In this case I would ask for smaller changes and justify every change. Have it look back upon these changes and have it ask itself are they truly justified or can it be simplified.

So I wondered, what if I built something more minimal?

I tend to work on small projects where all the code fits in the context window. So I built a harness with a single tool: the ability for an LLM to edit my files.

The core algorithm:

1. Read every file in the project directory

2. Concatenate them with the user's prompt into a single message

3. Send it to any language model, instructed to respond with a message and a list of edits — each edit being a search and replace on a specific file

4. Apply the edits

I tried getting the LLM to generate unified diffs, but it kept producing malformed ones. Search and replace turned out to be far more reliable — the model copies text verbatim from the files already in its context.

This can be extended with a few more features:

1. A CLI that lets you continue the conversation or reset it. Dumber LLMs need this so it can see the mistake they made.

2. Automatic retries — if an edit fails, re-send the current files and have the LLM fix its own search blocks.

3. File deletion. I made it so that an empty search-and-replace on an empty file deletes it.

The LLMs I tested were local ones running through the LM Studio API. Gemma 4 was pretty bad. Qwen 3.5 was significantly better.

Where is the source code? You should be able to get any coding agent to generate you one using this post.

Comments URL: https://news.ycombinator.com/item?id=47787383

Points: 4

# Comments: 0

I would be surprised if this silicon area could not be better utilized for something else

The braces being so far and tucked away make it really bad for c like languages

Those are examples of soft corruption

But you can use that to steer Meta. Explain how doing x (which also helps the community) makes them more money.

They think the higher fov is a pure win as it allows them to see more but in reality it is a tradeoff making targets in front of them smaller.

I don’t ever use it with user input, but use it often when building SPA without frameworks

You can’t run your own email server. All other large email providers will consider your self hosted emails as spam by default. It understandable why they took this stance (due to actual spam) but it is also awfully convenient it also increases their market power.

We are now at the whim of large corps even if we get a custom domain with them.