarking to show how fast it is:
Here is a benchmark comparing it to [Repomix][1] serializing the Next.js project:
time yek
Executed in 5.19 secs fish external
usr time 2.85 secs 54.00 micros 2.85 secs
sys time 6.31 secs 629.00 micros 6.31 secs
time repomix
Executed in 22.24 mins fish external
usr time 21.99 mins 0.18 millis 21.99 mins
sys time 0.23 mins 1.72 millis 0.23 mins
Maybe I don't understand the usecase but I'm curious why speed matters given that LLMs are so slow (?)
22 minutes for a medium-sized repo is probably slow enough to optimize.
Adding my take to the mix, which has been working well for me: https://github.com/ClaireGSB/project-context.
It outputs both a file tree of your repo, a list of the dependancies, and a select list of files you want to include in your prompt for the LLM, in a single xml file. The first time you run it, it generates a .project-context.toml config file in your repo with all your files commented out, and you can just uncomment the ones you want written in full in the context file. I've found this helps when iterating on a specific part of the codebase - while keeping the full filetree give the LLM the broader context; I always ask the LLM to request more files if needed, as it can see the full list.
The files are not sorted by priority in the output though, curious what the impact would be / how much room for manual config to leave (might want to order differently depending on the objective of the prompt).
i guess I shouldn’t be surprised that many of us have approached this in different ways. it’s neat to see already multiple replies of the sort I’m going to make too, which is to share the approach I’ve been taking, which is to concatenate or to “summarize” the code, with particular attention on dependency resolution.
[chimeracat](https://github.com/scottvr/chimeracat)
It took the shape that it has because it started as a tool to concatenate a library i had been working on into a single ipynb file so that I didn’t need to install the library on the remote colab, thus the dependency graph was born (as was the ascii graph plotter ‘phart’ that it uses) and then as I realized this could be useful to share code with an LLM, started adding the summarization capabilities, and in some sort of meta-recursive-irony, worked with Claude to do so. :-)
I’ve put a collection of ancillary tools I use to aid in the pairing with LLM process up at https://github.com/scottvr/LLMental
This is hilarious https://github.com/scottvr/retree
There are a lot of them. I collected a list of cli tools that does this:
https://prompt.16x.engineer/cli-tools
I also built a GUI tool that does this:
Has anyone build a linter that optimizes code for an LLM?
The idea would be to make it more token efficient and (lower accidental perplexity), e.g. by renaming variable names, fixing typos and shortening comments.
It should probably run after a normal linter like black.
Good idea - couldn't the linter also be an LLM?
I have a very simple bash function for this (filecontens), including ignoring files based on gitignore & binary files etc. Piped to clipboard and done.
All these other ways seem unnecessarily complicated...
I also feel like this can be done in a few lines of shell script.
Can you share your function, please?
I am doing something similar for my gitpodcast project:
def get_important_files(self, file_tree):
# file_tree = "api/backend/main.py api.py"
# Send the prompt to Azure OpenAI for processing
response = openai.beta.chat.completions.parse(
model=self.model_name,
messages=[
{"role": "system", "content": "Can you give the list of upto 10 most important file paths in this file tree to understand code architechture and high level decisions and overall what the repository is about to include in the podcast i am creating, as a list, do not write any unknown file paths not listed below"}, # Initial system prompt
{"role": "user", "content": file_tree}
],
response_format=FileListFormat,
)
try:
response = response.choices[0].message.parsed
print(type(response), " resp ")
return response.file_list
except Exception as e:
print("Error processing file tree:", e)
return []
How does this compare to a tool like RepoPrompt?
Error: yek: SHA256 mismatch Expected: 34896ad65e8ae7c5e93d90e87f15656b67ed5b7596492863d1da80e548ba7301 Actual: 353f4f7467af25b5bceb66bb29d9591ffe8d620d17bf40f6e0e4ec16cd4bd7e7 File: /Users/... Library/Caches/Homebrew/downloads/0308e13c088cb787ece0e33a518cd211773daab9b427649303d79e27bf723e0d--yek-x86_64-apple-darwin.tar.gz To retry an incomplete download, remove the file above.
