55 thoughts on “Pixel Scroll 7/6/23 The Felix Felis Flattus Sat On The Mat”

1. (8) Return of the Caped Crusaders was fun.

2. Boy, this is out hours earlier than I expect.

3. mark: I had to get up for the 6:00 a.m. local time unveiling of the Hugo ballot, and suddenly I had a lot of surplus hours on my hands to finish the Scroll!

4. I got a subscriber notification. I’m thinking of having it framed.

   1) Haven’t we been here before? Mind you, living here in the sunlit uplands of post-Brexit Britain, I am rather inclined to think of 2015 as The Good Old Days.

   The Scalzi hate is baffling. I mean, I’m not a Scalzi fan myself, but what part of “popular writer gets nominated for popular awards” is hard to understand, here?

5. (2) I can’t afford to watch anything, so has the Snake Princess (or is it Priestess?) showed up yet?

6. Wait, nominating a slate for an SF award because the people on the ballot are “too liberal” — I’m sure I’ve heard of someone trying that before.

7. (1) There is some kind of unofficial rule about Twitter users whose avatar is a bust of a Greek philosopher or writer… I didn’t take it seriously at first, but I’m beginning to think that rule has merit.

   Also, I’m getting annoyed with the usual “SF is in trouble because books I don’t like are getting awards…” song and dance routine. Bleh. But I have even less patience for people who toss around words like “woke” or “blue hair” (or worse, “pedo”) in their “criticisms.”

   If you don’t like a book or an author, fine. But this goes beyond hating a book to some kind of infection. (Do they want to borrow my Clotrimazole?)

   Some people really do need to push themselves away from their screens. (If I’m telling you to get away from your screens, then you know you are online too often!)

   (5) If only there were people who accepted money for writing articles. Oh, wait…

8. (1) Gwenna and I were all set to sign up for the Dragon Preservation Society, until we realized it’s only about protecting a popular award from being won by popular authors, and in particular the popular one we like who is inexplicably hated by Certain People.

   (5) LLM-generated garbage that can’t get simple, easily checked details right probably isn’t a good path toward boosting revenue for a site whose users will absolutely notice.

   (6) Given that most material “freely available on the internet” isn’t copyright-free, I’m betting on the authors being right that the content of their books was stolen. Generating accurate plot summaries is a pretty persuasive point, given that LLMs routinely get wrong easily checked details wrong, because there’s no intelligence there to say, “hey, maybe I should double check these facts.”

9. Batman: Return of the Caped Crusaders and Batman vs. Two-Face are fun. With Adam West and Burt Ward reprising their characters the animation style was styled after the 1966 TV Series and the Comic-Book Batman ’66 series. Julie Newmar reprised her role as Catwoman and William Shatner voiced Two-Face.

   I showed both at Balticon in the Video Room and the audience was entertained and appreciative.

10. On (4): These days I read about a lot more new films than I see, but it sure looks sounds the Our Hero(ine) Wanders Among Multiple Timelines bit has already turned into an overworked cliché. Since many of these were in development in parallel, it may be a matter of really unfortunate coincidence rather than copying, but I’m left with little desire to see any of this ilk.

11. (8) Mystery Men series? I know and am very fond of the 1999 film, which featured Geoffrey Rush as the villain Casanova Frankenstein, but I can find no evidence of a series. Which is a pity…

12. Looking forward to new B5.

    Cat Valente was GoH at last weekend’s con and was, of course, delightful. I’d signed up for her kaffeeklatch, but my ride left before that.

    Also, an award being called a “Dragon” and this year featuring a kaiju novel? What part of “DUH” don’t they understand?

    @Anne Marble: What you said.

13. I’m going to pay attention to that class-action lawsuit. I also need to put a notice on my blog concerning what my new publisher’s lawyer said they couldn’t police – that everything on my blog that I write is NOT public domain and anyone training their chatbot with anything on it is plagerism and copyright violation.

    Hey, google, here’s a $1B lawsuit. Ready to talk?

14. Oops, posted in the wrong thread accidentally. Disregard.

15. 8) Birthdays: I first encountered John Byrne as a writer and artist in the late ’70s, and have followed his career since — I’m not sure I rise to the level of “fan” of his, but I certainly like both his art and his writing.

    And yes, I realize that the listings of people’s work in the Birthdays is the product of personal choice and not intended to be complete or even an overview.

    That said, I find the ones given here to be, as a selection…eccentric.

    (As in, if you asked me to name 10 things that Byrne did in his career, I could probably do it in, say, 30 seconds? And my list would not include even one of these.)

