Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The AI Unbundling

stratechery.com

Post reply on HN
Print view View original

101–110 of 207 posts

Re: The AI Unbundling

#101

Post by sinenomine »

slowmovintarget wrote: post #88
dwohnitmok wrote: post #64

Earlier quoted context omitted.

> Because that's not true at all. The AI can't even draw hands yet. To say nothing of its ability to handle multiple people and objects interacting in complex scenes. This seems to be purely an issue of the size of the network. Parti ( https://parti.research.google/ ) demonstrates that as the number of parameters increases, with no change to the underlying architecture, a lot of these problems simply go away. Basical…

Doesn't "the AI" train on art produced by people? "Just expand the dataset, just increase the parameters" seems like it should hit a wall fairly quickly... and still not be very good, because deep learning systems have no insight.

Every instagram, facebook, and tiktok photo with associated text data is a potential pair for training.

In the smartphone age, the case for data hunger looks pretty weak.

Re: The AI Unbundling

#102

Post by simiones »

echelon wrote: post #92
simiones wrote: post #84

Earlier quoted context omitted.

Is this sarcasm, or do you think this is somehow impressive?

Not sarcasm. I'm working in this space and you can check my bio. If you're not excited by these rapidly improving results, then I don't know what to tell you. I think you can extrapolate what lies just ahead if you think about it. I've been a tech pessimist for the last two decades of my life, and I can strangely now state that this is the most important and exciting time of my entire life and career. Everything we k…

I can only wish you luck and say that personally, I'll believe it when I'll see it (that is, when I first end up enjoying a piece of art - book, game, movie, song etc. - where AI generation played a major part in its creation).

Re: The AI Unbundling

#103

Post by _aavaa_ »

simiones wrote: post #95
_aavaa_ wrote: post #87

Earlier quoted context omitted.

And you are free to search through the whole catalog of google results until you find an owl that looks exactly like you want. Though this is going to get harder as you want something more specific than a simple owl. But the approach for stable diffusion is just as easy whether you want just "an owl", or "an owl in X's style with A, B, and C"

Changing the prompt until it generates what I want is not that different from changing my search terms until the result I want is closer to the top. Now, I should of course note that search engines already employ ML techniques to actually interpret search terms, so to some extent the point is moot - ML is important to actually solving this problem.

But searching on google doesn't "generate" anything, If your image isn't on the web, there's nothing to bring "closer to the top".

Re: The AI Unbundling

#104

Post by erispoe »

dwohnitmok wrote: post #64
Jevon23 wrote: post #47

Earlier quoted context omitted.

Are you under the impression that right now, as of today, the publicly-available AI models are ready to replace humans for all types of art outside of scientific and technical illustration? Because that's not true at all. The AI can't even draw hands yet. To say nothing of its ability to handle multiple people and objects interacting in complex scenes. I'm concerned that the discussions about AI art on forums like HN…

> Because that's not true at all. The AI can't even draw hands yet. To say nothing of its ability to handle multiple people and objects interacting in complex scenes. This seems to be purely an issue of the size of the network. Parti ( https://parti.research.google/ ) demonstrates that as the number of parameters increases, with no change to the underlying architecture, a lot of these problems simply go away. Basical…

People who buy art do not buy it because of the technical execution. You may need to execute a piece in some way to get a desired effect, but the technique is the mean not the goal.

This is not to take away from the achievements of AI. It's that creating pictures adhering to a prompt with some degree of creativity is very little of what art is. Maybe it will replace some part of commissioned illustrations where the artist's name does not matter (e.g. some avatar pic?).

We still value, financially, some material goods for much more than they cost to produce. Or for much more than their almost identical mass-produced counterparts.

Re: The AI Unbundling

#105

Post by sinenomine »

Jevon23 wrote: post #76
dwohnitmok wrote: post #64

Earlier quoted context omitted.

> Because that's not true at all. The AI can't even draw hands yet. To say nothing of its ability to handle multiple people and objects interacting in complex scenes. This seems to be purely an issue of the size of the network. Parti ( https://parti.research.google/ ) demonstrates that as the number of parameters increases, with no change to the underlying architecture, a lot of these problems simply go away. Basical…

Well, we’ll see how it performs, if it’s ever made public. The 20B images don’t look that much more impressive than what SD is already doing (aside from the ability to render text), and in some cases they look worse. It’s hard to tell because the resolution is so small, but even in the 20B “astronaut riding a horse through a pond” image, it looks like his hands are still nonsensical.

This nitpick about hands sounds desperate. Here we are, with a tech so powerful that it overshadows the default hype it's surrounded by (no small feat, most technologies fail to live up to the hype as you know) ... and the critics merely move the goalpost a tiny bit further, even if the tech scales so well as to make their new goalpost irrelevant in a year.

