4 ms·

It’s not about the learning part, it’s about the copyright and money part. If you learn how to play song X by The Rolling Stones, you cannot just make money pla

• albertzeyer 4y ago
  But Copilot (or a human) is not copying some existing code. It (Copilot or human) just used existing code to learn. So it is only about the learning part, nothing else. Yes, if Copilot (or a human) would copy existing code, that would be a copyright violation. But none of the arguments here are about that. It's just about the learning.
  • BrandonJung 4y ago
    it is possible that on a long snippet that you could be copying code from non-permissive code. In addition GitHub themselves have said that this is not resolved and will likely need to be adjudicated in the courts. The only way to avoid this is to train on only fully permissive code. We know here at Tabnine having done this for 5+ years.
    • albertzeyer 4y ago
      And again, the same applies for a human as well. It's just the same. I'm pretty sure I have done that by accident already in the past without noticing. This is not so unusual when you write some code with very common patterns. And then this even can apply for code which you have not seen before. E.g. write some bubble sort function. Very likely you will find exactly the same code online.