To me AI code generators are the equivalent of crypto tumblers or mixers for digital coins. You can pretend all you want that the output is 'clean' but we all know it came from somewhere else and wasn't actually generated by the software, just endless little snippets that other people made.
Ownership of AI-Generated Code Hotly Disputed
61–70 of 205 posts
Re: Ownership of AI-Generated Code Hotly Disputed
#62“…modify[ing] its AI model so that it traces attribution and gives credit to the original authors of the code, adding the associated copyright notices and license terms in the process…Biderman says is technologically feasible.” Is it really feasible? What does “traces attribution” even mean here? It’s not emitting “code”, it’s emitting individual tokens that each were found throughout the input corpus. The “code” is…
Honestly, I would be fine with verbatim chunks of code if it could gaurantee a compatible license and did the copyright notices properly.
That always seemed like a problem with open source, you should theoretically be able to copy bits of code from hundreds of projects to make a new one, but keeping track of the licenses makes it too much of a pain. So the closest we really see is people vendorizing libraries.
Re: Ownership of AI-Generated Code Hotly Disputed
#63by reducing credit, copilot reduces incentive to create and publish free code. biting the hand that feeds it. exact same problem exists with GPT3 and others. big tech slashing and burning, ruthlessly exploiting the least empowered people in the tech economy. neat hack.
How does copilot reduce the incentive to publish free code?
Re: Ownership of AI-Generated Code Hotly Disputed
#64by reducing credit, copilot reduces incentive to create and publish free code. biting the hand that feeds it. exact same problem exists with GPT3 and others. big tech slashing and burning, ruthlessly exploiting the least empowered people in the tech economy. neat hack.
I used to put most of my code on public repos in Github, Gitlab, etc. Since Copilot came out, I've moved most of my projects to a Git server on my home network.
The prospect of knowledge work being resold in this way feels icky. Perhaps I'd feel different if there was a simple, well-defined way to opt out of contributing to the training corpus. Opt-in would be even better.
Re: Ownership of AI-Generated Code Hotly Disputed
#65Regardless of what side you stand on, this case is sorely needed, so everybody knows what the legalities are.
Re: Ownership of AI-Generated Code Hotly Disputed
#66Earlier quoted context omitted.
> It’s not emitting “code”, it’s emitting individual tokens ... The “code” is the arrangement of those tokens, but that is determined by the weighting of the whole network This theory of operation is not borne out in reality. It's been clearly displayed that these tools are emitting verbatim copies of existing code (and its comments) in their input. It's even being seen in image generation, where NatGeo cover images…
> It's even being seen in image generation, where NatGeo cover images are reproduced in their entirety or where stock photo watermarks are emitted on finished images. Can you cite sources? I’ve heard this claim repeatedly but have yet to see a good example.
https://youtu.be/kqPKNksl9hk?t=80 https://news.ycombinator.com/item?id=33061707 <- Watermarks (there's been a few linked here on HN)
Re: Ownership of AI-Generated Code Hotly Disputed
#67Earlier quoted context omitted.
Yes, as long as it doesn't disregard the original creator's intentions in the formulation.
The original creator's intentions do not override IP law and neither should they. Otherwise any deluded fool with extreme views on copyright could claim ownership way beyond that which the law currently offers.
Sure, respecting intellectual property laws is exactly what must be considered here.
Re: Ownership of AI-Generated Code Hotly Disputed
#68https://en.m.wikipedia.org/wiki/Information_wants_to_be_free
I propose that information naturally wants to degrade. Paper decomposes. Bits flip. File formats are replaced and lost. Storage mediums degrade. It's all an extension of the universe trending towards entropy. It actually takes quite a bit of effort to store, then distribute information precisely and broadly. There's a lot of infrastructure, effort, and money involved, and still information degrades and disappears ove…
I cede your larger point that information availability requires upkeep, but I don’t think the tendency for information to degrade is usually attributable to entropy. It’s probably true enough in some cases, but more often than not it’s going to be general interest by the population that determines the accessibility of information—and that crosses boundaries and regulatory efforts like copyright. The more intense the interest and/or awareness (generally speaking) the more likely the information will find its way into the public domain.
It’s why you can easily find literature written hundreds or thousands of years ago in multiple languages with little effort, but some memo stored in a tape archive from the 1970s is likely gone forever. “Bitrot” is more likely to be a function of diminishing interest rather than increasing entropy.
Re: Ownership of AI-Generated Code Hotly Disputed
#69Earlier quoted context omitted.
> It’s not emitting “code”, it’s emitting individual tokens ... The “code” is the arrangement of those tokens, but that is determined by the weighting of the whole network This theory of operation is not borne out in reality. It's been clearly displayed that these tools are emitting verbatim copies of existing code (and its comments) in their input. It's even being seen in image generation, where NatGeo cover images…
> It's been clearly displayed that these tools are emitting verbatim copies of existing code (and its comments) in their input. Which makes sense when you consider that the sort of code that is getting reproduced verbatim is usually library functions which developers may copy and paste verbatim comments and all into their project, especially when you prompt the AI with the header of a function that has been copied an…
So that should make it easy to attribute, yes?
Re: Ownership of AI-Generated Code Hotly Disputed
#70“…modify[ing] its AI model so that it traces attribution and gives credit to the original authors of the code, adding the associated copyright notices and license terms in the process…Biderman says is technologically feasible.” Is it really feasible? What does “traces attribution” even mean here? It’s not emitting “code”, it’s emitting individual tokens that each were found throughout the input corpus. The “code” is…
They probably mean using a code search engine to check all snippets. The simplest thing would be am n-gram filter. A more advanced approach would use a code similarity neural net. It's not principled attribution, just locating the most similar example in the training set.