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

31–40 of 207 posts

Re: The AI Unbundling

#31

Post by jeffreyrogers »

awillen wrote: post #14
hertzrat wrote: post #5

Isn’t it likely that patent law is going to lock ai down? One company will finally crack something general-enough in a way that gets it way ahead, will patent it, and will use that monopoly to gain control of the whole space, and the monopoly will never go away because all further improvements will also get patented by the ai legal team? The brief openness of ai right now is a glitch in the system that will get irone…

The problem with patenting AI is that you have to describe how something works in order to patent it. The issue presently is that you can describe relatively obvious things (e.g. one click checkout) and patent those. AI is really a black box, which should make patenting specific implementations of it very difficult. Even if you do manage to patent something like a diffusion model of image creation, in order to enforc…

The patent would apply to the model's architecture rather than the trained values so the black box nature of deep learning doesn't matter too much.

Re: The AI Unbundling

#32

Post by andreilys »

time_to_smile wrote: post #22
andreilys wrote: post #19

Earlier quoted context omitted.

I’m curious why do you think that AI generation will plateau at its current quality level?

Same reason that self driving cars have plateaued. The things that are really important for that last mile are also really hard and we're not really sure how to get there. The output between all the major AI image synthesis tools, while impressive, is quite clearly very high dimensional interpolation, and it shares the same limitations in that space as it does the 2d space. We've already thrown insane amounts of data…

Self driving cars and generating AI are two completely different problem sets.

Particularly as one deals with the world of atoms, where the cost of a mistake is much higher.

There are already models that work to “aesthetically” rate images, and the results are quite promising.

To think we won’t see leaps in quality over the coming months/years is narrow sighted.

Re: The AI Unbundling

#33

Post by gizmo »

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 from A to B. First you had to call the cab company now you use an app. It's an improvement, but not truly disruptive. Not like level 5 self-driving cars would be.

Amazon and Walmart by contrast did disrupt entire industries. Independent book shops and mom&pop retailers saw the writing on the wall. They knew they couldn't survive facing this kind of competition, and largely, they didn't.

Stable Diffusion (and similar tools) fall in this last category of truly disruptive technologies. It's going to destroy the livelihoods of the majority of independent artists in a way that looks inevitable to me. These new tools boost artist productivity by 100x and that means good artists will be producing much more art than ever before. This pushes cost down and quality expectations way up. Some artists will adapt and thrive in this new environment, but the majority won't. It won't be long until making a living with photoshop will become as hard as making a living playing guitar. This is good for society but bad for many individual artists.

There will be a backlash. People will insist that SD art isn't real art. Artists will fight back, and lose. Because SD isn't going anywhere. This is what disruption looks like, and it isn't pretty.

Re: The AI Unbundling

#34

Post by andreilys »

wpietri wrote: post #28
andreilys wrote: post #21

Earlier quoted context omitted.

How do you think artists “higher up the food chain” got there? Inevitably they started on the lower rungs and worked their way up. If these tools displace the need for lower level/apprenticeship like work, then it won’t be a surprise if later on we have less quality artists who weren’t able to get the necessary experience early on.

Do you have some evidence of this being how we are currently getting quality artists? Talking with one digital artist, my impression was that stock art plus the ever-declining ad CPM rate eliminated most of the lower-level work years ago.

I don’t see why AI generation would be restricted to stock art. There are plenty of very impressive results already that will displace artists further down the chain.

I think of it more like, If an artist can support themselves doing commissions of low quality art so they can support their high quality art, that’s better than if they had to drive Uber.

https://mobile.twitter.com/xsteenbrugge/status/1558508866463...

Re: The AI Unbundling

#35

Post by JackFr »

Having recently watched the Warhol Diaries on Netflix, I can't help but think of the examples of Andy Warhol, Keith Haring and Jean-Michel Basquiat and think that calls for the death of the artist are premature. The funny thing is that I don't find Keith Haring or Jean-Michel Basquiat's work particularly compelling, and from the a technical perspective, I don't think either of them were blazing new paths, but what they (and Andy Warhol) offered was something Stable Diffusion simply cannot offer, that is, something new.

You can't ask for "in the style of Keith Haring' before there's a Keith Haring

Re: The AI Unbundling

#36

Post by time_to_smile »

andreilys wrote: post #32
time_to_smile wrote: post #22

Earlier quoted context omitted.

Same reason that self driving cars have plateaued. The things that are really important for that last mile are also really hard and we're not really sure how to get there. The output between all the major AI image synthesis tools, while impressive, is quite clearly very high dimensional interpolation, and it shares the same limitations in that space as it does the 2d space. We've already thrown insane amounts of data…

Self driving cars and generating AI are two completely different problem sets. Particularly as one deals with the world of atoms, where the cost of a mistake is much higher. There are already models that work to “aesthetically” rate images, and the results are quite promising. To think we won’t see leaps in quality over the coming months/years is narrow sighted.

Coinbase stock keeps going up despite it being widely acknowledged that crypto is fundamentally a scam, so anything is possible.

Though it's more likely in a the coming months/years people's ability ignore reality will just become stronger and we'll all believe that a person waving their 4 wiggly fingers and staring at you with strange fish eyes is hyper realism at it's finest.

We've entered the "emperor's new clothes" reality at this point, so nothing would surprise me.

Re: The AI Unbundling

#37

Post by soared »

JackFr wrote: post #35

Having recently watched the Warhol Diaries on Netflix, I can't help but think of the examples of Andy Warhol, Keith Haring and Jean-Michel Basquiat and think that calls for the death of the artist are premature. The funny thing is that I don't find Keith Haring or Jean-Michel Basquiat's work particularly compelling, and from the a technical perspective, I don't think either of them were blazing new paths, but what th…

It seems it is possible to achieve that, but it may look different.

Since we still have to give text prompts to generate images, and these prompts need qualifiers (x subject “oil painting”) it seems possible that a prompt giver could generate images in a specific way to achieve their own unique style.

Difficult now since the tools are new and don’t have a ton of features, but down the road if the tools are built out more and prompts get more nuanced I could see “prompt givers” as “artists with their own unique style”

Re: The AI Unbundling

#38

Post by soared »

wpietri wrote: post #28
andreilys wrote: post #21

Earlier quoted context omitted.

How do you think artists “higher up the food chain” got there? Inevitably they started on the lower rungs and worked their way up. If these tools displace the need for lower level/apprenticeship like work, then it won’t be a surprise if later on we have less quality artists who weren’t able to get the necessary experience early on.

Do you have some evidence of this being how we are currently getting quality artists? Talking with one digital artist, my impression was that stock art plus the ever-declining ad CPM rate eliminated most of the lower-level work years ago.

There are still plenty of entry and lower/mid level graphic artists working at creative agencies, ad agencies, etc. it’s possible some may be replaced but their skills would still likely be needed if AI is doing most of the work for edits, adjustments, exporting out of .PSD, etc

Re: The AI Unbundling

#39

Post by mejutoco »

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 two, because some art is not painting pretty pictures.

In the modern art interpretation artists will not lose to SD. SD will enable them to do different things. There are many examples of famous artists that commission the execution of an artwork fully, without painting, sculpting or doing any other work. 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.

I am not an expert in art. My point is, in modern art SD will be one tool more (although it might create new influences) at the service of the human that gives it meaning.

Re: The AI Unbundling

#40

Post by Jevon23 »

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.

Post reply on HN