xmorse

it probably because the author uses the useless implementation of the Zed agent

This is an API I built for my agents to let them read files in public GitHub repos without relying on the official GitHub API.

There are no rate limits. The API is built on top of sqlite on Cloudflare Durable Objects.

The first time you fetch an url it downloads the repository tar file and stores the files contents in the sqlite database.

It then uses sqlite to list, read and search the files.

Agents can paginate files using search params like start and end to only read a portion of a file. They can also use globs to only search or read a portion of a repo

I am not. No one ever heard of their managed postgres and they already started calling it with an acronym

The answer is an open secret

This post is just ChatGPT slop

It's because these industries don't create their own benchmarks. The only ones creating evals are the AI company themselves or open source software engineers

This is a good time to implement "context", a way to pass down the call stack parameters instead of having to add a io argument to every function

You replied to an AI generated text, didn't you notice?

I would like to have some kind of interfaces instead of having to always use anytype, so compiler errors will appear in the call site instead of the called function body

Next time you try Rosetta and it fails can you open an issue with the error messages? I think this task would be perfect for o3 to work on, you basically have to browse hundreds of issues on GitHub and find other people that got into the same problem

Is this basically the same thing as Orbstack?

that's why i added fast at runtime too

x86 runs fine in Apple Silicon too thanks to Rosetta

Will macOS also get support for the self hosted back end at some point?

that issue is ridiculous, what did you expect randomly increasing the pointer of an array?

This could be the greatest programming language development of the last 10 years. Finally a language that compiles fast and is fast at runtime too.

I still come back to Cursor to read diffs, I really need a split editor diff view to understand the changes

I love fly but you really have to work harder on these things, the landing page is not something you need a marketing team for

The landing page shows all logos of small companies, including one that is migrating away from them (Turso)

https://fly.io/

what about functions that return a callback for disposal? like Redux store.subscribe, there are thousands more examples

Probably not

Why didn't they go with `defer` which is the more ergonomic and established pattern?

`defer` also doesn't need to change the API of thousands of objects to use it, instead now you have to add a method any resource like object, or for things that are not objects, you can't even use this feature.

A Fresh update

I am using Zed too, it still has some issues but it is comparable to Cursor. In my opinion they iterate even faster than the VSCode forks.

It still works

Running models locally is very expensive in terms of memory and scheduling requirements, maybe instead they should host their model on the Cloudflare AI network which is distributed all around the world and can have lower latency

does it really matter if it takes 1 hour to compile?

does it? a program would be able to spawn another program only with the exports available to him

I would love having gpui bingings for JS using NAPI, now using it from Rust is too slow, Rust feedback loop is too long

Basically same level as Cursor, one thing missing is the Apply Cursor model which sometimes makes better edits in the code