Technology

You've made a lot of confident assertions without supporting them. Just like an LLM! :)

ML techniques have been very useful in compression, yes, but it's sort of nuts to say that a data structure that encodes only (sometimes overly so for certain regions of its latent space/embedding space/semantics space/whatever you want to call it right now) relationships between values rather than value sequences themselves as storing contiguous copyright protected works is storing partiularized creative works in particularly identifiable manner.

On the other hand, it's hard to have a serious discussion with people who insist that building a LLM or diffusion model amounts to copying pieces of material into an obfuscated database. And then having to deal with the typical reply after explanation is attempted of "that isn't the point!" but without any elaboration strongly implies to me that some people just want to be pissy and don't want to hear how they may have been manipulated into taking a pro-corporate, hyper-capitalist position on something.

It's just sad seeing the stupid morons arguing for this as if they had anything to gain.

The real money shot here... How did we get to a point where people will argue against common working slave good?

There is a pattern too... Iraq, Afghanistan, israeli genocide, bailouts. Anytime there is money to be made for the regime, we got solid 30% of population working as hard for zealots.

Them 2 decades later when the two wars failed, we can't find a single guy who support either war around 🤡

The same is somehow now shilling we "shouldn't invafe ukraine but Israeli needs tools to defend themselves"

I'm getting really tired of saying this over and over on the Internet and getting either ignored or pounced on by pompous AI bros and boomers, but this "there isn't enough free data" claim has never been tested. The experiments that have come close (look up the early Phi and Starcoder papers, or the CommonCanvas text-to-image model) suggested that the claim is false, by showing that a) models trained on small, well-curated datasets can match and outperform models trained on lazily curated large web scrapes, and b) models trained solely on permissively licensed data can perform on par with at least the earlier versions of models trained more lazily (e.g. StarCoder 1.5 performing on par with Code-Davinci). But yes, a social network or other organization that has access to a bunch of data that they own, or have licensed, could almost certainly fine-tune a base LLM trained solely on permissively licensed data to get a tremendously useful tool that would probably be safer and more helpful than ChatGPT for that organization's specific business, at vastly lower risk of copyright claims or toxic generated content, for that matter.

I don't feel it is. They aren't saying that their physical requirements should be free (computers, engineers, programmers, electricity, etc...) which is what is being used for the analogy (cheese, ingredients, etc...).

It would be better to claim "I run a sandwich shop and couldn't afford to run it if I had to pay for every recipe, idea, and technique I use in the business."

Now, it's not as simple as this, and I'm not claiming it is. But this example isn't anywhere near correct. It's like the old claim that pirating something is the same as stealing it. The usage on one thing doesn't equal the loss of something physical.

It's one of those reasons why laws about this are difficult. Too strict and no one would be able to do "fan"-anything and many other issues ("if it uses AI" takes out many digital tools, etc...), too loose and you don't really have laws at all.

In the same way that a person can learn the material and also use that knowledge to potentially plagiarize it, though. It's no different in that sense. What is different is the speed of learning and both the speed and capacity of recall. However, it doesn't change the fundamental truths of OP's explanation.

Also, when you're talking specifically about music, you're talking about a very limited subset of note combinations that will sound pleasing to human ears. Additionally, even human composers commonly struggle to not simply accidentally reproduce others' work, which is partly why the music industry is filled with constant copyright litigation.

Yep, its definitely not possible that nice small businesses like universal and sony would sue without an actual case in order to try and crush competitors with costs.

And free cotton candy and rainbows for everybody!