I do think that once the hype dies down, people are going to sober up and realize that you still need actual human artists in the loop. The problem is that thanks to wonders of capitalism, companies working on AI tech are overselling the capabilities to secure funding, and companies are buying into it because they want to cut employee costs. From what I’m seeing though, the tooling and workflows around stuff like stable diffusion are already getting really sophisticated. With stuff like control nets, selective inpainting, and so on.
In practice this should be looked at a tool akin to photoshop, where you need a lot of skill and training to produce professional results. Anybody can type in a prompt and get a random picture out of it, but in most cases you actually want a specific result with a particular style, lighting, composition, and so on. So, once the initial euphoria fades, I expect we’ll see this stuff become another tool that artists use.
The companies have a vested interest in keeping the hype going as long as possible, so that more and more real artists are forced to quit their career (or their lives) so that people are forced to use the AI art, because there aren’t enough human artists around to actually do the necessary work. That way they can control 100% of the art market, instead of just selling another art tool to artists only.
Thing with AI is they don’t even need be the ultra intelligent genie in a bottle to have an impact. It’s the quantitative potential which will shape the workforce. I’m speaking on behalf of the creatives, can’t speak for IT but I assume it’s similar.
Sorry I had giant spiel I was working on but honestly it just deserves to be developed further into an essay lol.
I’ll just list some points I guess.
-Portfolio based jobs, learn as you go. Professional artists right now with years of work experience should be fine, as the art from ai is just slop once you start paying attention to anything other than just how detailed it is. Symbolism, meaning, telling a story through the scene. Ai can’t do composition, let alone that. Not to mention of the greatest parts of art is the human connection, but still drawing/painting/whatever is hard to make accurate. And so can do that, eventually, but it can fit dirt cheap and doesn’t strike. Even though beginner artists can do the higher level symbolism and composition, where they fall flat is the detail and AI acts a barrier to entry.
-Smaller artist “start ups” if you will have a horrible selection process with how much ai junk you have to soft through. Ai has totally bloated the portfolio process.
-Most main stream garbage is already recycled so AI isn’t even going to change the main stream, but again finding the quality content will be harder and harder.
-Silver lining is petite bourgeois jobs being the creatives and IT (also already left leaning) will finally be losing their special status.
-I think in a socialist state AI art could be kept in as a source of raw material for a giant artist organization to train younger artists to work with, so not only are artists elevated to a more abstracted position in the process, but then we won’t have AI art clogging up every single space.
Sorry I’ve been meaning right around this fir a while so this really just acts as a little brain dump. Also ran out of time lol. I’m a goofball.
That’s how I generally see things as well. I expect, as you said, the biggest immediate impact will be on entry level jobs. I think this dynamic will be similar for programming jobs as well as artists. AI can crap out code right now, but you need somebody with a good general understanding to read it, validate it, and to make sure it fits with the business requirements. A senior developer can use an AI generator to save some time this way, and I can see how at least some entry level coding positions could be replaced in the near future.
However, this creates a long term problem since the only way to get more senior developers is by investing the time into training up junior devs. I can see how once the dust settles jobs for both artists and devs will likely evolve into figuring out how to utilize machine learning models effectively. But there’s gonna be a lot of chaos in the near term I expect.
I do think that once the hype dies down, people are going to sober up and realize that you still need actual human artists in the loop. The problem is that thanks to wonders of capitalism, companies working on AI tech are overselling the capabilities to secure funding, and companies are buying into it because they want to cut employee costs. From what I’m seeing though, the tooling and workflows around stuff like stable diffusion are already getting really sophisticated. With stuff like control nets, selective inpainting, and so on.
In practice this should be looked at a tool akin to photoshop, where you need a lot of skill and training to produce professional results. Anybody can type in a prompt and get a random picture out of it, but in most cases you actually want a specific result with a particular style, lighting, composition, and so on. So, once the initial euphoria fades, I expect we’ll see this stuff become another tool that artists use.
The companies have a vested interest in keeping the hype going as long as possible, so that more and more real artists are forced to quit their career (or their lives) so that people are forced to use the AI art, because there aren’t enough human artists around to actually do the necessary work. That way they can control 100% of the art market, instead of just selling another art tool to artists only.
Unfortunately that’s very much the state of things, and why capitalism needs to be abolished.
Thing with AI is they don’t even need be the ultra intelligent genie in a bottle to have an impact. It’s the quantitative potential which will shape the workforce. I’m speaking on behalf of the creatives, can’t speak for IT but I assume it’s similar.
Sorry I had giant spiel I was working on but honestly it just deserves to be developed further into an essay lol.
I’ll just list some points I guess.
-Portfolio based jobs, learn as you go. Professional artists right now with years of work experience should be fine, as the art from ai is just slop once you start paying attention to anything other than just how detailed it is. Symbolism, meaning, telling a story through the scene. Ai can’t do composition, let alone that. Not to mention of the greatest parts of art is the human connection, but still drawing/painting/whatever is hard to make accurate. And so can do that, eventually, but it can fit dirt cheap and doesn’t strike. Even though beginner artists can do the higher level symbolism and composition, where they fall flat is the detail and AI acts a barrier to entry.
-Smaller artist “start ups” if you will have a horrible selection process with how much ai junk you have to soft through. Ai has totally bloated the portfolio process.
-Most main stream garbage is already recycled so AI isn’t even going to change the main stream, but again finding the quality content will be harder and harder.
-Silver lining is petite bourgeois jobs being the creatives and IT (also already left leaning) will finally be losing their special status.
-I think in a socialist state AI art could be kept in as a source of raw material for a giant artist organization to train younger artists to work with, so not only are artists elevated to a more abstracted position in the process, but then we won’t have AI art clogging up every single space.
Sorry I’ve been meaning right around this fir a while so this really just acts as a little brain dump. Also ran out of time lol. I’m a goofball.
That’s how I generally see things as well. I expect, as you said, the biggest immediate impact will be on entry level jobs. I think this dynamic will be similar for programming jobs as well as artists. AI can crap out code right now, but you need somebody with a good general understanding to read it, validate it, and to make sure it fits with the business requirements. A senior developer can use an AI generator to save some time this way, and I can see how at least some entry level coding positions could be replaced in the near future.
However, this creates a long term problem since the only way to get more senior developers is by investing the time into training up junior devs. I can see how once the dust settles jobs for both artists and devs will likely evolve into figuring out how to utilize machine learning models effectively. But there’s gonna be a lot of chaos in the near term I expect.