5 ms·

Does that make the code uncopyrightable? Non-human authorship?

by Animats 6mo ago

redwall_hp 6mo ago
The courts have determined that, yes, and that is the position of the Copyright Office. And the Supreme Court has rejected appeal, so that's the standing precedent. Realistically, look forward to SOX style audits and having to maintain evidence of how much of a code base has human authorship vs machine generation. Or reject slop. I can't wait for: * The first company to do perjury for litigating over a nonexistent copyright for machine generated code. * The first company to get nailed to the wall for reverse engineering and replicating high profile copyrighted code, like Windows.
- circuit10 6mo ago
  Having a tool involved isn't the same as being entirely generated by a tool For example, without any AI, if I generate a lookup table for the sine function in my code, that table may not be copyrightable because it was machine-generated, but it doesn't somehow make the rest of the code not copyrightable either "Co-authored by" doesn't imply it was entirely machine-generated
Dylan16807 6mo ago
If it's actually co authored then you should be fine on copyright. And of course dumb messages that aren't true won't affect copyright.
- lelanthran 6mo ago
  > If it's actually co authored then you should be fine on copyright How so? All your outoutput is now legally partly owned by Microsoft?
  - Dylan16807 6mo ago
    The bot (and therefore microsoft) doesn't get any copyright at all.
    - lelanthran 6mo ago
      > The bot (and therefore microsoft) doesn't get any copyright at all. But then neither do you, for every commit that was marked with copilot.
      - Dylan16807 6mo ago
        What makes you say that? If a monkey uses a typewriter, there's no copyright. If I use a typewriter with a monkey, I get copyright and the monkey doesn't. Why would the monkey need copyright for me to get copyright?
        lelanthran 6mo ago
        > If I use a typewriter with a monkey, I get copyright and the monkey doesn't. Right, because monkeys cannot be granted copyright. If you use a typewriter along with Microsoft, the resulting copyright will be owned jointly. This story isn't about a monkey claiming co-authorship, it's about Microsoft claiming co-authorship.
        Dylan16807 6mo ago
        > This story isn't about a monkey claiming co-authorship, it's about Microsoft claiming co-authorship. I don't understand your position. Your previous post was agreeing that Microsoft wouldn't get copyright on copilot output, wasn't it? I said the bot doesn't get copyright and you said "neither do you". Why are you now saying Microsoft would get copyright? Copilot is a robo-monkey owned by Microsoft.
  - tardedmeme 6mo ago
    No. Legal ownership doesn't depend on whether aislop edited your commit message.
  - baobabKoodaa 6mo ago
    If I write "I own lelanthran's car", does that make me the legal owner of your car? No?
    - lelanthran 6mo ago
      > If I write "I own lelanthran's car", does that make me the legal owner of your car? If I counter sign agreement, certainly. How do you think that sales of both movable and immovable property work?
      - baobabKoodaa 6mo ago
        That's not what we were talking about. We were talking about a third party modifying your document without your consent (and sometimes even without your knowledge). You write git commit "Fix bug" and then a third party swoops in the night and modifies that with "Co-authored by: Microsoft".
        lelanthran 6mo ago
        > That's not what we were talking about. We were talking about a third party modifying your document without your consent (and sometimes even without your knowledge). You write git commit "Fix bug" and then a third party swoops in the night and modifies that with "Co-authored by: Microsoft". Right, and how is a court supposed to differentiate between the cases when copilot was not sharing your typewriter and cases when it was?
        baobabKoodaa 6mo ago
        I don't know! How is the court supposed to differentiate between a forged document and an authentic one? A question for the ages.
cookiengineer 6mo ago
The real question is why Anthropic was able to use DMCA takedown requests "in good faith" against the Claude leaks when their own CTO claimed it is a 100% slopcoded codebase, and they themselves argue that all LLM generated code is transformed enough to not be copyrightable. Which they have to state without being able to turn back because they violated millions of book and software licenses during training. Make it make sense.
- naruhodo 6mo ago
  Truth, law and consequences (for the capital class) are so last year.
- ashirviskas 6mo ago
  What? Training is not inference. Reading books is not the same as writing.
  - cookiengineer 6mo ago
    Maybe read up on how transformers, their encoders and decoders, and the attention matrix works? https://arxiv.org/abs/1706.03762 https://arxiv.org/abs/1706.03762
- tardedmeme 6mo ago
  You can lie. You can get away with lying. You can lie to judges. You can get away with lying to judges. You can profit from getting away with lying to judges. A judge isn't involved, anyway. The leaker would have to take you to court and then prove that your request was in bad faith and that they didn't infringe copyright. Competent programmers understand how to tell the computer what needs to happen. Really good programmers understand how the computer executed the code, and take advantage of it - they know about speculative execution and cache prefetching. Competent lawyers know what the law says. Really good lawyers understand how the law is executed, and take advantage of it - they know when it won't be enforced.