Removed & tried again this was the result. Is the SHA256 mismatch a security concern?
Oh totally forgot about homebrew installer. I'll fix it ASAP. Sorry about that.
Edit: Working on a fix here https://github.com/bodo-run/yek/pull/14
You can use the bash installer on macOS for now. You can read the installer file before executing it if you're not sure if it is safe
I have to add https://github.com/simonw/files-to-prompt as a marker guid.
I think "the part of it" is key here. For packaging a codebase, I'll select a collection of files using rg/fzf and then concatenate them into a markdown document, # headers for paths ```filetype <data>``` for the contents.
The selection of the files is key to let the LLM focus on what is important for the immediate task. I'll also give it the full file list and have the LLM request files as needed.
Like many people, I built something similar, llmctx[1], but the chunking feature seems really interesting, I have to look into that for llmctx.
One thing I have with llmctx that I think is missing in yek is a “Claude mode”, as it outputs the concatenation in a format more suitable to provide context for the Anthropic LLM.
That's neat ! I've built a transient UI to do this manually[0] within emacs, but with the context windows getting bigger ang bigger, being more systematic may be the way to go.
The priorization mentioned in the readme is especially interesting.
There is also https://repo2txt.simplebasedomain.com/local.html which doesn't require to download anything
This has some interesting ideas that I hadn’t seen in the other similar projects, especially around trying to sort files according to importance.
(I’ve been using RepoPrompt for this sort of thing lately.)
Does anyone know of a more semantically meaningful way of chunking code in a generalizable way? Token count seems like it'd leave out meaningful context, or include unrelated context.
I have an LLM summarize each file, and feed in a list of summaries categorized by theme, along with the function signatures and a tree of the directory. I only paste the full code of the files that are very important for the task.
Sorry if it's not very obvious, where does Yek fit with existing coding assistants such as Copilot or Continue.dev?
Is it purpose-built for code, or any text (e.g., Obsidian vault) would work?
This can be a piece of your own AI automation. Every task has a different need so being able to program your own AI automation is great for programmers. Any text based document works with this tool. It's rather simple, just stitching fils together with a dash of priority sorting
You can do this all in the browser: https://dropnread.io/
What is the use-case here? What is a "chunk"? It looks like it's just an arbitrary group of files, where "more important" files get put at the end. Why is that useful for LLMs? Also, I see it can chunk based on token count but... what's a token? ChatGPT? Llama?
Note, I understand why code context is important for LLMs. I don't understand what this chunking is or how it helps me get better code context.
token counting is done by crate that I'm using. I agree that not all LLMs use the same tokenizer but they are mostly similar.
Chunking is useful because in chat mode you can feed more than context max size if you feed in multiple USER messages
LLMs pay more attention to the last part of conversation/message. This is why sorting is very important. Your last sentence in a very long prompt is much more important the first.
Use case: I use this to run an "AI Loop" with Deepseek to fix bugs or implement features. The loop steers the LLM by not letting it go stray in various rabbit holes. Every prompt reiterates what the objective is. By loop I mean: Serialize repo, run test, feed test failure and repo to LLM, get a diff, apply the diff and repeat until the objective is achieved.
Got it, thanks.
> in chat mode you can feed more than context max size if you feed in multiple USER messages
Just so you know, this is false. You might be using a system that automatically deletes or summarizes older messages, which would make you feel like that, and would also indicate why you feel that the sorting is so important (It is important! But possibly not critically important).
For future work, you might be interested in seeing how tools like Aider do their "repo serializing" (they call it a repomap), which tries to be more intelligent by only including "important lines" (like function definitions but not bodies).
Thanks for this! I have the exact use-case and have been using a Python script to do this for a while.
I have a feeling theres some native unix commands that should cover this. So looking at the current scope of this tool, i think it would need more features.
Anyone has a bash script that covers this use case?
Sure, find, cat and split would do the job.
This is really fast! Serialized 50k lines in 500ms on my Mac
This looks promising. Hopefully much faster and less naive than Repomix
`tree --gitignore && cat .py && cat templates/`
why Dropbox when you can rsync, huh? ;)