16. I’m sorry to report that “frens” has been coopted and is now an alt-right dog whistle. If somebody tells you that Kirby is shaped like a fren, it is one thing. If somebody says “Frens, I have news”, they’re probably a Pepe fan. And yes, this is why we can’t have nice things.

    https://en.wiktionary.org/wiki/fren

17. (6) To win anything in a lawsuit, don’t you have to show you’ve been damaged? If a computer reads your book, what (in dollars) is the damage? How do you put a price on it?

18. (8) I think “Keith Urban” should read “Karl Urban”.

19. (3) I have seen the group effort nominations and their efforts to get all or many members listed now, and something clicked for me now. There is no sense of collective accomplishment here. It goes for Strange Horizons, and arguably was part of the AO3 meltdown too.

    When a sports team wins a trophy, it is the teamthat wins it; not the individual members. Why is that mentality so hard for sf fans to grok? (Well, I guess I know, sf fandom grew up out of a liberal (in the original sense) and individualistic tradition. I just wish we could get some more anarchist collective insight here, in that we can do things as collectives as we can’t do as individuals.

20. 1)Have fun with those/that.

    5)I remember a time when Io9 wasn’t all that full of integrity or accuracy or quality. I think they’re just upset because AI just proved it is capable of churning out IO9 level work.

    8) Love Mystery Men

21. @bill–

    (6) To win anything in a lawsuit, don’t you have to show you’ve been damaged? If a computer reads your book, what (in dollars) is the damage? How do you put a price on it?

    The chatbots are using the authors’ copyrighted material to make money without their consent to use that material. A price gets put on the use of the creator’s (writer, musician, artist, etc.) every single day. The chatbots aren’t reading the books for personal enjoyment. They aren’t “intelligent” in any sense. It’s just an applied statistics model. The people putting the chatbots out there, and often the people using them, are knowingly exploiting other people’s creative work for their own profit.

    You don’t get to just grab other people’s stuff for your own benefit without their permission. And no, buying a single copy doesn’t grant you any rights beyond reading that copy.

22. 1) What is more pathetic, the fact that it is not against the rules or that it is usless to not get a certain work nominated (Scalzi was a finalist last year)

    3) There is a contrast between Strange Horizon and AO3, leaving aside that honoring everyone with AO3 would have been imposible. The question between individuum und collective is perhaps one between western fans and chinese fans, but this is a dificult worldcon, imho. I think there are two problems, one I understand that certain things can make problems for a con, I respect that, but imho Strange Horizon is not unreasonable here. Second problem is comunication, which is a problem for this worldcon, that is not limited to Strange Horizon.

23. 3) I do not think it is unfair to have a limit on how many names are listed for a finalist. We do not expect to display all the names of those who have worked on a movie or TV Series for Best Dramatic Presentation, so I don’t think it is necessary for a Fanzine either.

    Agree with Karl-Johan that it is the team that should be praised.

24. Goldfarb, here’s the deal. You don’t like my choices in the John Byrne Birthday.

    So instead of criticising them, you write him next year. Yes, do some work. You spend a lot of time, kvetching about these Birthdays, so why not instead occasionally do one?

    Filers who are useful write up writers like Robert Heinlein. And they do a brilliant job of it.

25. The output of generative AI [sic] tools can’t be copyrighted, but the output may be identical to (or close enough for legal purposes) a copyrighted work that a reasonable reading of the output would determine that it was, in fact, plagiarism. So if ChatGPT prints the text of an existing story (“In the style of Peter S. Beagle, write a short story about a professor who lives with a rhinoceros. The rhino thinks it is a unicorn. They like to discuss Nietzsche.”), that short story has just been stolen.

    Two additional issues which complicate matters: (1) the LLMs are unable (says the developers) to cite their sources. I call BS on this one, because there are log files somewhere. There are always log files somewhere. I suspect the developers consider those to be their own intellectual property, which is ironic. (2) The tools which are supposedly able to determine if a given text was created by a LLM are worse than the tools which create those texts, and they are in some cases quite problematic.

    All of which is to say, we live in interesting times.

26. 8) My first encounter with John Byrne was his work on Fantastic Four and Alpha Flight (I missed him on X-Men by THIS MUCH). I had no idea that he’d worked on Hellboy — that definitely seems like something I need to track down.

27. 8) John Byrne is one of those artists whose work you really recognized. I’d say The Dark Phoenix Saga is the one I’d really remember him from. I mean, Crisis on Infinite Earths had an enormous influence, but that was mostly because how many comics were affected. It wasn’t that good as a comic.

    Lately, Byrne has been known as a bit of an a-hole – no one likes a Company Man and bigotry is never fun – so I haven’t really tracked what he is working on. And his Superman comic with Big Barda was icky (even if he otherwise was a perfect fit for drawing Superman).

