[IgorPartola]

I think this is the correct answer. IANAL but the copilot code vs the copilot training data are different things and licensing for one shouldn’t affect the other, right? And the fact that training data happens to also be code is incidental.

[8note]

One view would be that copilot the app distributes GPL'd code, in a weird encoding. Training the model is a compilation step to that encoding

[keonix]

I assume the code is a derivative work of training data because given different data code would be also different (neuron weights)

[IgorPartola]

If I read a GPL implementation of a linked list and then write my own linked list implementation, was my neural network in my brain a derivative work of the GPL code?

[keonix]

Sure it is, you brain is not software though

[IgorPartola]

So as long as I read GPL code, then rewrite it from memory and feed it to copilot to train it I can unGPL anything?