zlacker

>>ssgodd+(OP)
I hope this results in Fair Use being expanded to cover AI training. This is way more important to humanity's future than any single media outlet. If the NYT goes under, a dozen similar outlets can replace them overnight. If we lose AI to stupid IP battles in its infancy, we end up handicapping probably the single most important development in human history just to protect some ancient newspaper. Then another country is going to do it anyway, and still the NYT is going to get eaten.

>>solard+Aj
Why can't AI at least cite its source? This feels like a broader problem, nothing specific to the NYTimes.

Long term, if no one is given credit for their research, either the creators will start to wall off their content or not create at all. Both options would be sad.

A humane attribution comment from the AI could go a long way - "I think I read something about this <topic X> in the NYTimes <link> on January 3rd, 2021."

It appears that without attribution, long term, nothing moves forward.

AI loses access to the latest findings from humanity. And so does the public.

>>aantix+1l
"Why can't AI at least cite its source" each article seen alters the weights a tiny, non-human understandable amount. it doesn't have a source, unless you think of the whole humongous corpus that it is trained on

>>make3+5m
that just sounds like "we didn't even try to build those systems in that way, and we're all out of ideas, so it basically will never work"

which is really just a very, very common story with ai problems, be it sources/citations/licenses/usage tracking/etc., it's all just 'too complex if not impossible to solve', which just seems like a facade for intentionally ignoring those problems for benefit at this point. those problems definitely exist, why not try to solve them? because well...actually trying to solve them would entail having to use data properly and pay creators, and that'd just cut into bottom line. the point is free data use without having to pay, so why would they try to ruin that for themselves?

>>pxoe+Wo
Just a question, do you remember a source for all the knowledge in your mind, or did you at least try to remember?

>>KHRZ+us
a computer isn't a human. aren't computers good at storing data? why can't they just store that data? they literally have sources in datasets. why can't they just reference those sources?

human analogies are cute, but they're completely irrelevant. it doesn't change that it's specifically about computers, and doesn't change or excuse how computers work.

>>pxoe+zt
Can't have your cake and eat it too.

1. If you run different software (LLM), install different hardware (GPU/TPU), and use it differently (natural language), to the point that in many ways it's a different kind of machine; does it actually surprise you that it works differently? There's definitely computer components in there somewhere, but they're combined in a somewhat different way. Just like you can use the same lego bricks to make either a house or a space-ship, even though it's the same bricks. For one: GPT-4 is not quite going to display a windows desktop for you (right-this-minute at least)

2. Comparing to humans is fine. Else by similar logic a robot arm is not a human arm, and thus should not be capable of gripping things and picking them up. Obviously that logic has a flaw somewhere. A more useful logic might be to compare eg. Human arm, Gorilla arm, Robot arm, they're all arms!