I’ve felt the same way. After years of patience and practice my kid has become a fantastic artist. I’m a bit concerned that at some point they might think “why bother?” and give it up after they see someone type some prompt in with “pencil, sketch” as a suffix.
To be honest, I already got to the “why bother” state by seeing on the internet how many human artists are so much better than me.
And, on the other hand, if you actually have a specific vision of what you want to create, it is often quite difficult, or actually impossible, to get an AI to generate an output that fulfills your particular vision.
> To be honest, I already got to the “why bother” state by seeing on the internet how many human artists are so much better than me.
I'm also learning to draw. As a counterpoint, I found it quite fun and enlightening to use artwork I like as practicing material. You can learn a lot by trying to copy parts of an artwork (without tracing and without uploading the result anywhere) : By copying, you still actually have to draw the artwork and deal with the same challenges that the original artist did - however you can use the original artwork as a sort of guide or cheat sheet on how those problems can be addressed.
I found it often even gave me a deeper appreciation for the original artist, because you see all the small, individual decisions they had to make to eventually get to the artwork - it felt sort of similar to reading code and trying to get into the head of the original developers.
You can also get creative and only copy parts of the artwork or stitch together a new piece from different sources - essentially doing by hand what image generators do automatically. (Just don't upload your new artwork anywhere without contacting the original authors first)
These threads remind me a lot of some of Isaac Asimov "Robots" series novels. He painted a world where Robots and AI solved a lot of things, so people didn't have to do things that "worried them". It seems some lost any meaning for living.
I've been an "amateurish" guitar player for 30 years. I play chords, I do some OK picking and I love trying to "play" songs I like, even if badly played (I like speed metal and shredding, which I know I will never master in my lifetime).
Nevertheless, I do it because I like it, I've seen Satriani, Oscar Lopez, Steve Vai play, and I know I'll never be like them. However I don't think "why bother". I enjoy playing basic guitar some afternoons in my living room, and my wife enjoys listening to me while reading a book. I've always played for me first, and because I enjoyed it.
It is a very rare opportunity that we people have the ability/resources to do something just because we enjoy it. With the advent of this wave of AI, we will need to learn to do more things that way.
As long as it’s humans you know it’s achievable. You can pursue their level and doing that has value. I kind of believe that learning about art and getting better is just a great exercise for oneself. It makes you a better person.
> To be honest, I already got to the “why bother” state by seeing on the internet how many human artists are so much better than me.
Same here, it's extremely hard to find motivation when I know every idea I have has been done better by someone who isn't me, or recently by a computer. In either case I can't meaningfully compete and if it's been done better, why do it at all?
This applies to art, but even more broadly than that. I get no satisfaction in creating something when I can see people who have done it far better than me via a quick search.
I'd still much rather see real art. It's kind of the difference between eating a homemade meal and eating manufactured crap out of a vending machine. And, from what I can tell, the art generators are basically just stealing other people's stuff. It'll probably be brought up as a real IP issue at some point.
No, that's just not how they technically work. AI generators train on an image archive, but the resulting model doesn't include the actual pictures in it, and often contains only a tiny influence from each source image.
I think the current popular model works out to somewhere around a byte in the model per sample image.
Well, if I were an artist who wants to make work publicly available (previews?) and charge for it, I would make sure the license I use to publish my work explicitly states that it's not allowed to use my work as part of training data for any ML algorithm. If they want my work for that, then pay for it.
Hell, I think I as a developer could do that... Gonna invent a new open source license.
If you think you'll get paid, you probably won't because AI trains on very large datasets, and so pretty much any price per image would quickly get too expensive.
If you think it'll hamstring AI, I doubt it'll be for long. People will just end up curating their dataset better, and images that allow for it will continue to be created.
Also, per https://opensource.org/osd -- "No Discrimination Against Fields of Endeavor", so this likely won't get much uptake in open source circles either.
Not being in the dataset maybe? There's many artists that hate that you can just write "in the style of" with their name.
So at least, if it does appear that it works with their name, they now either have been paid that amount, or they can sue for that amount of lost revenue. At least not everything would be lost.
I'm not a lawyer though, so really no idea whether this could even be possible as it's seems like currently it's fair use.
With very limited exceptions you could go to a photographer and say "I want a photo that looks like this Vogue cover" and get it done.
> So at least, if it does appear that it works with their name, they now either have been paid that amount, or they can sue for that amount of lost revenue. At least not everything would be lost.
Don't be silly. If that logic worked, Microsoft could have sued Linus Torvalds for providing an OS for free. Or if you want to be a bit more exact, Sun Microsystems, IBM, or any other Unix vendor.
Encoding and compression also have the property that they vastly change the data _representation_ of information (often changing it completely). Yet, the underlying thing reamins the same in a way that's easily recognizable by a human. This is similar - sure, a transform is being applied that changes the representation of the data. But, ultimately, it's just mashing together large chunks of other people's IP.
all AI art is confined to digital medium (at least for now, 3D printing might change that). You can't see the brushstrokes like in a real life, physical painting. Dancing is one that will be very hard to copy, because you need actual humans for it.
