zlacker

[parent] [thread] 4 comments
1. bradch+(OP)[view] [source] 2023-12-27 14:55:44
On the other hand, you could also argue that if AI takes all financial incentives from professionals to produce original works, then the AI will lose out on quality material to train on and become worse. Unless your argument is there’s no need for anything else created by humanity, everything worth reading has already been written, and humanity has peaked and everyone should stop?

Like all things, it’s about finding a balance. American, or any other, AI isn’t free from the global system which exists around us— capitalism.

replies(2): >>logicc+k2 >>Salgat+hYc
2. logicc+k2[view] [source] 2023-12-27 15:08:07
>>bradch+(OP)
>financial incentives from professionals to produce original works

People produce countless volumes of unpaid works of art and fiction purely for the joy of doing so; that's not going to change in future.

replies(2): >>sensan+X6 >>cozzyd+0a
◧◩
3. sensan+X6[view] [source] [discussion] 2023-12-27 15:33:59
>>logicc+k2
Anecdotal but I know lots of creatives (and by creatives I also include some devs) who've stopped publishing anything publicly because of various AI companies just stealing everything they can get their hands on.

They don't mind sharing their work for free to individuals or hell, to a large group of individuals and even companies, but AIs really take it to a whole different level in their eyes.

Whether this is a trend that will accelerate or even make a dent in the grand scheme of things, who knows, but at least in my circle of friends a lot of people are against AI companies (which is basically == M$) being able to get away with their shenanigans.

◧◩
4. cozzyd+0a[view] [source] [discussion] 2023-12-27 15:51:04
>>logicc+k2
Why should OpenAI be the one making money off their hard work, even if they do it for free?
5. Salgat+hYc[view] [source] 2024-01-01 15:24:17
>>bradch+(OP)
The whole "AI training blackhole" thing is a myth. As long as humans are curating the content generated by ML, the content generated is still valid training data. Remember, for every ML generated image you see online, someone had to go through countless attempts to get it to create exactly what they wanted.
[go to top]