Comment by latexr | Hacker News Reader

Comment by latexr | Hacker News Reader

Parent Full thread

> It is very tiring to say “I don’t necessarily disagree with you about AI ‘art’, but in my field—which you do not understand, and in which the underlying build process is often not the creative output—AI presents very real productivity gains” for the umpteenth time.

That’s not a justification. If a company were poisoning the water to your home as a byproduct, would you be satisfied if they told you “we don’t necessarily disagree with you about polluting the water, but in our field—which you do not understand, and in which the underlying build process is often not the water pollution—what we’re doing presents very real productivity gains”?

> I am skeptical of there being sufficient data to build “ethical” training datasets

Then you don’t build any. What fucked up world we live in where people think it’s OK to be unethical because they want something and can’t think of any other way to do it. What monumentally selfish rotten babies.

is there nothing else that we take for granted as a convenience to humanity that also has serious negative externalities? just AI ?

That is called whataboutism, and it’s really annoying how prevalent it has become on HN discourse.

https://en.wikipedia.org/wiki/Whataboutism

Other things being bad doesn’t make it OK that this thing is and no one is saying that.

My argument was purposefully general, you’re the one who chose to bring it back to AI. Don’t argue in bad faith.

this is not whataboutism at all. This is my conjecture:

AI having negative externalities is not the sole deciding factor in deciding to eradicate it

then my argument: negative externalities are tolerated in many areas where we deem the topic is of enough value that the externalities are managed. there are many examples of things with negative externalities, much much larger than those of AI, that aren't being considered for eradication.

so your argument is implicitly, "AI is not worth these negative externalities"

which is a common opinion among people who dont use AI at all.

> If a company were poisoning the water to your home as a byproduct, would you be satisfied if they told you “we don’t necessarily disagree with you about polluting the water, but in our field—which you do not understand, and in which the underlying build process is often not the water pollution—what we’re doing presents very real productivity gains”?

no but I also wouldn't declare that whatever it is that company does should have its entire industry eradicated. Poor industrial practices can be mitigated while not abandoning the product being manufactured.

Looks like you have a very utilitarian approach to it, in that case, would you be fine with eradicating just diffusion and video models?

By your logic, they don't particularly contribute to scientific productivity/advances: it's not like you can use stable diffusion to generate an architectural diagram.

GenIA for images and videos is only really used to copy artstyles, make fake menus and create misinformation so wouldn't you agree that eradicating this part would not only preserve your described "useful applications" but also allow providers to allocate more resources to "useful" and "scientific" AI?

I know that local models have become decent enough that this isn't really doable, but I'm only proposing this thought experiment because you seem to think people are saying this should be all or nothing. But the truth is, if you get a genie to wipe away just diffusion models, the entire internet gets much better (or at least back to regular levels of bad), no one loses anything, and a whole lot of people stop complaining/campaigning against the productivity gains in your field.

Reply on news.ycombinator.com