Hacker News Reader

Comment by lofties | Hacker News Reader

That's quite bleak, and I hold an opposing view. My main view is people are generally quite good to discern AI cruft from "human-made" paintings, games, and music. As soon as it will be impossible to make money from AI cruft, the AI cruft will disappear. As a (hobby) game developer myself, no AI tool I have used has ever made me feel "wow, this is great". It has always required multiple iterations of feedback. I tried tools like Meshy too. I ended up hiring a part-time 3D designer from Indonesia, who is a joy to work with.

View on HN

We are good at distinguishing it, but that doesn’t mean the unmitigated flow of Super Intelligence-generated images, music, etc won’t degrade the spaces we share and exhaust people from filtering from the art that they do want.

I can tell an image is AI generated almost instantly, but it nonetheless stomped places like the Photographers of Threads channel. Not that it was great to begin with, but it was something. More and more I’m motivated to only discover photographers via photo books.

For example, I no longer generally enjoy videos of animals doing cute things. LLMs have taken the wonder of that from us, because though animals do really unexpected things sometimes, it's more common these days for it to be someone puppetting an AI in the hunt for internet points. I discard anything post-2022 or so, unless it comes from a known good source. (which also generally has to have a presence pre-2022 in order to feel safe)

LLMs are butchering whimsy and wonder.

LLMs are butchering particular (digital) channels for whimsy and wonder, and it's easy to feel like those channels are everything but they haven't been that way for long and can easily change.

My view: every large online space was basically ruined by an ever-growing advance of inauthentic content pressures starting around 2015. The pools were already pissed in, LLMs are just producing enough piss that, finally, everyone can see it. It's a nuisance in the moment but in the medium term, the human social response will be for the better.

Much of make it by human talk is giant cope.

we insist that humans can do things better in some way or do things that ai can't do. some of that is true right now, but can't assume that will hold.

and yes it requires iterations and taste right now.

the idea that it can't happen in the future. or that humans are special in some way (which you are not claiming, but is a common claim) is the cope.

But it's missing the point: it is not always about "better" or "worse".

And in fact humans are special in a very straight-forward way: we share one thing with each other that no AI will share with us, which is our common biological ancestry and its consequences.

So, when looking at a human-made painting - which could look much worse than flawless AI counterpart - I'm not just looking at a painting, but I'm also thinking about what it took to produce, how much the artist was struggling to balance the need of survival with the need to create, how big was the price they had to pay, how it's similar or different to my own life, whether I'd be willing to pay such a price, and a lot more.

OTOH I don't care at all about the "struggle" of the ML model. I don't care about GPUs and TPUs sweating while climbing their "mountains". It might produce outputs that are aesthetically pleasant, but it will never be art.

Reply on news.ycombinator.com