Technology

58108 readers

3888 users here now

This is a most excellent place for technology news and articles.

Our Rules

Follow the lemmy.world rules.
Only tech related content.
Be excellent to each another!
Mod approved content bots can post up to 10 articles per day.
Threads asking for personal tech support may be deleted.
Politics threads may be removed.
No memes allowed as posts, OK to post as comments.
Only approved bots from the list below, to ask if your bot can be added please contact us.
Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago

MODERATORS

[email protected]

165

Why The New York Times might win its copyright lawsuit against OpenAI (arstechnica.com)

submitted 7 months ago by [email protected] to c/[email protected]

36 comments fedilink hide all child comments

Why The New York Times might win its copyright lawsuit against OpenAI::The AI community needs to take copyright lawsuits seriously.

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 4 points 7 months ago (2 children)

It looks like someone hasn't seen the video of Copilot spitting out the Quake inverse sqrt algorithm verbatim.

[–] [email protected] 0 points 6 months ago (1 children)

While it got popularised as "Carmack's reverse" the algorithm is actually significantly older.

Also you'd have to show that it was literally copy+pasted, including comments and all, to even have a chance at a copyright claim: Algorithms are not subject to copyright, similar to how story structures aren't. This is like saying "I asked an author to write a book and they plagiarised the hero's arc!". And even if it was copied straight-out you'd have an uphill battle to fight, to wit, wikipedia is quoting the thing verbatim.

That said copilot seems to be severely over-fitted in places, and I don't like the thing one single bit, and the only thing it's generally good at is writing code faster that shouldn't have been written in the first place, but inverse sqrt isn't a good example.

[–] [email protected] 3 points 6 months ago* (last edited 6 months ago) (2 children)

It didn't just get the gist if the algorithm though, it literally had the same magic number (which isn't even the most optimal iirc), the same COMMENTS (//what the fuck?), same variable names, etc.

It didn't produce the algorithm logically, it copied it.

Wikipedia is also adhering to the GPL license of the code. Copilot is not, especially if it's working on proprietary code or adding an MIT license header to copied GPL code (lol)

[–] [email protected] 1 points 6 months ago

I had bing chat spit back at me the question I posted on stack overflow the day before. You know, the example code I provided which didn't exactly work as I wanted.

[–] [email protected] 1 points 6 months ago (1 children)

It didn’t produce the algorithm logically, it copied it.

The magic number is part of the logic of the thing.

But yes as said copilot is overfitted. Inverse sqrt still isn't a good example, it's nearly as bad as Oracle trying to claim to have found copyright infringement in Android's standard Java library by saying that Math.average or whatnot is identical. There are way better examples of why copilot is fucked up.

[–] [email protected] 1 points 6 months ago

The magic number is part of the logic, yes, but that's not even the best magic number for the job iirc, and nobody remembers how they got it.

I just used this as an example because it's incredibly clear that it was copied verbatim (again, comments like "what the fuck?" showing up, you can't tell me it came up with that itself)