Hacker News Reader

Comment by bonzini | Hacker News Reader

It was, but according to Wikipedia they sorted that out by moving the proprietary neural network out of the executable. If this is correct, it means that the source is available and the nature of the changes (including removing the original authors) should be visible.

View on HN

https://github.com/official-stockfish/Stockfish/blob/master/...

Stockfish _is_ currently licensed under GPLv3. Chessbase is allowed to sell a modified copy, but only under the same license and they have to provide the source.

The neural network is just data, but if it's part of the executable you need to provide the raw data that is then linked into the executable.

I think it can be argued that the set of weights of a neural network is a program, but not “the preferred form of the work for making modifications to it”.

That would mean the training set and training algorithm (and random number generator seed) would have to be distributed under the GPL, too.

I don't think it is always feasible to make the learning process entirely deterministic, especially for distributed reinforcement learning.

Binaries weren't historically deterministic either. This isn't a problem.

They did the same thing though. Two differently trained neural networks will have different outcomes, in this case they'd pick different chess moves in some special cases.

I’m fairly sure there were (maybe still are) linkers that, given duplicate symbol names, would happily grab one of them without checking whether they had the same code.

Add in parallel linking, and program behavior can change.

Reply on news.ycombinator.com