This Ars Technica article did a good job of answering a lot of the questions I had:
https://arstechnica.com/science/2023/06/nanograv-picks-up-si...
https://arstechnica.com/science/2023/06/nanograv-picks-up-si...
Astrobites articles are usually written by post-grad students working in the relevant area, with the writing level targeted at undergrad astrophysics students. This one was written by NANOGrav collab members, all (or at least most) of whom directly worked on the results that will be all over the press today and tomorrow, and maybe hits a level for an advanced undergrad student or so.
The journal these papers are published in is open access: no paywall.
The Ars article focused on the NANOGrav raft of papers; NANOGrav is just one of the several pulsar timing array collaborations that are in this synchronized un-embargoing of papers. There is substantial overlap between, say, the NANOGrav papers in the Ars Technica article and the joint InPTA/EPTA set which includes https://arxiv.org/abs/2306.16225 https://arxiv.org/abs/2306.16214 and https://arxiv.org/abs/2306.16226 but also enough non-overlap to encourage reading all of them (and to refresh one's foundational knowledge by consulting e.g. Burke-Spolaor et al 2019 "The astrophysics of nanohertz gravitational waves", <https://link.springer.com/article/10.1007/s00159-019-0115-7>).
Unfortunately I find myself griping about a focus on -- to be rude -- only the USA-based-institutions-dominated collaboration (<https://nanograv.org/collaboration/partner-institutions>, although of course there are non-USA and non-North American institutions and researchers in that collab too). Relatively USA-institution-light collabs will chase press attention in different markets and in different ways, of course, but there is scant sign of them in the Ars article (which at least names and links to IPTA) or in the WaPo one (which only lists "Europe, India, Australia and China", not naming the actual collabs).
The inter-collab two year plan on synchronization of paper releases (the "3P+ agreement") itself is something of a story; I wonder if it's going to get much attention in anyone's press.
That said, I don't mind either Ouellette's or Achenbach & Jaggard's reporting. I think they serve amateur enthusiasts well enough that their summaries of the results and science don't need to be nitpicked. Other than that both articles use the unhelpful cliché "fabric of space[-]time" (the latter also has "fabric of the cosmos", "fabric of space and time", because variety is important right?), grrrrrrrr.
They can and do replace submitted URLs under numerous conditions, including not only clickbait but superior articles.
I'll do so here, though being ~15 h after initial submission, it's largely a moot point.
Ranting over poor link quality in the thread itself is not useful in this regard.
Why arent we using midjourney for obligatory space article pictures?
World’s most replaceable job award goes to Astronomy editorial artists
You? I have absolutely no idea why you did not do this and I wish you would! It'd be even better if you could make some kind of "Astronomy.com's RSS feed -> yieldcrv's midjourney prompt" but one step at a time.
To me it’s especially clear that using AI to make art is immoral when you use it for a purpose that you would have previously hired a human artist for. Harvesting work from artists without permission and then using that to replace them is pretty scummy.
Same with Getty’s
is that the approved use then?
Now back to the morality of the Getty and Adobe models: did those artists know their work was going to be used for training AI art models? From what I understand there are many artists who licensed their art to Adobe Stock who are upset that their work is being used for this purpose.
Keep in mind this is a usage of the art that literally didn’t exist when the artists licensed their art to Adobe. I would argue it’s a fundamentally different use case.
and once that’s done all arguments disintegrate as people stop going to artists and photographers anymore either way?
> It's not immoral, it's just [how] economies work.
Do you believe that every aspect of "economies" (or, to simplify the question, let's say just current day capitalism to exclude things like communism and historic economies from the equation) must automatically be moral, or perhaps amoral since you only said "not immoral"?
Seems to me like either a badly thought out claim, or a bad faith argument to justify your main claim about AI art, rather than an actual justification for believing AI art is not immoral.
(Side note - from the POV of my comment anyway, though I suppose technically my comment is the side-note and this is back on the main topic: if you genuinely haven't noticed any debate over the morality or not or AI art, you've not been following AI coverage in mainstream & tech news publications, nor reading the huge number of HN threads where people disagree over whether or not AI image generators training on human art without compensating the human creators should be considered immoral IP theft or should be considered the same as a human studying great artists while working on improving their own art).
Same with Getty’s
is that the approved use then? or does the goal post move
But thanks for assuming that I'm arguing in bad faith despite not making an argument for either side!
Thanks for your contributions
It makes sense you can't clone exact types of art, but mixing and mashing two or more to make something totally new and derivative is fine.
As for AI art specifically, the problem isn’t really that artists are being replaced. The problem that their own work is being used to replace them, and yet they are not getting credited or compensated by midjourney (etc).