28. @Karl-Johan:

    When a sports team wins a trophy, it is the team that wins it; not the individual members.

    And yet this is not universally true, nor is there an inherent reason that it should be.

29. @Hampus — Did Byrne work on Crisis? I thought that was Perez, but I also wasn’t reading DC at the time. I remember when he came over to DC to pick up the Superman reins — that was a Big Deal at the time.

30. @Hampus @Joe H.
    Byrne went to DC after Crisis by Wolfman and Perez was over. It was the reset of the DC Universe due to Crisis that led to Byrne’s retooling of Superman.

    He did draw Legends, the mini-series that was the crossover after Crisis. I quite liked his art on that at the time(and still do), it had a spark that didn’t seem present on Superman and his Marvel work just before he went to DC. Or maybe it was the inking that was reminiscent of Austin inking Byrne on X-Men.

31. @Stephen Rochelle: It’s not remotely true in cycling, either, although is also (almost always) a team sport.

    @bill: For about the tenth time, no computer is “reading” a book: the text is being statistically analyzed to allow LLMs to generate additional text modeled on the input. Using deliberately misleading terminology doesn’t change this.

32. 6) As we know, chat gtp + co are trained on text and produce an internal statistical model which can be used in a stochastic process to decide which word comes next given a string of prior words. Once the next word is chosen, it gets added to the priors and the process repeats.
    Given this, it seems much more likely that a chat gtp-produced plot summary is going to be based on other plot summaries that formed the training material. After all, it’s not intelligent, and it being able to consume an entire novel and summarise it would be a tremendous achievement in AI which I don’t think we’re anywhere close to reaching.

    That said, I don’t doubt that it and others like it have been trained on copyrighted materials. Some people find this morally troubling: why should the owners of Chat GTP profit off the work of others? Is this very different from a professional reviewer getting paid to comment on the work of others? ChatGTP has not ‘copied’ or ‘stored’ the training set – it would be nigh on impossible for it to regurgitate any given work verbatim because of the stochastic choices it makes combined with the fact that any given prior string has many possible completions. Given this, how different is this from a budding author learning from the works of their heroes and producing their own work that is perhaps too much in the style of, eg, Tolkien?

33. If you’d like to review them for here, I’ll purchase the Kindle editions for you. If interested, say so and I’ll post my email address here so you can contact me off list.

34. Don’t forget all the strips of Funky Winkerbean that Byrne drew over the years (including the last week of the strip before it was retired, after 50 years, late last year).

35. trained on text and produce an internal statistical model which can be used in a stochastic process

    Every one of us will soon give our transformer trained the way we think best the first five installments of Song of Ice and Fire, then let our avatars duke it out on social media with the other authors of Winds of Winter and Dreams of Spring.

36. The first place I ran into John Byrne’s work was in the back of Charlton’s E-Man comics, with his supporting series Rog-2000. Good stuff.

    @PhilRM — I was using “reading” as shorthand. Replace my word with your language, and my point still stands.

    @Lis Carey — I don’t disagree with the bulk of what you are saying, but it doesn’t make much difference to the question I’m asking. Suppose you write a story and publish it. Scenario one — no LLM/AI ever looks at your story or incorporates it into any “fiction” it generates. Scenario two — your story gets scooped up along with hundreds/thousands of others by an LLM/AI, and a work of “fiction” generated by the LLM/AI uses your story as data input. How, apart from you being (legitimately) affronted, is your world any different after Scen. 2 than Scen. 1? Most importantly, can you demonstrate that you somehow have lost money, that court-ordered damages will restore?

    @Stephen Rochelle
    As the wikipedia article you linked to says, “it is the only trophy in professional sports that has the name of the winning players, coaches, management, and club staff engraved on it” — it is the exception, not the rule.
    “nor is there an inherent reason that it should be.” I would argue otherwise — in team sports, many successful teams work very hard to elevate the team over the individual, thus the cliché “there is no I in team”.

37. @bill–

    @Lis Carey — I don’t disagree with the bulk of what you are saying, but it doesn’t make much difference to the question I’m asking. Suppose you write a story and publish it. Scenario one — no LLM/AI ever looks at your story or incorporates it into any “fiction” it generates. Scenario two — your story gets scooped up along with hundreds/thousands of others by an LLM/AI, and a work of “fiction” generated by the LLM/AI uses your story as data input. How, apart from you being (legitimately) affronted, is your world any different after Scen. 2 than Scen. 1? Most importantly, can you demonstrate that you somehow have lost money, that court-ordered damages will restore?