Not to mention, the kid may enjoy making art, and his/her art may give others joy, even if it's not a profitable career. It can still be a fun hobby - synthesizers have been around for nearly half a century now, but people still play guitar :)
I'm not sure I'd hang my hat on a particular current inability of AI, particularly inabilities where there's no obvious impediment to AI doing it. You're just begging for AI to eventually gain that ability, and now you're back where you started. Hooking AI up to physical actuators and getting it to work properly strikes me as a "this will 100% happen, I guarantee this is already being worked on" situation; basing your self-worth off the inability of AI to do those things seems like a ticking time bomb.
Seems like this is missing the point of dance, at least part of which is specifically watching humans. Right now, a pitching machine can throw harder than a human, but we still restrict baseball teams to only having humans on them. An automated motorbike can cover 100 meters faster than any human, but we don't allow motorbikes to compete in an Olympic sprint competition. A mechanical sumo wrestler wouldn't need to be intelligent, just heavier than the heaviest competitor, and it could already win against any human right now, but we don't do that.
Similarly, people still play chess, go, and Starcraft even though AI can beat them. Human Jeopardy matches still get televised even though Watson could beat them. We still have math olympics even though Wolfram Alpha could have won all of them decades ago.
I think human dancers are safe for as long as there are humans.
Exactly. Things in life are not valuable pursuits just because an AI can't do them, nor do they stop being valuable once an AI can participate. OP seemed to suggest that AI being confined to digital art means that physical art still has value, but it doesn't follow at all. If I strap a paintbrush to a computer with AI, it doesn't devalue human painting. They're not related in that way.
Indeed, this is already happening. This company is using human artist works for now, but the leap to a model producing such art is quite small at this point: https://art.art/blog/portia-de-rossi
alright then, I'll fallback to my backup - humans are cheaper.
They don't require mining rare earth metals, they're more versatile, and uh... oh gosh it's only a matter of time before we're all replaced by cheap carbon based robots. It's been real <3
> all AI art is confined to digital medium (at least for now, 3D printing might change that). You can't see the brushstrokes like in a real life, physical painting.
For now. There's no technical reason why we can't have CNC painting machines - there's plenty of existing tech in manufacturing that has the required precision and range of motion. I think we don't see such machines because there is no business need for them - or at least, there wasn't.
To date, the kind of digital images you'd want to reproduce using a CNC painter would be drawn by actual artists, using their hands and - most likely - tools like pens and drawing tablets. The work of an artist is similar enough to actual painting that, if there's a need for the 3D texture, they can just... use real brush on real canvas. There technically is a market for cheap reproductions, but AFAIK this one is covered by "painter farms" - warehouses full of low-paid painters mass-producing paintings, in countries with cheap labor costs.
Now, if you could get the ML model produce "original" work, encoded as a time series of brush strokes (either straight from the model, or by another model turning JPEGs into painting instructions), then this just might provide an incentive for CNC painters. It might even start as a hobby thing - there's lots of people who would build such machine, the way they build 3D printers now, so they could produce their own paintings at home. Those same people will be unlikely to pay a far-away painter farm instead.
> Dancing is one that will be very hard to copy, because you need actual humans for it.
AFAIR that was already done, to a passable level, by Boston Dynamics, a few years ago. Sure, that was a one-off carefully choreographed stunt, but it demonstrated that 1) robotics can handle dancing just fine, and 2) the result is actually enjoyable to watch. The missing bit is software, but given the ML work being done in recent years, in particular all those video editing/generating models that start by parsing human movement from input videos, I think it won't be hard for someone to start generating dance moves.
But what you’ll not have is a human behind it, with a human story and struggles. But yes, I could see mass produced paintings a la CNC becoming common. It will be more expensive than prints but only a few levels above ikea art since it’ll be mass produced.
Every mind sees things different. The artist struggles to put that to a medium. The successful expression of their vision is extremely common, but the ability for anyone to even iterate on another artist's work is basically nonexistent. Even with AIs cranking out art, the obsessed will continue to fight to share their mind.
Imagine he might choose (or already does it) to do it professionally and struggle to make ends meet because “AI is just good enough. People still consume the media we produce. Why hire artists?”
AI art is not nearly there yet, and has very significant areas in which it struggles a lot. The funny thing about it is that just like normal art there are easy and hard things to accomplish, but it's nigh impossible for a person new to this concept to tell which is which: https://xkcd.com/1425/
Overall, the easy part of AI art is generic portraits where you either don't care about the details, or want to use existing, popular things. Eg, you can ask for Harry Potter. And you can ask for a random character. But if you have a personal, distinctive design you made and want to bring to life, then it can get very tough pretty fast.
In fact, the best possible user for an AI generator would be an artist that's also capable of training a custom model. That opens up way more possibilities. Here's what this looks like:
