I disagree with this in the specific case of Stable Diffusion, like another user said all the images I've seen look both samey and technically inferior (ie misuse of soft lines even on objects with hard edges or a mess of large-scale composition or people with hands with too many or too few fingers), but allow me to put my big-boy austrian pantaloons on and attempt to argue on the harder presumption that Stable Diffusion indeed has few drawbacks.
It might run something like this: The entire point of invention, I think we can agree, is (either to do something new or) to save labor, and thus cost, on something you're doing anyway. In a world without the invention, we get one picture out of each artist per unit time of labor it took them to make the picture. In a world with the invention, though, we get one picture (for free) plus
whatever other thing they decided to do with their labor in that unit of time, meaning their is more value in the world. This is how Capitalism inverted poverty
over the past two centuries despite (or should I say due to) organizing the law "in favor of" property and its owners, especially productive property; under such a system, it's in the producer's best interest to make what the consumers want, and one thing consumers want is plenty, meaning low prices. So capitalist used their resources to develop labor-saving devices and methods, which allowed them to lower prices by firing some of their workers. But, as we saw, those fired workers do not permanently fall out of the bottom of the world, any more than the luddites of England's textile factories did, because the same people who dumped them out onto the curb dump pallets of now much cheaper goods beside them. This is what we in the biz call a price signal
, an indication of the structure of the market's sellers and the valuation of its buyers; a market price of art, or fabric, or whatever else, that's untenably low for a worker in that field is also a indicator of malinvestment
, a message to the worker that labor, the currency of the rarest resource of all, the limited time of a human life, in that field would be untenably inefficient in comparison to all of people's demands that they are not currently getting fulfilled. Thus, they should shift their labor they invest for money to an investment vehicle that will produce more of what people want per unit so the graph keeps going down while they make money.
As a useful side effect, they, too, then get to turn around and enjoy the increased bounty as much as much as anyone else does. A portion of Orwell's otherwise very worker-oriented The Road to Wigan Pier
describes how even the lower class of the North of England, living filthy lives in literally decayed urban projects, can splurge every so often on smart suits and dresses from decent quality outfitters, and how this overjoys them more efficiently than anything else they buy because the tailors of a generation ago simply could not sell them clothes, or the textilers of two generations fabric, they wanted at prices they could hope to afford. Then the prices of automation (rather unceremoniously, it is absolutely true) told these clothesmakers to go find new uses for their dexterous hands, and all of Manchester, including the clothesmakers, could dance in the street in quality clothes on the weekends. Fast-forward to today, after the clothing industry learned that lesson and embraced change, and the whole first world is wearing quality clothes every day, produced by the best paying positions in front of high-tech machines in the third world. (Seriously. As an aside, it's difficult to overemphasize how much "sweatshops" are loved in the quarter of the population where five bucks a day is an average living. Serger lessons are the best charity you give orphans to save them from starving to death under a bridge, as a buddy of mine experienced first hand.) Our artists may no longer be able to spend six hours drawing in exchange for a viable amount of money, but they, just as much as everyone else, no longer have to scrounge up the six hours of their own time or money if they want a picture of something specific for themselves, like they would have to during the time they're not working for others. The whole world is awash in all the art it wants, rather than just what it could afford/spare the time for, while more efficiently engaging the labor it was using before.
In the long run, it's a win-win-win as stark as that poverty graph. The inverse, then, is that we're ultimately losing out significantly if we try to put the brakes on technological progress. Is the transition difficult for the few who find themselves on the wrong end of each instance automation? Absolutely, but it's both not a permanent sentence, so to speak, and is in service of an objectively improved cause for the many, that they too reap the rewards from now that they've joined the many, in addition to now producing labor output no-one had before at the same worldwide food demand as before. The first is something you can't improve otherwise except by growing the population, which unfortunately then counterbalances by worsening the second. More demand for food then means either higher prices, more farmers, or it just pushes the obligation to labor-save down to the farmers themselves. It's not a coincidence that, as prosperity has risen, food production has exceeded population growth while the share of farmers has shrunk; their labor-savings produce our prosperity, which generationally is also their prosperity, since almost all of us used to be them.
Clearly, for wealth to happen, labor-saving must be made somewhere, since this is the only consistent way it has ever happened. California sure got rich off the gold rush, but no other State can adopt that as their strategy. Our only actual choice is to either press labor-saving down to some subsections of the world, usually the most "necessary" ones like agriculture, and ignore the slump into diminishing returns (the descent of the percentage of farmers is indeed slowing), or we can be evenhanded and not declare some sectors too noble and romantic for market-endorsed innovation of technology. My estimation is that the only justifiable choice, in those terms, is to take the side of the fastest labor-saving possible, which concretely means allowing people to try innovating anywhere, and then specifically endorsing wherever they realize they can find success.