Does that make the code uncopyrightable? Non-human authorship?

by Animats 6mo ago

Does that make the code uncopyrightable? Non-human authorship?

The courts have determined that, yes, and that is the position of the Copyright Office. And the Supreme Court has rejected appeal, so that's the standing precedent. Realistically, look forward to SOX style audits and having to maintain evidence of how much of a code base has human authorship vs machine generation. Or reject slop. I can't wait for: * The first company to do perjury for litigating over a nonexistent copyright for machine generated code. * The first company to get nailed to the wall for reverse engineering and replicating high profile copyrighted code, like Windows.

circuit10 6mo ago
Having a tool involved isn't the same as being entirely generated by a tool For example, without any AI, if I generate a lookup table for the sine function in my code, that table may not be copyrightable because it was machine-generated, but it doesn't somehow make the rest of the code not copyrightable either "Co-authored by" doesn't imply it was entirely machine-generated

Dylan16807 6mo ago

If it's actually co authored then you should be fine on copyright. And of course dumb messages that aren't true won't affect copyright.

lelanthran 6mo ago
> If it's actually co authored then you should be fine on copyright How so? All your outoutput is now legally partly owned by Microsoft?
- Dylan16807 6mo ago
  The bot (and therefore microsoft) doesn't get any copyright at all.
  - lelanthran 6mo ago
    > The bot (and therefore microsoft) doesn't get any copyright at all. But then neither do you, for every commit that was marked with copilot.
    - Dylan16807 6mo ago
      What makes you say that? If a monkey uses a typewriter, there's no copyright. If I use a typewriter with a monkey, I get copyright and the monkey doesn't. Why would the monkey need copyright for me to get copyright?
      - lelanthran 6mo ago
        > If I use a typewriter with a monkey, I get copyright and the monkey doesn't. Right, because monkeys cannot be granted copyright. If you use a typewriter along with Microsoft, the resulting copyright will be owned jointly. This story isn't about a monkey claiming co-authorship, it's about Microsoft claiming co-authorship.
        Dylan16807 6mo ago
        > This story isn't about a monkey claiming co-authorship, it's about Microsoft claiming co-authorship. I don't understand your position. Your previous post was agreeing that Microsoft wouldn't get copyright on copilot output, wasn't it? I said the bot doesn't get copyright and you said "neither do you". Why are you now saying Microsoft would get copyright? Copilot is a robo-monkey owned by Microsoft.
- tardedmeme 6mo ago
  No. Legal ownership doesn't depend on whether aislop edited your commit message.
- baobabKoodaa 6mo ago
  If I write "I own lelanthran's car", does that make me the legal owner of your car? No?
  - lelanthran 6mo ago
    > If I write "I own lelanthran's car", does that make me the legal owner of your car? If I counter sign agreement, certainly. How do you think that sales of both movable and immovable property work?
    - baobabKoodaa 6mo ago
      That's not what we were talking about. We were talking about a third party modifying your document without your consent (and sometimes even without your knowledge). You write git commit "Fix bug" and then a third party swoops in the night and modifies that with "Co-authored by: Microsoft".
      - lelanthran 6mo ago
        > That's not what we were talking about. We were talking about a third party modifying your document without your consent (and sometimes even without your knowledge). You write git commit "Fix bug" and then a third party swoops in the night and modifies that with "Co-authored by: Microsoft". Right, and how is a court supposed to differentiate between the cases when copilot was not sharing your typewriter and cases when it was?
        baobabKoodaa 6mo ago
        I don't know! How is the court supposed to differentiate between a forged document and an authentic one? A question for the ages.

cookiengineer 6mo ago

The real question is why Anthropic was able to use DMCA takedown requests "in good faith" against the Claude leaks when their own CTO claimed it is a 100% slopcoded codebase, and they themselves argue that all LLM generated code is transformed enough to not be copyrightable. Which they have to state without being able to turn back because they violated millions of book and software licenses during training. Make it make sense.

naruhodo 6mo ago
Truth, law and consequences (for the capital class) are so last year.
ashirviskas 6mo ago
What? Training is not inference. Reading books is not the same as writing.
- cookiengineer 6mo ago
  Maybe read up on how transformers, their encoders and decoders, and the attention matrix works? https://arxiv.org/abs/1706.03762 https://arxiv.org/abs/1706.03762
tardedmeme 6mo ago
You can lie. You can get away with lying. You can lie to judges. You can get away with lying to judges. You can profit from getting away with lying to judges. A judge isn't involved, anyway. The leaker would have to take you to court and then prove that your request was in bad faith and that they didn't infringe copyright. Competent programmers understand how to tell the computer what needs to happen. Really good programmers understand how the computer executed the code, and take advantage of it - they know about speculative execution and cache prefetching. Competent lawyers know what the law says. Really good lawyers understand how the law is executed, and take advantage of it - they know when it won't be enforced.