Re: The AI Unbundling

#106

Post by simiones »

_aavaa_ wrote: post #103
simiones wrote: post #95

Earlier quoted context omitted.

Changing the prompt until it generates what I want is not that different from changing my search terms until the result I want is closer to the top. Now, I should of course note that search engines already employ ML techniques to actually interpret search terms, so to some extent the point is moot - ML is important to actually solving this problem.

But searching on google doesn't "generate" anything, If your image isn't on the web, there's nothing to bring "closer to the top".

Sure, but chances are, it is already on the web.

And of course, it's also possible that the image I want can't be generated by SD/DALL-E/etc.

Re: The AI Unbundling

#107

Post by karaterobot »

mejutoco wrote: post #39
gizmo wrote: post #33

Although we talk a lot about disruption, only very few technologies are truly disruptive. You can tell by the panic and awe in the air whether you're dealing with real disruption or incremental change. Dropbox made filesharing easier. It's a good product, but not disruptive. Nobody panicked that Dropbox would make their job redundant. Uber was hard on the taxi industry, but fundamentally you still have drivers taking…

> People will insist that SD art isn't real art. Artists will fight back, and lose. When talking about Stable Diffusion and art there are usually two different aspects of art. I am not going to try to define art, but sometimes we refer to art as illustration, or stock images (what SD puts in danger) and some as broad modern art. I am no trying to say that one is more valuable than the other, but want to qualify these…

This comment is right on the money in making the distinction between illustration/stock images, and fine art, let alone installation or performance art. It's a distinction not made often enough in these conversations. These functions are performed by different people, for different reasons, and they're used in different ways.

Sadly, the market for the kind of art that the these models cannot disrupt is relatively small compared to the kind it can. There aren't a lot of people making a family-supporting living doing installation art. Most people I know with art degrees do illustration, photography, or are part of a video game or film asset production pipeline (or they draw tattoos, but that's a different matter). If I were them, I would be looking at the next generation of this technology as a potential threat to my livelihood. I don't want to be alarmist, but it is a possibility, and it'd be weird to dismiss it.

One other thing I'll tack on here is that I find it fascinating that the kinds of skills required to be "good" at using these image generation models — "prompt engineering" if you like — are largely different than the ones required to create art from scratch. You can be a great studio painter, but not be able to "talk to" Stable Diffusion at all. Likewise, you may have zero artistic ability in the traditional sense, but be a prodigy at getting the computer to spit out what you are imagining, or something even better than that. If AI generated art is determined to be a kind of art (as I believe it will be) the parameters of what we call artistic ability may change.

Re: The AI Unbundling

#108

Post by johndfsgdgdfg »

mejutoco wrote: post #39
gizmo wrote: post #33

Although we talk a lot about disruption, only very few technologies are truly disruptive. You can tell by the panic and awe in the air whether you're dealing with real disruption or incremental change. Dropbox made filesharing easier. It's a good product, but not disruptive. Nobody panicked that Dropbox would make their job redundant. Uber was hard on the taxi industry, but fundamentally you still have drivers taking…

> People will insist that SD art isn't real art. Artists will fight back, and lose. When talking about Stable Diffusion and art there are usually two different aspects of art. I am not going to try to define art, but sometimes we refer to art as illustration, or stock images (what SD puts in danger) and some as broad modern art. I am no trying to say that one is more valuable than the other, but want to qualify these…

> I remember an example of an artist paying illegal immigrants to hold a wall (that could not stand by itself) in a gallery, to touch on social issues.

Do you happen to have a link? This sounds amazing. I wish more artists and mainstream would come forward to help immigrants. The way the states treat the immigrants is inhumane. We should open our borders, not build more walls.

Re: The AI Unbundling

#110

Post by ljp_206 »

Jevon23 wrote: post #40
gizmo wrote: post #33

Although we talk a lot about disruption, only very few technologies are truly disruptive. You can tell by the panic and awe in the air whether you're dealing with real disruption or incremental change. Dropbox made filesharing easier. It's a good product, but not disruptive. Nobody panicked that Dropbox would make their job redundant. Uber was hard on the taxi industry, but fundamentally you still have drivers taking…

>It's going to destroy the livelihoods of the majority of independent artists in a way that looks inevitable to me. Why is SD going to destroy the livelihoods of artists when machine language translation hasn't put human translators out of work yet? I don't think there's been any industry that's been ended by AI yet, and yet people are strangely confident that art is going to be the first.

I am becoming more and more convinced that many techy folks near the AI scene saw that SD et alia can create an image that convincingly (and perhaps even indistinguishably) looks like a very nice digital painting, and based on that data point alone are calling artists obsolete.
Post reply on HN