    In the first scenario, my intellectual property is stolen, and I don’t get paid for the use of it in a project that the thief expects to profit from.

    In the second scenario, I do get paid for my work. I’ll decline to speculate on the monetary value of a story I never wrote, but point out that economic decisions like that are made routinely. Getting paid the agreed economic value of a work of fiction, not just for its first publication, but for reprints and adaptations to other forms, is how fiction writers make whatever portion of their income comes from their writing. Theft of their intellectual property impacts them quite directly, especially if they are writers who sell regularly, are supporting themselves adequately but who are not, you know, Stephen King or Nora Roberts. The theft directly robs them of income that would help buy food and pay bills.

38. 1
    As frustrating as I find this on a personal level, on a research into human nature level, the continued obsession with ideological purity in popular awards is indicative. Of something.

    I mean, first there’s the envy factor. Why am I not the recipient of such acclaim and remuneration? Then there’s the umbrage: how dare the majority dismiss my tender moral feelings! Then there’s the appeal to sympathy: join me, pretty please, your support of my biases is what keeps me from having to seek gainful employment. (Or so I imagine.)

    Anyhoo. Have election resentments come to resemble award resentments, or are the latter the traditional emotional pressure valves of the former, and the last several years have seen a failure of the coping mechanism?

    Anyhoo. 5

    I guess I didn’t think humans would sell out their own species quite so quickly. I mean, at least hold out for a simulation of a delicious steak.

39. (1) I may finally have a reason to vote for the Dragon awards – though perhaps in a Null-A fashion (very fannish!)

40. @Lis Carey
    It looks like you’ve reversed the numbering of the scenarios — I’m proceeding on that assumption.

    “my intellectual property is stolen,” — How? (assuming that the copy of your story that was scanned was bought and paid for)

    “I don’t get paid for the use of it in a project that the thief expects to profit from”
    The copyright laws don’t require that you get paid for every use of your work; they only say that you get to control who makes copies of it. Unless the output of the AI/LLM includes verbatim quotes of your story, or is so closely related to your story that it constitutes a “derivative work”, no new copies are made.

    I still don’t see how you or any other author whose works are used as input can document specific damages based on them breaking the copyright laws.

41. 1) I note that in the replies to the tweet in question, some of the tweeters ARE Invoking the High days of the Puppies. So they are well aware of the antecedents.

42. @Brown Robin: It’s mostly envy at base, with a thick coating of whiny snowflake ideology. Their fee-fees were hurt that people who vote for Hugos weren’t down with terrible-to-mediocre RWNJ gun porn with objectified wome, truly bad takes on religion, and mostly White (indeed, “yt” as our friends with melanin say) male, straight-cis heroes. Because that’s boring and their competition writes much more interesting stories, in plot, ideas, characters, and prose quality.

    Also their fee-fees were hurt that they weren’t allowed to hurt other people’s feelings. Also they have a bug up their ass about Scalzi, who DGAF.

    After being roundly rejected by the Hugos, Pups started up the Dragons so they could have an award guaranteed to go to Their Kind. Although as soon as people realized anyone can vote, and it truly became a public award, it turns out that people who vote for Dragons don’t like their crap either. Being SWM, they’re gobsmacked that this could happen. ????

    That and the fact their favorite publisher only has distribution in the US and Canada (plus a few specialty stores overseas), which means the majority of humans haven’t read any of it. Even where they can read English.

    They’re closed-minded, old-fashioned (not popular in a future-leaning genre), and jingoistic (not popular with anyone outside the US, and a number of Americans).

    @bill: The fact that the writer did NOT give ChatGPT or any other LLM permission to copy their story NOR integrate it into their derivative output means it’s defacto copyright infringement. Simple.

    There is no I in team, but there is a me. And meat.

43. @bill–Yes, I did get that confused. Sorry!

    In the first scenario, no one stole my intellectual property. It didn’t get scooped up for the LLM (let’s note, it is in no way, shape, or form an AI. It’s just applied statistics. That’s also why the LLM isn’t “reading” anything.)

    In the second scenario, it DID get stolen–because it was used without anyone either purchasing or otherwise obtaining my permission.

    Note that purchasing a copy of the story to read is not remotely the same as getting permission for use of the story. The LLM isn’t “reading” the story; it’s just scanning it for words and sentences to reuse. Even you have admitted that you know this.

    It’s a content scraper, not a reader.

    “my intellectual property is stolen,” — How? (assuming that the copy of your story that was scanned was bought and paid for)

    Again, to emphasize a point you’ve repeatedly either ignored or badly misunderstood, buying a copy does not get you rights to reuse the material.

