3 ms·

The fact that they trained on those code bases means they ingested the code. IMO they used libraries and didn't give attribution. AI doesn't actually "learn",

The fact that they trained on those code bases means they ingested the code. IMO they used libraries and didn't give attribution. AI doesn't actually "learn", it makes inferences based on the reinterpretation of data fed to it by a human. their model is a derivative work.

Dracophoenix 4y ago
I would think this point was already settled in the Google vs Oracle case. https://en.m.wikipedia.org/wiki/Google_LLC_v._Oracle_America,_Inc https://en.m.wikipedia.org/wiki/Google_LLC_v._Oracle_America.... > The inquiry into the “the purpose and character” of the use turns in large measure on whether the copying at issue was “transformative,” i.e., whether it “adds something new, with a further purpose or different character.” Campbell, 510 U. S., at 579. Google’s limited copying of the API is a transformative use. Google copied only what was needed to allow programmers to work in a different computing environment without discarding a portion of a familiar programming language. Google’s purpose was to create a different task-related system for a different computing environment (smartphones) and to create a platform—the Android platform—that would help achieve and popularize that objective. The record demonstrates numerous ways in which reimplementing an interface can further the development of computer programs. Google’s purpose was therefore consistent with that creative progress that is the basic constitutional objective of copyright itself. Pp. 24–28. Creating a model out of training data is transformative. If we apply the Google assessment to Co-Pilot, then Co-pilot is non-infringing as it is only outputting jumbled snippets of source code. The expectation would be that the programmer will decide what to edit and what to compile. The only contention I could see is if Co-pilot outputs whole source code verbatim.
- w4ffl35 4y ago
  which it does apparently https://news.ycombinator.com/item?id=27710287 https://news.ycombinator.com/item?id=27710287
  - Dracophoenix 4y ago
    A year ago, Co-pilot was in beta and wasn't production-ready. I don't think it should be held against Microsoft.
    - w4ffl35 4y ago
      what makes you think the product no longer trains on code it shouldn't be touching?
      - Dracophoenix 4y ago
        Training on open-source code isn't a problem in and of itself. As I said before, the training is likely transformative. Co-pilot operates as little more than a statistical recommendation engine. The biggest issue is the final output. I don't know if Microsoft did anything in particular to solve the issue of verbatim outputs. But whatever the case, I'm of the view that since Co-pilot is now production-ready, Microsoft may now be held legally responsible should such problems arise.

Dracophoenix 4y ago

I would think this point was already settled in the Google vs Oracle case. https://en.m.wikipedia.org/wiki/Google_LLC_v._Oracle_America,_Inc https://en.m.wikipedia.org/wiki/Google_LLC_v._Oracle_America.... > The inquiry into the “the purpose and character” of the use turns in large measure on whether the copying at issue was “transformative,” i.e., whether it “adds something new, with a further purpose or different character.” Campbell, 510 U. S., at 579. Google’s limited copying of the API is a transformative use. Google copied only what was needed to allow programmers to work in a different computing environment without discarding a portion of a familiar programming language. Google’s purpose was to create a different task-related system for a different computing environment (smartphones) and to create a platform—the Android platform—that would help achieve and popularize that objective. The record demonstrates numerous ways in which reimplementing an interface can further the development of computer programs. Google’s purpose was therefore consistent with that creative progress that is the basic constitutional objective of copyright itself. Pp. 24–28. Creating a model out of training data is transformative. If we apply the Google assessment to Co-Pilot, then Co-pilot is non-infringing as it is only outputting jumbled snippets of source code. The expectation would be that the programmer will decide what to edit and what to compile. The only contention I could see is if Co-pilot outputs whole source code verbatim.

w4ffl35 4y ago
which it does apparently https://news.ycombinator.com/item?id=27710287 https://news.ycombinator.com/item?id=27710287
- Dracophoenix 4y ago
  A year ago, Co-pilot was in beta and wasn't production-ready. I don't think it should be held against Microsoft.
  - w4ffl35 4y ago
    what makes you think the product no longer trains on code it shouldn't be touching?
    - Dracophoenix 4y ago
      Training on open-source code isn't a problem in and of itself. As I said before, the training is likely transformative. Co-pilot operates as little more than a statistical recommendation engine. The biggest issue is the final output. I don't know if Microsoft did anything in particular to solve the issue of verbatim outputs. But whatever the case, I'm of the view that since Co-pilot is now production-ready, Microsoft may now be held legally responsible should such problems arise.