Hacker News Reader

Comment by paxys | Hacker News Reader

To start, you have to treat copyrighted works on Github as a whole (say the entire project/repository or at least substantial parts of it) rather than look at the few lines that copilot copied in isolation. With that context:

1. The derived work is transformative. It uses the original as inspiration but the end result is substantially different.

2. Only a trivial amount of the work is copied (a few lines of code out of thousands/millions).

3. It is hard to prove that the creators of the original work were damaged by copilot (for example did they lose any customers or revenue?) This is normally the single biggest test in such cases.

View on HN

> you have to treat copyrighted works on Github as a whole (say the entire project/repository or at least substantial parts of it

But that's not how copyright works here, so this analysis seems irrelevant.

> 1. The derived work is transformative. It uses the original as inspiration but the end result is substantially different.

It's not. Changing variable names isn't a substantial difference.

> 2. Only a trivial amount of the work is copied (a few lines of code out of thousands/millions).

That doesn't matter. Relative size doesn't matter, absolute size does.

> 3. It is hard to prove that the creators of the original work were damaged by copilot (for example did they lose any customers or revenue?) This is normally the single biggest test in such cases.

??? No? You appear to have no clue how copyright works or what copyright infringement does. It's damaging solely in that you don't have the right to reproduce it without my permission. It's mine.

Well you can have your own interpretation of copyright law, but the Supreme Court – who ruled in favor of Google exactly on the basis of these three tests I shared above – will disagree with you.

Ruling on those basis in a case doesn't make that the only test for whether something infringes copyright or not. You have to qualify for all of those things.

Also, there were four tests.

The general principle of the ruling is that:

- it was mostly about APIs; organization rather than implementation.

- it was sufficiently transformative

- it was a small amount of code of insubstantial value

- it was serving a different market

So, I suppose you have indeed proven that we can have different interpretations of copyright law.

Taking a small but substantial piece of implementation code and using it in a similar way for a similar purpose to solve a similar problem would appear to fail all of those tests, and at least to me smells like rancid infringement.

Well you can have your own interpretation of copyright law, but the Supreme Court – who ruled in favor of Google exactly on the basis of these three tests I shared above – will disagree with you.

Reply on news.ycombinator.com