When you write about technology—or most subjects, really—there are certain topics that the internet seems to hand down like homework, demanding that Twitter’s vast volunteer army and the semi-professional Substack corps weigh in with a full spectrum of hot takes.
Second paragraph reminds me of an observation made by (I believe) Senator Ben Sasse, which went something along the lines of: individual fulfillment can be obtained through meaningful production, but not consumption.
Which, as an architect, leads me to be somewhat concerned about architecture education in light of this new technology. The ease with which one can generate some ideas for building designs with DALL-E, however superficial, may be too big of a temptation during a formative time during which one should be learning the craft of design through drawing and thinking and drawing and thinking again. That’s when one’s imagination is put to the test, not an AI machine.
It may prove less disruptive for urban planning, but time will tell. Well, what do you think?
It might vary, depending on which design student uses it. I'm thinking of the similarities between AI art and surrealist movement, which used all sorts of "randomizers" such as automatic writing and collage / frottage, etc. Yet, painters like Salvador Dali still went to school and learned how to draw. I read an essay recently which considers AI art just another tool in an artist's repertoire, which might be illuminating for this discussion: https://covidianaesthetics.substack.com/p/kunstlosigkeit-or-fully-algorithmic
Good point, and thanks for sharing that article. Although despite having attended the same graduate institution as the author(s), I fear the reach of my vocabulary pales in comparison and I'm probably not grasping the full nuance of the argument.
Mileage will vary, sure. I'm still casually critical about this, since it reminds me of another trend further in adolescence: the increasing inability of designers to draw anything whatsoever by hand. Not all, of course, (never all), but the trajectory of the...let's say "Average James"...is worth reflecting on.
One productive response, which I don't think is addressed in the Kunstlosigkeit article, would be for designers to do a more robust and critical analysis of how DALL-E actually generates what it does. What happens if we really start playing with...designing...the code. Of this subject I have not encountered much, perhaps because they're buried under all these think pieces about DALL-E and internet "culture". Maybe I'm not looking hard enough.
I think your idea about "designing the code" is a promising avenue for thinking about this. If DALL-E is yet another tool for artists or designers, then the task is to understand it as fully as possible and tweak it according to one's goals. This would also demystify it: Instead of seeing it as a quasi-human intelligence, we could accept it as the mechanism it really is, and then simply use it (as we would any other advanced tool). William's reference to "randomizers" gets at this a bit.
I see the concerns about its impact on architecture (as well as art). It feels like whatever the result is when the dust settles will be more of a change in degree than a change in kind - there are already other ways to unimaginatively generate superficial design ideas using software, maybe this just streamlines them. It feels like the writing equivalent of Googling every concept and linking to the top result instead of reading and researching more deeply. I'm sure it will impact urban planning too, at least the design component (although much of urban planning is not design so I'd expect the impact to be less pervasive).
Questions like what is lost when designers stop knowing how to draw by hand seem like enduring questions in the industrial and post-industrial age. I think the value of pure craft will remain but there will always be new techniques that suggest it's unnecessary. I like to believe that the value of good work (in a holistic sense) will endure as well, even if there are exceptions, but it will be harder to define as straighforwardly as it once was, and there will be a lot of noise to cut through in order to find the signal.
Speaking of craft & AI, it’s plainly obvious to me that the DALL-E images have a similar...let’s call it style. There’s a hauntedness to them, strange dark depths and spooky unexpected curves. Maybe there is something of an uncanny valley thing going on?
Anyway, I think a healthy % of population could identify most of the DALL-E produced images as such without knowing the source. At least for now. I say most because there are a few where, well, a smaller few are pretty damn good.
One thing it surely can’t do: produce a workable master plan. Partly because (human) politics gets involved, but also because it can’t handle the specificity and complexity strictly in terms of visual output. And for that reason I’m inclined to side with you on the degree vs. kind prediction.
Yes, I'd definitely say it's the uncanny valley effect - similar to the AI that finally beat the world Go champions, observers said the moves the AI made seemed alien or counterintuitive. Will probably take us a while to get used to that.
Good point about the inability to produce a master plan. At a certain point, tasks come into play that are so fundamentally human that they are almost inherently non-automatable (in addition to being significantly more complex).
For now, DALL-E seems to be better at producing amusement than any more powerful artistic statement...but who knows, maybe that will change as it evolves.
It seems there is a good deal of concerned discussion about how "AI art is going to replace human artists," but it's all coming from the pundits. Are artists themselves as worried about AI art as the twitter / thinkpiece crowd?
Yes - I agree with Kevin's comment - I think it's more of a talking point than a genuine concern. There have been plenty of technological threats to artists in the past, some impacted how they were paid or eliminated/shrunk certain categories of artistic production but obviously the broader endeavor has continued on.
I tried to allude to that in the post - this is not the most serious threat to artistic production. I saw a tweet (after publishing this) comparing DALL-E to Google image search and I think that's spot on. People who needed a stock image could already grab one online rather than commissioning an illustration, and a Google search could already probably get you closer to what you need than an AI prompt could.
I know you weren’t asking me, but I think most artists are too busy making art to be worried about such things. Similar to writers that proclaim (insert platform here) dead, while everyone else is still working.
