4 ms·

I'd be far more amenable to corporations training their AI on my content of those exact same corporations hadn't spent the last two decades aggressively defendi

I'd be far more amenable to corporations training their AI on my content of those exact same corporations hadn't spent the last two decades aggressively defending their own IP with DRM and multi-million dollar lawsuits. So they can go fuck themselves, or alternatively they can make their super advanced AI reproduce my copyright statement and license every time it copies my code. Which shouldn't be difficult at all.

kerpotgh 4y ago
[flagged]
- dang 4y ago
  Would you please stop breaking the site guidelines so we don't have to keep banning you? I appreciate your good comments but bad ones destroy more than good ones contribute. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
oehpr 4y ago
Indeed, they can go fuck themselves. But if copyright law gets put in place against AI, the only ones who will be able to enjoy the benefits of AI are the people who can go fuck themselves. The gatekeepers that will just add yet another line to their 10 billion line long EULA that will require you give up the rights to any art you possess for AI training, to the exclusive use of mega co. inc. ect.
nhchris 4y ago
Your stance, if it becomes law, guarantees that only the giant corporations you hate will be able to negotiate licenses to train their AI. It guarantees that libre AI tools will be left far behind. Is that what you want?
- snickerbockers 4y ago
  i think it's arrogant and selfish to assume that everybody has a stake in AI research. i want the license that the software i create (which has nothing to do with AI) to be obeyed. anyways, there's a lot of ways that AI researchers could engage with IP owners to come up with a fair way to use their work, but nobody's making that effort. If my content is part of an AI's training set (and especially if that AI has a tendency to output excerpts of its own training set verbatim, as github's copilot has been shown to do) then it's not unreasonable to set terms and conditions, which could restrict how the content is used for training and what sort of compensation (if any) I deserve. I'm of the opinion that it's time for new versions of GPL and CC licenses to be created which will enumerate how content can be used for AI training.
  - nhchris 4y ago
    > it's arrogant and selfish to assume that everybody has a stake in AI research I didn't assume anything, just described the likely consequence of your preferences.
kdnvk 4y ago
Which corporations are you referring to?
- snickerbockers 4y ago
  mostly Microsoft but i'm sure there's plenty of hypocrisy coming from the others too.

kerpotgh 4y ago

[flagged]

dang 4y ago
Would you please stop breaking the site guidelines so we don't have to keep banning you? I appreciate your good comments but bad ones destroy more than good ones contribute. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html

oehpr 4y ago

Indeed, they can go fuck themselves. But if copyright law gets put in place against AI, the only ones who will be able to enjoy the benefits of AI are the people who can go fuck themselves. The gatekeepers that will just add yet another line to their 10 billion line long EULA that will require you give up the rights to any art you possess for AI training, to the exclusive use of mega co. inc. ect.

nhchris 4y ago

Your stance, if it becomes law, guarantees that only the giant corporations you hate will be able to negotiate licenses to train their AI. It guarantees that libre AI tools will be left far behind. Is that what you want?

snickerbockers 4y ago
i think it's arrogant and selfish to assume that everybody has a stake in AI research. i want the license that the software i create (which has nothing to do with AI) to be obeyed. anyways, there's a lot of ways that AI researchers could engage with IP owners to come up with a fair way to use their work, but nobody's making that effort. If my content is part of an AI's training set (and especially if that AI has a tendency to output excerpts of its own training set verbatim, as github's copilot has been shown to do) then it's not unreasonable to set terms and conditions, which could restrict how the content is used for training and what sort of compensation (if any) I deserve. I'm of the opinion that it's time for new versions of GPL and CC licenses to be created which will enumerate how content can be used for AI training.
- nhchris 4y ago
  > it's arrogant and selfish to assume that everybody has a stake in AI research I didn't assume anything, just described the likely consequence of your preferences.

kdnvk 4y ago

Which corporations are you referring to?

snickerbockers 4y ago
mostly Microsoft but i'm sure there's plenty of hypocrisy coming from the others too.