Hacker News new | ask | show | jobs
by TeMPOraL 1812 days ago
Honestly, I feel most people don't care about that. What they do care about, is the risk of Copilot making the user liable for copyright infringement. Even a possibility of it spewing out non-public-domain code should be considered a showstopper for any use of Copilot-generated code in a commercial project.

Can Copilot produce licensed code verbatim, in enough quantities to matter, with a license your business would be infringing? Yes. Can you easily tell by looking at the output? No. Could someone end up suing you over it? Maybe, if they cared enough to find out. Can you honestly tell your investors, or a company you seek to be acquired by, that nobody else can have valid copyright claim against your code? No.
1 comments
emodendroket 1812 days ago
> Can Copilot produce licensed code verbatim, in enough quantities to matter, with a license your business would be infringing? Yes. Can you easily tell by looking at the output? No. Could someone end up suing you over it? Maybe, if they cared enough to find out. Can you honestly tell your investors, or a company you seek to be acquired by, that nobody else can have valid copyright claim against your code? No.

Well aren't all your assertions exactly the point of contention?

link
TeMPOraL 1812 days ago
Well, the "enough quantities to matter" part wasn't tested in courts yet, but I fail to see a way to rule for "No" here in a way that wouldn't gift us an universal way to turn any code into public domain, destroying source code licensing as a concept. Other than this part, the first two claims have already been demonstrated, and the rest follow from them.
link
emodendroket 1812 days ago
But that is in fact the most fundamental question here. And I’m not fully sold on the idea either that this is going to happen in real-world usage or that a single function in a massive program constitutes a large enough portion to be infringing.
link
TeMPOraL 1812 days ago
Quake's square root function wasn't the only, or the largest, example of code Copilot reproduces verbatim. Among others I've seen to date is someone generating a real "About" page with PII information of some random software developer.

How much code is enough to infringe is a tricky question, though. It's not only a function of size, but also of importance/uniqueness - and we know that Copilot doesn't understand these concepts.

link
shakna 1812 days ago
> ... or that a single function in a massive program constitutes a large enough portion to be infringing.

As part of the sequences of rulings in Google vs Oracle, the 9-line rangeCheck function, in the entirety of the Android codebase, was found to be infringing.

link