A ton of lower tier artists are going to be replaced by this. I've already seen people say that once they get access to this tool (assuming commercial use is allowed and why wouldn't it) they'll stop hiring artists. I'm an artists but how could I ever say that this technology shouldn't be used if it does what the customer wants faster, cheaper and probably better. There is no right to be an artist and make money with painting pretty pictures. That said, I haven't yet found a clear answer to is what kind of images have they used to train the model with. If they really just scanned instagram and the rest of the internet, what are the legal implications here, or what ought they be? Are they in the process of creating a superhuman art machine that learned from millions of artists just like me without us knowing and now OpenAI launches it as a product to make a profit?
Isn't there already extremely high competition already to create image content and the skill to work with retouch, composition, layout, design, text etc valuable? If you have the expections to input anything and it doesn't get it right you want to some kind of advanced professional tools.
@@hombacom Yes there is and once something like Dall-E 2 hits the market and can be used commercially there will be even more competition because a big part of the market will be satisfied with the results they directly get from Dall-E 2. Those would be low paying customers but that's how you usually get a foot in a professional industry: you take on the low tier jobs and slowly make a name for yourself. Absolutely do I believe that there will remain art jobs that work with Dall-E 2 or similar AI tools but fewer people will be able to make a living. A single artist might be able to fix Dall-E 2 creations for 10 customers in a week so that they're satisfied instead of 10 artists taking a week each to create something from scratch. Very good for the customers and that one artist (if he enjoys that work) but obviously not for the other 9 artists.
@@LarsRichterMedia its still have a lot of errors in many example result images and confused about more complex descriptions so no hurry to expect big change over night but interesting tool to learn from for future.
@@hombacom even though I learned programming myself I recognize that it's a long time ago and I'm not the one who can convincingly project the development and progress of this tech into the future. If you compare the first and the second version though I don't understand how anyone could believe that "magical" progress won't happen. It already did with the second version which is a gigantic leap compared to the first as far as I understand it and that took a year or two. If anything I'll always expect and prepare myself for huge leaps coming rather quickly.
@@LarsRichterMedia it’s mind blowing to experience a unique constructed image of this quality that nobody ever seen before. Even more crazy if it can be close to real-time and you could sculpt anything instantly. But as an artist you can pay attention to details, why is the composition chosen and where do you want the attention to be. Do you want to tell a story or just have something decorative. Can you create a sequence of images that feels both non-repetitive and connected to each other. I hope visual knowledge always will be helpful and the form maybe less important.
Where I see this headed is artists producing original characters and scenes as samples and putting together things like panel comics. You can imagine playing with the iterations and language to craft the flow of a story from scene to scene. Ultimately, human + tool is superior to tool alone.
This is going to sound ridiculous but it needs to be talked about. OpenAI are so focused on wokery and not possibly offending anyone anywhere ever, that with dalle-2 and beyond they will leave a huge sector of the art world still thriving for human artists that still want a job, and that is nude art. Everything from traditional art to CGI sculpted to nude photography, and like I said it's a huge sector of the art world you only have to check one of the major sites such as devianart. Rest assured OpenAI will do their up most to neuter their AI when it comes to this topic. This is really strange because 1. I thought this was original AI art not copying and pasting of real photos. 2. If nude art offends you, either don't ask for it in the prompt or don't look for it.
![Megan Thee Stallion - Mamushi (Remix) [feat. TWICE] [Official Audio]](http://i.ytimg.com/vi/5c0_u9PsrsQ/mqdefault.jpg)
A ton of lower tier artists are going to be replaced by this. I've already seen people say that once they get access to this tool (assuming commercial use is allowed and why wouldn't it) they'll stop hiring artists. I'm an artists but how could I ever say that this technology shouldn't be used if it does what the customer wants faster, cheaper and probably better. There is no right to be an artist and make money with painting pretty pictures. That said, I haven't yet found a clear answer to is what kind of images have they used to train the model with. If they really just scanned instagram and the rest of the internet, what are the legal implications here, or what ought they be? Are they in the process of creating a superhuman art machine that learned from millions of artists just like me without us knowing and now OpenAI launches it as a product to make a profit?
Isn't there already extremely high competition already to create image content and the skill to work with retouch, composition, layout, design, text etc valuable? If you have the expections to input anything and it doesn't get it right you want to some kind of advanced professional tools.
@@hombacom Yes there is and once something like Dall-E 2 hits the market and can be used commercially there will be even more competition because a big part of the market will be satisfied with the results they directly get from Dall-E 2. Those would be low paying customers but that's how you usually get a foot in a professional industry: you take on the low tier jobs and slowly make a name for yourself. Absolutely do I believe that there will remain art jobs that work with Dall-E 2 or similar AI tools but fewer people will be able to make a living. A single artist might be able to fix Dall-E 2 creations for 10 customers in a week so that they're satisfied instead of 10 artists taking a week each to create something from scratch. Very good for the customers and that one artist (if he enjoys that work) but obviously not for the other 9 artists.
@@LarsRichterMedia its still have a lot of errors in many example result images and confused about more complex descriptions so no hurry to expect big change over night but interesting tool to learn from for future.
@@hombacom even though I learned programming myself I recognize that it's a long time ago and I'm not the one who can convincingly project the development and progress of this tech into the future. If you compare the first and the second version though I don't understand how anyone could believe that "magical" progress won't happen. It already did with the second version which is a gigantic leap compared to the first as far as I understand it and that took a year or two. If anything I'll always expect and prepare myself for huge leaps coming rather quickly.
@@LarsRichterMedia it’s mind blowing to experience a unique constructed image of this quality that nobody ever seen before. Even more crazy if it can be close to real-time and you could sculpt anything instantly. But as an artist you can pay attention to details, why is the composition chosen and where do you want the attention to be. Do you want to tell a story or just have something decorative. Can you create a sequence of images that feels both non-repetitive and connected to each other. I hope visual knowledge always will be helpful and the form maybe less important.
Where I see this headed is artists producing original characters and scenes as samples and putting together things like panel comics. You can imagine playing with the iterations and language to craft the flow of a story from scene to scene. Ultimately, human + tool is superior to tool alone.
I think you’re right on that, human + tool = more effective
That is scary, but the way film was scary to a novelist, or streaming to filmmakers.
Good point. I suppose everything had its use case. Maybe what makes art special is the human thought that goes into it.
This is going to sound ridiculous but it needs to be talked about. OpenAI are so focused on wokery and not possibly offending anyone anywhere ever, that with dalle-2 and beyond they will leave a huge sector of the art world still thriving for human artists that still want a job, and that is nude art.
Everything from traditional art to CGI sculpted to nude photography, and like I said it's a huge sector of the art world you only have to check one of the major sites such as devianart. Rest assured OpenAI will do their up most to neuter their AI when it comes to this topic. This is really strange because 1. I thought this was original AI art not copying and pasting of real photos. 2. If nude art offends you, either don't ask for it in the prompt or don't look for it.