Disclaimer: I love MidJourney. It's the most provocative and addictive tool I've played with on a computer for many a year. Largely agree with AW the OP, but t

Disclaimer: I love MidJourney. It's the most provocative and addictive tool I've played with on a computer for many a year. Largely agree with AW the OP, but think the question of "what art are we talking about" needs more air. A lot of the turmoil around these tools appears to be about the threat they pose to the careers of artists and designers (and soon, most related domains). But some of the simple versions of this story (not this one) overlook the nuances IMO. Some of those being, - the distinction between "fine" art in the commercial "art world" gallerist sense, and graphic design/illustration/commercial art - that professional commercial artists are going to adopt and exploit these tools with the same intrinsic advantages they have using with other tools - that in a very short time there is already an explosion of work that uses these tools as one (large) element to produce synthetic work across media that would have been prohibitively expensive/time consuming without them Even as a bystander, but with a background and interest as a serious and modestly successful artist in a different medium, it's been remarkable to see just how quickly people with fewer kids, better ideas, and more time than me, seized upon these tools not as simple image-making widgets, but as sophisticated tools to exploit to generate a stream of imagery to be applied to create Other Things, both as proofs of concept, commercially, and as yes "real art" in the gallerist sense. These things are quite obviously disruptive in a way most VC only dream of. Idle other comments: The "eerie" uncanny sense that scaling up the visual cortex abstraction stack toward "grandmother neurons" that these systems has, is IMO considerably more unsettling than the impact on specific industries. What we can see (literally) in these systems is a visualization (in the data viz sense) of how much further along towards AGI these things may soon be. Their failure modes are more interesting to me than their successes. The ramifications for how they fail and how they are of course unaware of their own failure, is a true cautionary tale. IMO those failures offer an excellent basis for "cautionary tales" and illustration to the general public of why we need oversight and governance for the deployment of these systems. It's one thing for them to produce "nightmare fuel" when asked to create kittens in the kitchen; it's another for a comparable (smaller less tuned closed-source closed-training-dataset...) system to be put in service approving mortgages, assigning credit risks, or (as made ProPublica famous) assessing whether someone is a candidate for bail. (Highly recommended: The Alignment Problem) Anyway. The singularity is already here, it's just unevenly distributed.