If the model was trained entirely using code that Microsoft has copyright over (for example: the MS Windows codebase, the MS Office codebase, etc.) then they could offer legal assurances that they have usage rights to the generated code as derivative works.

Without such assurances, how do you know that the generated code is not subject to copyright and what license the generated code is under?

Are you comfortable risking your company's IP by unknowingly using AGPLv3-licensed code [1] that Copilot "generated" in your company's products?

[1] https://en.wikipedia.org/wiki/GNU_Affero_General_Public_Lice...

To me this whole thing is like pandora's box and it will not in any way be put back into the box. In the long run isn't arguing about the code it generates and how it generates it mostly tilting at windmills? I've already met new / junior programmers that have used copilot and chatgpt to help them see how to approach certain problems or try to get better framing for what they couldn't quite get into the most accurate words to google.

I too would prefer these tools embody the ideal: no license violation, perfect citation of where the archetypes of the code came from. I've commented here today (amongst some great FOSS software engineers) to see if a genuine respectful conversation can be had about how just like torrents this one isn't going to be put back in the box no matter how many legal precedents attempt (or succeed) in cutting off heads of the hydra. It's utility seems like it will steamroll any attempts to stop or slow it down.

Am I wrong? Is it a fools errand to ask?

We’ll hang back until other companies have litigated their way to some legislation around it.

I would prefer to see full license attributions included in generated responses, though. Something that then also wouldn't be that difficult to generate a licenses file from?

Amazon's CodeWhisperer has a "reference tracker" that tells you the license of training data code if the generated response is within some similarity threshold, but that's still not good enough imo.

The result looks like my own code and is utilizing the already existing parts of my application. The code it writes for me solves problems that you cannot find a standard solution for anywhere and is definitely not something that could be attributed.

How Copilot is trained is an issue but answering the question "where did this code come from and what license is it under" would be impossible.

I understand that some folks don't believe the same thing, and use copyleft licenses so that their code can't be re-used in a closed way, and that's fair. Github shouldn't be training their product on copyleft licenses.

It's fair to call out its misuse of certain licenses, but "the equivalent of money laundering for violation of Open Source licenses" is simply inaccurate, as many licenses allow this type of re-use explicitly.

Double-checking whether the generated part is a verbatim copy negates the speed advantage.

Possible infringements from similarity are even harder to search.

Was looking for a way to instruct CodePilot to abide by the following rules:

- Only use Apache v2, MIT or BSD licensed work for its recommendations. (Or a specific license set)

- Only use code trained on public repositories.

- Provide code attributions of the source code where the recommendations originate from.

I'm not sure if the last point is possible given these GPT type architectures but it would really help during code reviews.

That's what a good chunk of people do anyway at work. No one really cares nor will care. We were already moving in that direction anyway this will just accelerate it.

I can’t even imagine the hilarity that would ensue if I went to my GC right now to ask permission with so much in limbo; it’d be suicide by conference call.

> Copilot for Business does not retain any telemetry or Code Snippets Data.

https://docs.github.com/en/copilot/configuring-github-copilo...

It's amazing at specific things, like being z context aware boilerplate generator or doing the scaffolding of an algorithm from a comment describing it.

That's really different than using libraries.

I am using it within JS/Vue and Python and really enjoy it a lot. You can write a simple comment like "upsert the object in the store, and if the item is incomplete recursively call a refresh fn until it is complete" and it will "magically" do the rest - down to understanding where the objects I am referring to are in my store, the attribute used to decide this ('completed_at'), down to the correct syntax for updating the data in a way that plays nice with vue reactivity.

It's also a stellar autocompleter in Python-land. I have been using more and more type annotations in my codebases, but even without that, it will usually guess the right attribute or function name.

I also dig the way it will automatically write a docstring for a function. It can sometimes be a great debugging method since I can just have it comment an undocumented fn to quickly glean what it is doing. I'll circle back to enhance the docstring usually too so it is less cookie cutter.

For writing blog posts it is really neat too, because it will help me write functions to illustrate certain points based on what I am trying to teach in the post. Most of the time I can come up with the most succinct and relevant example, but sometimes I cannot and this tool does a good job helping there.

It's not perfect, but the fact that I can type a few words, hit tab, and then correct any mistakes is really a magic experience sometimes.

My company (big UK-based tech company) had an “all employees” sort of e-mail saying the use of Copilot, ChatGPT et al was not allowed for anything work-related or using company equipment due to unclear licensing model of the generated code.

I find many of these blanket rules silly, but in this case I find it is sensible to wait before polluting our products with this autogenerated code.