3 ms·

> No-one expected creative AIs to come for the artist jobs first, but here we are! Maybe that's because we never really thought about it. In hindsight, it's on

> No-one expected creative AIs to come for the artist jobs first, but here we are! Maybe that's because we never really thought about it. In hindsight, it's only logical. For an artistic rendering, correctness doesn't matter much nor does understanding the model. For flying a plane or driving a car or just transforming code from one language to another, it very much does.

astrange 4y ago
There are some kinds of “correctness” that matter. When you’re doing something new, you can only really do one new weird thing at once. Too many weird things and it falls apart or distracts the viewers. (The “cow tools” Far Side comic is a dumb example.) Since these AIs can’t count, that extra weird thing is probably going to be people with too many fingers. I actually have a personal collection of weird Midjourney images because if you ask for wide aspect ratios it starts generating multiples of the same thing and fusing them together…
abigail95 4y ago
Surely the best people to create new art from models like this would be artists themselves? Wouldn't this create more jobs for artists? It's just a different tool
- TulliusCicero 4y ago
  That depends on how easy it is to use the new tools. Getting higher level languages and various libraries/frameworks didn't make coders obsolete, because there was still tons of work to do at a "higher level" that required algorithmic/engineering-style thinking. Yes, previously difficult tasks got a lot easier, but that just meant that even more ambitious projects could be tackled, and that was useful. In contrast, it may be the case with art that the things a relative newbie can spit out with the help of image generation models are "good enough" for most use cases. Yes, a real artist would be able to squeeze out even more, but many companies may be willing to go with the option that's enormously cheaper and still seems to produce okay results.
  - selestify 4y ago
    Is the same not true for programming? There's a lot of non-programmers that create programs in Excel, or use no code solutions. A relative newbie can still get a Wordpress website up and running, you only need professional SWEs for the heavy duty stuff.
    - roca 4y ago
      For better or worse, you very very quickly pile up complexity when programming in a way that novices can't deal with (and experts too, often). What we're going to find out is whether novice users of image generator AIs frequently get stuck with inadequate results. IMHO the answer is likely to be "no", especially after we've iterated on the tools for a few more years.
    - TulliusCicero 4y ago
      Agreed, but the rub is that there was a ton of demand for "heavy duty stuff"; indeed, the improvement in code tooling actually increased that demand, because software programs became so much more powerful. It's not clear that you'll see something similar here, that easier-to-produce high complexity art will greatly increase demand for said art.
noneeeed 4y ago
I don't know about that. A lot of the images I've seen SD generate have been interesting but "incorrect" in really important ways that indicate it doesn't _understand_. Things like people's limbs not being connected correctly, or three legs, or a wheel on a car not being in the right place, or elements of perspective that are just off. All things a real artist would never do unless it was intentional. It's like some of the stuff that GPT-3 generates, it sounds extrememly plausible and realistic but then you read it properly and realise that parts just don't make any sense. However I don't think these things are show-stoppers, I think we'll see a lot of artists getting good at providing the "feedstock" and guidance to these systems to deal with those quirks and generating some really interesing work.
evouga 4y ago
Agreed. Symbolic AI dominated the field for 50 years: thanks partially to an accident of history (the highly influential work of Minsky, Newell, etc) and partially due to lack of the data needed to try anything else. Now we’ve seen essentially 50 years of research on data-driven methods compressed into 5. In retrospect it makes sense that applications to which symbolic reasoning is especially ill-suited (including artistic rendering) would be the lowest-hanging fruit.
SyzygistSix 4y ago
It isn't coming for the artists' jobs. Commercial illustrators sure. Francis Bacon has a quote that is apt about how he paints the things that cannot be described with words.
- drewbeck 4y ago
  A lot of artists work as commercial illustrators. In fact most artists don’t have artist jobs at all cuz there’s no money in it!
- alxlu 4y ago
  A lot of artists make a living doing commercial illustration.

astrange 4y ago

There are some kinds of “correctness” that matter. When you’re doing something new, you can only really do one new weird thing at once. Too many weird things and it falls apart or distracts the viewers. (The “cow tools” Far Side comic is a dumb example.) Since these AIs can’t count, that extra weird thing is probably going to be people with too many fingers. I actually have a personal collection of weird Midjourney images because if you ask for wide aspect ratios it starts generating multiples of the same thing and fusing them together…

abigail95 4y ago

Surely the best people to create new art from models like this would be artists themselves? Wouldn't this create more jobs for artists? It's just a different tool

TulliusCicero 4y ago
That depends on how easy it is to use the new tools. Getting higher level languages and various libraries/frameworks didn't make coders obsolete, because there was still tons of work to do at a "higher level" that required algorithmic/engineering-style thinking. Yes, previously difficult tasks got a lot easier, but that just meant that even more ambitious projects could be tackled, and that was useful. In contrast, it may be the case with art that the things a relative newbie can spit out with the help of image generation models are "good enough" for most use cases. Yes, a real artist would be able to squeeze out even more, but many companies may be willing to go with the option that's enormously cheaper and still seems to produce okay results.
- selestify 4y ago
  Is the same not true for programming? There's a lot of non-programmers that create programs in Excel, or use no code solutions. A relative newbie can still get a Wordpress website up and running, you only need professional SWEs for the heavy duty stuff.
  - roca 4y ago
    For better or worse, you very very quickly pile up complexity when programming in a way that novices can't deal with (and experts too, often). What we're going to find out is whether novice users of image generator AIs frequently get stuck with inadequate results. IMHO the answer is likely to be "no", especially after we've iterated on the tools for a few more years.
  - TulliusCicero 4y ago
    Agreed, but the rub is that there was a ton of demand for "heavy duty stuff"; indeed, the improvement in code tooling actually increased that demand, because software programs became so much more powerful. It's not clear that you'll see something similar here, that easier-to-produce high complexity art will greatly increase demand for said art.

noneeeed 4y ago

I don't know about that. A lot of the images I've seen SD generate have been interesting but "incorrect" in really important ways that indicate it doesn't _understand_. Things like people's limbs not being connected correctly, or three legs, or a wheel on a car not being in the right place, or elements of perspective that are just off. All things a real artist would never do unless it was intentional. It's like some of the stuff that GPT-3 generates, it sounds extrememly plausible and realistic but then you read it properly and realise that parts just don't make any sense. However I don't think these things are show-stoppers, I think we'll see a lot of artists getting good at providing the "feedstock" and guidance to these systems to deal with those quirks and generating some really interesing work.

evouga 4y ago

Agreed. Symbolic AI dominated the field for 50 years: thanks partially to an accident of history (the highly influential work of Minsky, Newell, etc) and partially due to lack of the data needed to try anything else. Now we’ve seen essentially 50 years of research on data-driven methods compressed into 5. In retrospect it makes sense that applications to which symbolic reasoning is especially ill-suited (including artistic rendering) would be the lowest-hanging fruit.

SyzygistSix 4y ago

It isn't coming for the artists' jobs. Commercial illustrators sure. Francis Bacon has a quote that is apt about how he paints the things that cannot be described with words.

drewbeck 4y ago
A lot of artists work as commercial illustrators. In fact most artists don’t have artist jobs at all cuz there’s no money in it!
alxlu 4y ago
A lot of artists make a living doing commercial illustration.