Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

GitHub Copilot Workspace: Technical Preview

github.blog

Post reply on HN
Print view View original

301–310 of 333 posts

Re: GitHub Copilot Workspace: Technical Preview

#301

Post by _lvbh »

freedomben wrote: post #256
_lvbh wrote: post #255

Earlier quoted context omitted.

There are lots of hackers bringing VsCode-only extensions to Neovim. I brought Copilot Chat to neovim back when it first became available to me and the community is still alive and well, though with a different maintainer. Tbh AI tools aren’t that useful, and those that are will get ported. Currently trying to bring Cursor to Neovim but that has proven much more difficult

I don't know that I can say thank you enough for your work! Thank you!!! Do you know why Github didn't make an official version? I'm definitely going to give yours a try, and I don't really care whether it's "official" or not, but rather I'm curious because looking for official neovim support is a useful signal of their priorities/intentions.

Link btw since it’s on a org: https://github.com/CopilotC-Nvim/CopilotChat.nvim

Re: official version

There was some discussion but the answer is that it’s just too difficult. Our version was reversed from a bunch of MITM and guesswork. The VSCode implementation has a “@workspace” command which involves complex tree-sitter integration, sending off all your code to be vectorized, and uses RAG to get the most relevant snippets. We obviously weren’t able to implement all features. They can get away with this in JavaScript because you can get just about anything via npm. In Lua, most things have to be from scratch. I had to spend a few hours just getting tiktoken working in Lua and even then it requires manual installation. The package management system with Lazy.nvim is very lacking.

Re: GitHub Copilot Workspace: Technical Preview

#302

Post by halfmatthalfcat »

Arainach wrote: post #287
halfmatthalfcat wrote: post #270

Earlier quoted context omitted.

People always say this but how else are you going to learn? I doubt many of us who are "rolling our own crypto" are actually deploying it into mission critical contexts anyway.

Asking an LLM to do something for you doesn't involve any learning at all.

I’m not talking about the LLM case, just the mantra of “don’t roll your own crypto” constantly. Comes off as unnecessarily gatekeepy.

Re: GitHub Copilot Workspace: Technical Preview

#303

Post by rsynnott »

halfmatthalfcat wrote: post #302
Arainach wrote: post #287

Earlier quoted context omitted.

Asking an LLM to do something for you doesn't involve any learning at all.

I’m not talking about the LLM case, just the mantra of “don’t roll your own crypto” constantly. Comes off as unnecessarily gatekeepy.

I mean, by that, people don't generally mean, literally, "never write your own crypto". They just mean "on no account _use_ self-written crypto for anything".

Re: GitHub Copilot Workspace: Technical Preview

#304

Post by rsynnott »

krainboltgreene wrote: post #198
Vuizur wrote: post #163

Earlier quoted context omitted.

>executing large-scale changes in entire repositories in 3 years You can look at SWE-Agent, it solved 12 percent of the GitHub issues of their test dataset. It probably depends on your definition of large-scale. This will get much better, it is a new problem with lots of unexplored details, and we will likely get GPT-5 this year, which is supposed to be a similar jump in performance as from 3.5 to 4 according to Altm…

This is a laughable definition of large-scale. It's also a misrepresentation of that situation: It was 12% of issues in a dataset for the top 5000 repositories pypy packages. Further "solves" is a incredibly generous definition, so I'm assuming you didn't read the source or any of the attempts to use this service. Here's one where it deletes half the code and replaces network handling with a comment to handle network…

> Here's one where it deletes half the code and replaces network handling with a comment to handle network handling

... Wait, that's not one that they considered a _success_, is it? Like, one of the 12%?

Re: GitHub Copilot Workspace: Technical Preview

#305

Post by Zambyte »

BoorishBears wrote: post #257
Zambyte wrote: post #199

Earlier quoted context omitted.

I really don't think they're being over sold that much. I'm running llama 3 8b on my machine, and it feels a lot like running claude 3 haiku with a much lower context window. Quality wise it is surprisingly nice.

Llama 3 just came out so they couldn't have used it, and Claude Haiku is the smallest cheapest closed source model out there from what I've seen. Github is likely using a GPT-4 class model which is two (massive) steps up in capabilities in Anthropic's offerings alone

Yeah I just mentioned Llama to point out that the open weight models have been really catching up.

Microsoft is almost certainly using GPT-4 given their relationship with ClosedAI, but I would definitely not put GPT-4 (nor Turbo) "two massive steps up" from Claude 3 Opus. I have access to both through Kagi, and I have found myself favoring the responses of Claude to the point where I almost never use GPT(TM) anymore.

Re: GitHub Copilot Workspace: Technical Preview

#306

Post by mountainriver »

singluere wrote: post #112

While I've not used this product, I've created somewhat similar setup using open source LLMs that runs locally. After having used it for about three months, I can say that debugging LLM prompts was far more annoying than debugging code. Ultimately, I ended up abandoning my setup and going in favor of writing code the good old fashioned way. YMMV

Maybe they know something about GPT5

Re: GitHub Copilot Workspace: Technical Preview

#307

Post by freedomben »

_lvbh wrote: post #301
freedomben wrote: post #256

Earlier quoted context omitted.

I don't know that I can say thank you enough for your work! Thank you!!! Do you know why Github didn't make an official version? I'm definitely going to give yours a try, and I don't really care whether it's "official" or not, but rather I'm curious because looking for official neovim support is a useful signal of their priorities/intentions.

Link btw since it’s on a org: https://github.com/CopilotC-Nvim/CopilotChat.nvim Re: official version There was some discussion but the answer is that it’s just too difficult. Our version was reversed from a bunch of MITM and guesswork. The VSCode implementation has a “@workspace” command which involves complex tree-sitter integration, sending off all your code to be vectorized, and uses RAG to get the most relevant s…

Thank you! I'll be trying this out in the next few days as soon as I get some time. I much appreciate your effort. I know you don't do it for the money, but I'll throw some sponsor bucks your way. I'd love to buy you a few beers (or coffees or whatever your love is) :-)

Re: GitHub Copilot Workspace: Technical Preview

#308

Post by freedomben »

packetlost wrote: post #278
freedomben wrote: post #240

Earlier quoted context omitted.

Interesting, latency has always been great for me. If it's going to work, it usually has suggestions within a second or two. I use the neovim plugin though so not on the typical VS-code based path.

I also used the neovim plugin. I'm on a fiber connection in the Midwest, so that is likely a factor. Latency was on the order of 2-5s consistently, which is way more than enough to interrupt my flow.

Interesting indeed! Do you normally experience high latencies >1s from your connection or is copilot more of an outlier? I have noticed that when I travel to the midwest I will get latencies around 70 to 90 ms rather than my current 30 ms, but It's not something I really notice too much, though that tends to be in major cities.

Re: GitHub Copilot Workspace: Technical Preview

#310

Post by krainboltgreene »

rsynnott wrote: post #304
krainboltgreene wrote: post #198

Earlier quoted context omitted.

This is a laughable definition of large-scale. It's also a misrepresentation of that situation: It was 12% of issues in a dataset for the top 5000 repositories pypy packages. Further "solves" is a incredibly generous definition, so I'm assuming you didn't read the source or any of the attempts to use this service. Here's one where it deletes half the code and replaces network handling with a comment to handle network…

> Here's one where it deletes half the code and replaces network handling with a comment to handle network handling ... Wait, that's not one that they considered a _success_, is it? Like, one of the 12%?

We unfortunately have no idea what they consider a success! That's just one of the most recent ones by some random user who wanted to use the program in the real world.
Post reply on HN