Talk:Loab
This is the talk page for discussing improvements to the Loab article. This is not a forum for general discussion of the article's subject. |
Article policies
|
Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL |
![]() | This article is rated C-class on Wikipedia's content assessment scale. It is of interest to multiple WikiProjects. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
I think this should be merged with the article on Stable Diffusion
[edit]It doesn't seem notable enough to have a separate article. 2A02:C7D:B612:AD00:80A3:5124:930B:CC7 (talk) 14:02, 7 September 2022 (UTC)
- I'm the artist and I actually haven't disclosed what tool I've generated the images with, for various reasons. 92.34.149.75 (talk) 19:46, 7 September 2022 (UTC)
- whatever tool you used was trained on Caitlyn Jenner 2603:8001:D300:A631:0:0:0:10D0 (talk) 18:44, 10 September 2022 (UTC)
- Hi there! Since you revealed your connection to Loab, please read our policy on conflict of interest editing if you wish to contribute to Wikipedia. Thanks! A diehard editor (talk | edits) 21:17, 2 October 2022 (UTC)
- Thanks! I haven't edited this article and don't plan on doing so. Only replied with what I have said in other interviews which has now been added I see. 92.34.149.75 (talk) 14:42, 9 October 2022 (UTC)
- Okay, makes sense. Right now, the article just says "unspecified text-to-image AI model", since you haven't specified which AI model you did use. It would also be a violation of WP:OR and WP:RS to say that it was generated by Stable Diffusion, without a source to prove it. A diehard editor (talk | edits) 15:12, 9 October 2022 (UTC)
- Thanks! I haven't edited this article and don't plan on doing so. Only replied with what I have said in other interviews which has now been added I see. 92.34.149.75 (talk) 14:42, 9 October 2022 (UTC)
- I'm the curator, and that's perfectly legitimate. What corporation wants to be adjacent to so much gore? Murder, Inc., maybe. kencf0618 (talk) 11:55, 8 September 2022 (UTC)
- This subject has gone viral and received attention from notable sources including Cnet and Rolling Stone. Notability doesn't appear to be an issue here. Bugfingers (talk) 19:36, 8 September 2022 (UTC)
I’d say no to merging. Subject seems notable in vars. popular media as of Nov. 2022. — Asdfjrjjj (talk) 18:33, 3 December 2022 (UTC)
Derivative works
[edit]I'm not sure how to add this or if I should wait for more works before creating a section, but Chuck Tingle has written a story about Loab. I assume that there will be plenty more works to come! Philpax (talk) 15:58, 9 September 2022 (UTC)
Categorisation
[edit]I have listed this as "fictional character", though not sure if that's the best/correct category. Perhaps it should instead be listed as an artwork? Bugfingers (talk) 14:18, 12 September 2022 (UTC)
this article does a bad job at explaining its subject
[edit]i came from the "See also" on Artificial intelligence art, but even after reading the article i had a very limited understanding of what it was about or how it came about. i then read the PCGAMER article ( https://www.pcgamer.com/ai-image-generaotr-loab-cryptid-supercomposite/ ) and it was a bit clearer (and the pictures helped). But, please, can someone rewrite this article? It is really really hard to understand. Thanks! --95.89.81.91 (talk) 14:38, 3 October 2022 (UTC)
- Thanks for telling us. Someone will eventually get to work rewriting it. Wikipedia is an open encyclopedia and anyone can contribute, so if you have the time, you can try rewriting it yourself. See WP:COPYEDIT for details. A diehard editor (talk | edits) 15:57, 3 October 2022 (UTC)
- I curated the article, and just now got back to it after several months. Laurens Verhagen was spot on, I think, in that Loab illustrates our limited understanding of AI. As with the Flashed face distortion effect it was completely unexpected; phenomenology has outrun theory. kencf0618 (talk) 09:37, 16 May 2023 (UTC)
Lead paragraph of article
[edit]Which paragraph do you think explains the subject best? IP user stated that the article "did a bad job at explaining its subject". I have already changed the lead to paragraph option 2, but I would like additional opinions.
1 | Loab is the name of a character that a Twitter user named Supercomposite claims to have discovered in an unspecified text-to-image AI model. Loab appears as a distressed macabre older woman, and Supercomposite describes it as being an unexpectedly emergent property of the software, discovered when asking the model to produce something "as different from the prompt as possible". |
2 | Loab is a fictional character appearing as a distressed macabre old woman. She was created by an unspecified text-to-image AI model in April 2022 by Twitter user Supercomposite, who described it as an unexpectedly emergent property of the software, discovered when asking the model to produce something "as different from the prompt as possible". |
-- A diehard editor (talk | edits) 22:37, 3 October 2022 (UTC)
- Your re-wording sounds fine to me. SilverserenC 23:08, 3 October 2022 (UTC)
- Taking Supercomposite's research at face value, I don't think we should be defining Loab as a "fictional character" in the WP:FIRSTSENTENCE definition. The Loab Twitter thread definitely feels somewhere on the spectrum from genuine AI image glitch to invented creepypasta, and by describing Loab as a fictional character Wikipedia is leaning more towards the latter, in a way that the main text of the article doesn't support.
