Live data from Hacker News

GitHub Copilot is now available for free

github.com

241–250 of 412 posts

Re: GitHub Copilot is now available for free

#241

Earlier quoted context omitted.

You're the second person in this thread to make this point, what are you using it for? I find the difference is basically negligible (in the sense that both get the busywork right and both fail at anything complicated)

yeah, Sonnet goes past that. 300+ line changes in 20 seconds. You have to review it, but generally it's right. It's infinitely faster than the time to look at docs and do it myself. Sure it's busywork. But it's a lot of busywork very fast.

Well it's definitely not infinitely faster since you're having to review it, but we're talking about the delta between Sonnet and Qwen/Mistral/Llama or whatever, not doing it manually.

I'm really curious what your problem domain is, like specifically what sort of code are you asking it to change and what changes are you asking for.

I just gave o1 and Sonnet a total layup question (optimization that had a huge win simply by filtering an array before sorting it vs the other way around) and neither model got the solution right, both of them came up with ~hundred lines of code, neither model's code worked on the first try. It took me like 10 minutes to refactor and optimize the code for a 6x speedup and it would take longer than that to debug the AI code to even make it run. (I spent 10 minutes prompting/editing to try to get the generated solutions to run)

Also the initial code was 11 sloc, my solution is 14 sloc, and claude was 70 sloc and o1 was 93. idfk, i just don't think we're there yet

Re: GitHub Copilot is now available for free

#243

Microsoft would need to pay me to use Copilot. Seems like a major scam for them to learn from our code and then tell us we can’t use it to make our own competing AI systems. “Limits on use of data from the AI Services. You may not use the AI services, or data from the AI services, to create, train, or improve (directly or indirectly) any other AI service.” From https://www.microsoft.com/en/servicesagreement#13r_AISer…

It's a good example of how ridiculous the AI training situation is. They claim it's fair use for them to steal all data they want, but you're not allowed to use AI data output, despite this data literally not being subject to copyright protections on account of lacking a human author. And especially Github. They already have an enormous corpus that is licensed under MIT/equivalent licenses, explicitly permitting them…

I wonder how GPLv3 and CC BY SA licenses should be considered when training AIs like this? The model is software, and if it's sufficiently different from the source, it's a derivative work, isn't it?

Re: GitHub Copilot is now available for free

#244

Microsoft would need to pay me to use Copilot. Seems like a major scam for them to learn from our code and then tell us we can’t use it to make our own competing AI systems. “Limits on use of data from the AI Services. You may not use the AI services, or data from the AI services, to create, train, or improve (directly or indirectly) any other AI service.” From https://www.microsoft.com/en/servicesagreement#13r_AISer…

psssst *it's a secret If You're Not Paying For It, You Become The Product (2012) https://www.forbes.com/sites/marketshare/2012/03/05/if-youre...

These days you will pay for it and still become the product :)

Re: GitHub Copilot is now available for free

#245

The real cost is that society at large is no longer contributing to the StackOverflow, so problems and solutions are all now stored in proprietary databases (which granted SO also was) but also now stored in an invisible proprietary database

I think you'd find society-at- small was contributing , with perhaps 10x larger yet still quite small number posting but watering down useful contributions, 100x that lurking, and 1000x that just drive-by copy-pasting from SO to their IDE.

"Useful contributions" is subjective. Not everyone is born a senior developer. Juniors, and even children who aren't even juniors yet ask questions on these channels.

Source: I bothered a lot of people on the Internet about C++ when I was child.

Re: GitHub Copilot is now available for free

#246

Earlier quoted context omitted.

A software developer's time is much more precious than wasting time on sub-optimal models. Open Weights models has it's place (in training custom agents and custom services), but if you are knowledge worker, using a model even 5% less than SOTA is extremely dumb

100% disagree with this take, the flexibility in controlling the prompt leads to QwenCoder2.5-32b outperforming gpt-o1 and claude sonnet 3.5 for nearly everything that I use it for (true for Gemma-27b and llama3.3-70b, though in this context I'm almost always using the former). A specialist model that's specifically prompted to do the correct thing will outperform a SOTA generic model with a one size fits all system…

Do you have any example prompts or suggestions for coming up with them?

Re: GitHub Copilot is now available for free

#247

Earlier quoted context omitted.

This. 1000x This. I watch people navigate code in VSCode and I want to pull my hair out. Things that I don’t even think about are hard and/or require just falling back to search. And before “there is a plugin for that”, I’m sure there is. I’m sure you can configure VSCode to be just as powerful as IDEA but the rank and file using it aren’t doing that work to install and configure a bunch of plugins. So, on average, V…

>I watch people navigate code in VSCode and I want to pull my hair out. For me it's the other way around, when I see someone using an IDE instead of a lean editor I see their struggle. Multiple seconds to open the IDE (sometimes tens of second), multi-hundred millisecond lag when opening a file, noticeable input lag. And when you have to edit a file your IDE doesn't understand, all you have is a bloated notepad. I kn…

> sometimes few different projects a day - and read files in dozens of programming language

+1 this is what brought me back to vscode after experimenting with goland. To me vscode better handles the heterogeneity of my daily work. In my workspace I can keep open: a golang codebase, a massive codebase consisting of yaml config files, a filesystem from a remote ssh connection, a directory of personal markdown notes, and directories of debug logs. In my experience jetbrains excelled at the single use case, but vscode won on its diversity.

I will say that the parent comment had me curious about goland again. But I suspect I really need to spend more time configuring my vscode setup. I spent years using emacs, and would love to have a helm-like navigation experience.

Re: GitHub Copilot is now available for free

#248

The real cost is that society at large is no longer contributing to the StackOverflow, so problems and solutions are all now stored in proprietary databases (which granted SO also was) but also now stored in an invisible proprietary database

I wonder if a purge and a fresh start for StackOverflow would renew interest.

I used to like reading StackExchange sites as a social media site--lots of interesting questions and clever answers. Today, votes have slowed down and the best answers are from 2017, and only niche questions can avoid being closed.

Re: GitHub Copilot is now available for free

#249
post #193
post #85

Earlier quoted context omitted.

Unfortunately sourcehut has its own cultural issues.

Maybe I’m out of the loop, what happened there?

I interpreted "culture" in two ways, one technical and one administrative

Technical: https://news.ycombinator.com/item?id=23038520 with the tl;dr of "patches over email is king, if you want fancy web stuff go elsewhere"

Administrative: https://sourcehut.org/blog/2022-10-31-tos-update-cryptocurre... https://news.ycombinator.com/item?id=33403780

Re: GitHub Copilot is now available for free

#250

Earlier quoted context omitted.

It's a good example of how ridiculous the AI training situation is. They claim it's fair use for them to steal all data they want, but you're not allowed to use AI data output, despite this data literally not being subject to copyright protections on account of lacking a human author. And especially Github. They already have an enormous corpus that is licensed under MIT/equivalent licenses, explicitly permitting them…

I wonder how GPLv3 and CC BY SA licenses should be considered when training AIs like this? The model is software, and if it's sufficiently different from the source, it's a derivative work, isn't it?

> it's a derivative work, isn't it

Short answer: unlikely

Serious answer: we'll only know whether it is when someone challenge it at court.

Post reply on HN