Second paragraph reminds me of an observation made by (I believe) Senator Ben Sasse, which went something along the lines of: individual fulfillment can be obtained through meaningful production, but not consumption.
Which, as an architect, leads me to be somewhat concerned about architecture education in light of this new technology. The ease with which one can generate some ideas for building designs with DALL-E, however superficial, may be too big of a temptation during a formative time during which one should be learning the craft of design through drawing and thinking and drawing and thinking again. That’s when one’s imagination is put to the test, not an AI machine.
It may prove less disruptive for urban planning, but time will tell. Well, what do you think?
It might vary, depending on which design student uses it. I'm thinking of the similarities between AI art and surrealist movement, which used all sorts of "randomizers" such as automatic writing and collage / frottage, etc. Yet, painters like Salvador Dali still went to school and learned how to draw. I read an essay recently which considers AI art just another tool in an artist's repertoire, which might be illuminating for this discussion: https://covidianaesthetics.substack.com/p/kunstlosigkeit-or-fully-algorithmic
Good point, and thanks for sharing that article. Although despite having attended the same graduate institution as the author(s), I fear the reach of my vocabulary pales in comparison and I'm probably not grasping the full nuance of the argument.
Mileage will vary, sure. I'm still casually critical about this, since it reminds me of another trend further in adolescence: the increasing inability of designers to draw anything whatsoever by hand. Not all, of course, (never all), but the trajectory of the...let's say "Average James"...is worth reflecting on.
One productive response, which I don't think is addressed in the Kunstlosigkeit article, would be for designers to do a more robust and critical analysis of how DALL-E actually generates what it does. What happens if we really start playing with...designing...the code. Of this subject I have not encountered much, perhaps because they're buried under all these think pieces about DALL-E and internet "culture". Maybe I'm not looking hard enough.
I think your idea about "designing the code" is a promising avenue for thinking about this. If DALL-E is yet another tool for artists or designers, then the task is to understand it as fully as possible and tweak it according to one's goals. This would also demystify it: Instead of seeing it as a quasi-human intelligence, we could accept it as the mechanism it really is, and then simply use it (as we would any other advanced tool). William's reference to "randomizers" gets at this a bit.
I see the concerns about its impact on architecture (as well as art). It feels like whatever the result is when the dust settles will be more of a change in degree than a change in kind - there are already other ways to unimaginatively generate superficial design ideas using software, maybe this just streamlines them. It feels like the writing equivalent of Googling every concept and linking to the top result instead of reading and researching more deeply. I'm sure it will impact urban planning too, at least the design component (although much of urban planning is not design so I'd expect the impact to be less pervasive).
Questions like what is lost when designers stop knowing how to draw by hand seem like enduring questions in the industrial and post-industrial age. I think the value of pure craft will remain but there will always be new techniques that suggest it's unnecessary. I like to believe that the value of good work (in a holistic sense) will endure as well, even if there are exceptions, but it will be harder to define as straighforwardly as it once was, and there will be a lot of noise to cut through in order to find the signal.
Excellent points!
Speaking of craft & AI, it’s plainly obvious to me that the DALL-E images have a similar...let’s call it style. There’s a hauntedness to them, strange dark depths and spooky unexpected curves. Maybe there is something of an uncanny valley thing going on?
Anyway, I think a healthy % of population could identify most of the DALL-E produced images as such without knowing the source. At least for now. I say most because there are a few where, well, a smaller few are pretty damn good.
One thing it surely can’t do: produce a workable master plan. Partly because (human) politics gets involved, but also because it can’t handle the specificity and complexity strictly in terms of visual output. And for that reason I’m inclined to side with you on the degree vs. kind prediction.
Yes, I'd definitely say it's the uncanny valley effect - similar to the AI that finally beat the world Go champions, observers said the moves the AI made seemed alien or counterintuitive. Will probably take us a while to get used to that.
Good point about the inability to produce a master plan. At a certain point, tasks come into play that are so fundamentally human that they are almost inherently non-automatable (in addition to being significantly more complex).
For now, DALL-E seems to be better at producing amusement than any more powerful artistic statement...but who knows, maybe that will change as it evolves.
It seems there is a good deal of concerned discussion about how "AI art is going to replace human artists," but it's all coming from the pundits. Are artists themselves as worried about AI art as the twitter / thinkpiece crowd?
Yes - I agree with Kevin's comment - I think it's more of a talking point than a genuine concern. There have been plenty of technological threats to artists in the past, some impacted how they were paid or eliminated/shrunk certain categories of artistic production but obviously the broader endeavor has continued on.
I tried to allude to that in the post - this is not the most serious threat to artistic production. I saw a tweet (after publishing this) comparing DALL-E to Google image search and I think that's spot on. People who needed a stock image could already grab one online rather than commissioning an illustration, and a Google search could already probably get you closer to what you need than an AI prompt could.
I know you weren’t asking me, but I think most artists are too busy making art to be worried about such things. Similar to writers that proclaim (insert platform here) dead, while everyone else is still working.
Yeah, that sounds about right.
Learned a new word today! Thanks for working “abstruse” in to the article.
Thanks - maybe the first time I've ever used it too.