- Adding "character" was actually my attempt to clarify the lead, but reflecting on it I think the previous wording of "image" might be more accurate. --Lord Belbury (talk) 08:09, 4 October 2022 (UTC)
- I just restored "fictional character" to this intro. It had been replaced with "an image" and I wanted to make clear this is not a single image, but instead a recurring character. I also removed "distressed macabre old woman" because that description would need attribution to a reliable source. This is a supposedly accidentally discovered character, so we don't know their age or that they in distress. That somebody may find images of old women in distress to be "macabre" is something that needs attribution to someone. Elspea756 (talk) 19:58, 14 October 2022 (UTC)
- Congratulations on spotting that WP:OR violation, regarding the description of "distressed macabre old woman". A diehard editor (talk | edits) 20:05, 14 October 2022 (UTC)
- Thank you. Elspea756 (talk) 20:23, 14 October 2022 (UTC)
- Calling Loab a "fictional character" sounds wrong to me for implying that there is some work of fiction here, when the article text and sources only describe an artist giving a name to a person they perceived as recurring in some images. Is an unexplained portrait a work of fiction? We wouldn't, I think, describe the woman depicted in the Mona Lisa as being a "fictional character". Lord Belbury (talk) 20:18, 14 October 2022 (UTC)
- Instead of "fictional character", how about "generated character"? Elspea756 (talk) 20:23, 14 October 2022 (UTC)
- That seems clearer to me, although I'm not entirely sure that we'd call the subject of the Mona Lisa a "character" either. --Lord Belbury (talk) 11:30, 15 October 2022 (UTC)
- I think we are fine with "character" and "fictional." The New Scientist article uses "character" twice in the title, "Why do AIs keep creating nightmarish images of strange characters? Image-generating AIs seem to produce mythical characters ..." and goes on to describe Loab as "Loab is a slightly different, but equally fictional beast" and ends with a quote from Mhairi Aitken at the Alan Turing Institute in London calling these sorts of images "myths about characters living in cyberspace."[1] Elspea756 (talk) 15:27, 15 October 2022 (UTC)
- That seems clearer to me, although I'm not entirely sure that we'd call the subject of the Mona Lisa a "character" either. --Lord Belbury (talk) 11:30, 15 October 2022 (UTC)
- Instead of "fictional character", how about "generated character"? Elspea756 (talk) 20:23, 14 October 2022 (UTC)
- Congratulations on spotting that WP:OR violation, regarding the description of "distressed macabre old woman". A diehard editor (talk | edits) 20:05, 14 October 2022 (UTC)
- I just restored "fictional character" to this intro. It had been replaced with "an image" and I wanted to make clear this is not a single image, but instead a recurring character. I also removed "distressed macabre old woman" because that description would need attribution to a reliable source. This is a supposedly accidentally discovered character, so we don't know their age or that they in distress. That somebody may find images of old women in distress to be "macabre" is something that needs attribution to someone. Elspea756 (talk) 19:58, 14 October 2022 (UTC)
Use of sources in article, ignoring their skepticism
[edit]I want to note here that this article seems to rely on 1) poor sources whose reliability I am questioning because they are not engaging in fact-checking of the claims of the twitter user who posted these images, and 2) A poor use of other sources, where we have ignored the skepticism in these sources. For example, a more reliable source with a reputation for standards and accuracy like the Rolling Stone article says "Supercomposite claims to have 'discovered' her," but we are not accurately reflecting the skepticism of these sources when we simply write that these images were "discovered with an unspecified text-to-image AI model". I am going to see what I can do to correct this, probably at least adding in a few "claims" and "said" to make clear that these are statements from a twitter account, not findings that have been subject to any fact-checking. Elspea756 (talk) 17:44, 14 October 2022 (UTC)
- Has the phenomenon been reproduced by independent sources? Xxanthippe (talk) 23:51, 25 November 2022 (UTC).
Yes to Stable Diffusion?
[edit]Supercomposite did an interview last month with The Mary Sue. What I'm trying to determine is whether her response to that first question is an acknowledgement that Stable Diffusion was the AI art generator used for Loab. What do y'all think? SilverserenC 04:26, 26 November 2022 (UTC)
Notability
[edit]I think Loab definitely has the required notability in AI art history, at least the meme and the story. The question where research ought to be found is replication: does the phenomenon actually exist? Using which AI art tools? Any actual peer reviewed research would validate the notability. But since that kind of stuff is a slow process, probably this article should be allowed to stay for another, say, 5 years? If by 2027 no research on the subject has been done, it can safely be deemed not notable and put to the trashbin of history where it may belong. --Sigmundur (talk) 08:51, 30 March 2023 (UTC)
- There is no requirement that any Wikipedia article must have academic research done on the topic for it to be notable. SilverserenC 12:42, 30 March 2023 (UTC)
- Memes, too, have their ambit and domain. Research is undoubtedly ongoing, and AFAIK no one has really come up with a hypothesis even. She bears watching. kencf0618 (talk) 09:40, 16 May 2023 (UTC)
- I just added an academic journal reference to the article, so it looks like you predicted it. :P SilverserenC 00:47, 17 May 2023 (UTC)
- Memes, too, have their ambit and domain. Research is undoubtedly ongoing, and AFAIK no one has really come up with a hypothesis even. She bears watching. kencf0618 (talk) 09:40, 16 May 2023 (UTC)
- WikiProject Artificial Intelligence articles
- C-Class Computing articles
- Low-importance Computing articles
- All Computing articles
- C-Class Technology articles
- WikiProject Technology articles
- C-Class software articles
- Low-importance software articles
- C-Class software articles of Low-importance
- Unknown-importance Computing articles
- All Software articles
- C-Class Computer science articles
- Low-importance Computer science articles
- WikiProject Computer science articles
- C-Class Cryptids articles
- Low-importance Cryptids articles
- WikiProject Cryptozoology articles