hackermanai

It should be more performant if actually handwritten I guess. But just lowering the implementation like this will make it faster than many other apps seemingly for free, but not having the abstractions from e.g. C or C++ will likely make it unmaintainable long term.

Typing latency is within one frame, so on par with my editor and Sublime Text etc. So, yes, it's fast.

Obv not as polished but still cool.

> Long term prospects for Sublime Text look good. The business is healthy and there is consistent work being done by a small but dedicated team.

I am glad to hear this. (As someone new also trying to sell a text editor)

Great inspiration.

100MB is not a giant file

> it isn't acceptable for software to start pulling in other software without my consent

Yes. That is a bit shady.

If you're on Mac: https://koieditor.com/compare/zed/

Sublime Text is measured "snappier" and use less memory than Zed.

https://koieditor.com/benchmarks/

There’s also Koi Editor.

No telemetry, no tracking, and much faster than Zed: https://koieditor.com/benchmarks/

Dev solely focused on performance and bug fixing.

So it seems.

> Claude one shots pretty much everything

What?

> An IDE at this point is best as a tool for code review.

I heard from a friend that most devs building serious stuff still write code. It's shocking but true. (No code review needed.)

At least it didn't succeed yet. They should drop a model somewhere, beating something else in some use case, and maybe people would use.

> You just say, "Go Build me x - it should have y,z features - and build it in golang for me" - and that's it - the 3-4 page Plan comes back - usually pretty credible - and then you click "build.".

What you're describing seems like a workflow for building toys only. There's currently no reality in which someone would actually know what the y,z features are before making them. A plan generated in 5min would likely suggest a suboptimal solution compared to what a good solution would look like (which might take a year or two to figure out, for a human, so still a week or so for SOTA models if at all possible). Building something in golang is cute, but hard to be convinced until more novel applications are being generated from prompts.

The data submitted by Cursor's users tho, that seems to be very valuable.

I think this is a cool approach, like a giant undo stack.

I have this line completion feature in koieditor.com as well, and it's hard to suggest "safe"/good completions at a low latency. Best approach I could think of is a second pass to verify first pass, but adds to latency, or change to better model, which often also impacts latency.

Many are pro AI, anti hype.

I think this approach is more common than the hype for actual work. I do something similar, many back and forth, then settle on something often with now known tradeoffs, written by hand to spot issues as a final guard/ keep consistent naming etc.

Not a textedit fork, but I am building a minimal text editor with niceties based on scintilla. So performance wise on par with Notepad++ et al.

> Usually performance was the reason for using native APIs rather than web views, but this doesn't seem to be true any more.

It's still true. There's no way around it, web views will always be slower.

Hackerman Text editor.

595 days and counting.

ᕙ(⇀‸↼‶)ᕗ

> actively reject multiple safety warnings

Is this like a popup? which most people actively accept without blinking

I think plugin/extensions should be a bit harder to run by default. I get the user friction from extra hurdles before using their plugins etc., but I don't think there is an actually safe way to execute arbitrary code, unaudited, without sandboxing, or other restrictions.

> “But Local Models Aren’t As Smart”

This is what makes me continuously doubt and rewrite the local-first approach to inline chat in my editor. Next edit/ code complete makes more sense due to latency advantage. But chat is hard.

It's fast and feels good to run locally, but output quality is just not ChatGPT etal.

> How does anyone at Microsoft/Github thinks this is ok?

Their whole marketing department probably.

> .. webview ..

That would make it an web app tho. Not sure if that's what OP wants when making Swift apps.

I recently tested input latency on Mac Mini M2 Pro, keypress to visible pixel (p95 value).

- Zed 36.8ms (26.3 avg) - Sublime 33.5ms (25.9 avg) - VSCode 31.7ms (20.6 avg) - Hackerman Text 15.3ms (12.5 avg)

All editors tested as stripped down as possible, minimal UI etc., empty file, no syntax highlighting. Mostly surprised by VSCode tbh.

Testing on sqlite3.c (with syntax highlighting): Zed 60.5ms, Sublime 37.1ms, VSCode 22.0ms, and Hackerman Text 18.4ms.

And soon™, https://hackerman.ai/

> How else are you going to have “a quick glance at code” across project files without using a new view for that?

By showing the text around match inline with the search result in the tree. Especially useful if you do not expect to edit every search result. If you do expect to edit every search result, then Zed's multibuffer is arguably better/ faster.