Hacker News Reader

Comment by ergonaught | Hacker News Reader

If it reaches a decision, I’m curious how it will affect “education”. If you study a hundred repos to learn how to do a thing, and then produce something of your own that happens to be similar (because that’s how you learned), are you under any particular obligations?

One hopes Oracle et al are not further inspired by such an outcome.

View on HN

Humans aren't machines, so it won't affect education at all. This argument is a distraction.

Humans are machines actually. Whether what Copilot does can actually qualify as "learning" vs. merely generating derivative works might actually be relevant.

Machines aren't humans.

Sure, but can machines learn and create new works the way we do? That's the only relevant question.

Learning from others is copyright infringement, better make sure none of your code contains any sequences longer than 150 characters that match any other code ever written.

> Learning from others is copyright infringement

Isn't this the reason for clean room implementations?

That's specific to reverse engineering a product that has the same features as an existing one though. Part of the concept of open source is allowing others to learn from your implementation. In many cases one doesn't want to include an open source project directly in their code base for a variety of reasons, maintainability being a big factor. I don't see a clean room implementation being a good solution to the problem of wanting to use accepted standards and practices which must be gleaned from experience working with or learning from other people's code. I don't think you can get far arguing that all code should be written in a vacuum with no ability to learn from outside sources.

Reply on news.ycombinator.com