    Authors and artists are finding easily recognizable portions of their works in LLMs and the similar visual arts NotAIs. This is plagiarism as lazy and transparent as the bottom feeders on KDP who published “books” with content ripped from books from real writers. They made the mistake of including Nora Roberts among their targets. She sued, because she could afford to and knew most of the lesser-known ripped-off authors couldn’t afford to.

    It’s still theft if your victim can’t afford to go after you.

    Studios get sued when they hijack someone’s work; that’s why most of them have protections in place to avoid doing it even accidentally. They can get sued even when they’ve changed the names and done a halfway decent job of filing off the serial numbers.

    Your certainty that the people behind the LLMs are stealing nothing of any actual value belies the fact that studios purchase options to even consider taking a good look at some for potential adaptation to film.

    They paid Andre Norton for The Beast Master, even though the result was unrecognizable and she insisted her name be removed.

    Authors have actual legal rights to their material, and someone wanting to make commercial use of it needs to purchase the right to do so.

    No matter how shitty you think the material is, how unimportant you think the author is, and no matter how much your notion of economic good relies on the little guys having only the rights the bigger guys decide to permit them.

44. @bill: I was using “reading” as shorthand. Replace my word with your language, and my point still stands.

    Okay, instead of “reading”, I’ll call it “illicit data-mining of copyrighted material for the purpose of intellectual property theft”.

45. 1) This Aristophanes dude…I poked around his website. He spends his free time trying to get ChatGPT to regurgitate white supremacist talking points. Tells us everything we need to know about his world view.

46. @Lurkertype

    The fact that the writer did NOT give ChatGPT or any other LLM permission to copy their story NOR integrate it into their derivative output means it’s defacto copyright infringement.

    Lawsuits go by de jure, not de facto, views of the world.

    There is no I in team, but there is a me. And meat.

    Ironic, isn’t it, that in a discussion of intellectual property, you use someone else’s jokes, without credit.

    @Lis Carey — As you and others have pointed out, all the LLM does is do a statistical analysis of an input text. It does not copy it. There’s nothing in the copyright laws that says the owner of a copyright controls the ability of other parties (or their computer software) to do statistical analysis.

    You keep talking about this as if it is self-evidently “theft” and it is not necessary to show in what way inputting a text to a chatbot or other program is a violation of copyright laws; as if simple “use” of a text is something a copyright owner controls. I think anyone who wants to win a lawsuit like this would have to go through this process, and would have to show specific damages, and not just say “they should pay me”. You keep skipping these steps.

47. @bill: There’s nothing in the copyright laws that says the owner of a copyright controls the ability of other parties (or their computer software) to do statistical analysis.

    That’s because before the deployment of LLMs it wasn’t an issue, so there’s no precedent : this is precisely why Tremblay and Awad have filed suit.

    Banning the use of copyrighted material as inputs to LLMs and forbidding their use to generate work is one of the core issues in the WGA strike.

48. @Cat Eldridge: It’s evident that you felt personally attacked, and I regret that. It’s quite true that you put in the work to do the birthday listings, and I don’t: I at least tried to acknowledge that in my original message. It’s unfortunately the nature of something frequent and daily like the Birthdays that people don’t praise the ones they agree with (the comment sections here would be full of nothing else) but only respond to ones where they feel there’s some issue.

    So let me say definitely: I enjoy reading the Birthdays, and I thank you for the effort that you put into them. I’d be happy to make some contributions; doing a writeup for Kurt Busiek, when September rolls around, is something that comes to mind. Perhaps you could make some suggestions for the nearer future?

    But this one for John Byrne does depart from the usual pattern. The overwhelming majority of the listings make mention of the things for which the listee is best known. To take a few examples, chosen more or less at random (but I did go and look at the listings here)

    Larry Niven’s listing mentions Ringworld and “Neutron Star”;
    Anne McCaffrey’s listing mentions Pern;
    Brandon Sanderson’s listing mentions the Cosmere;
    Ursula Le Guin’s mentions Earthsea and the Hainish novels.

    So then, do you mention Hellboy and Blood of the Demon because these are the equivalent works from Byrne? Of course not, that would be absurd. But the suggestion is created, from the usual pattern, that you do think so; it’s that dissonance which prompted my original post.

49. @PhilRM

    Banning the use of copyrighted material as inputs to LLMs and forbidding their use to generate work is one of the core issues in the WGA strike.

    It is entirely appropriate for private parties to agree to things in contracts that laws do not cover.

50. @bill: It is entirely appropriate for private parties to agree to things in contracts that laws do not cover.

    It is also entirely appropriate for courts to rule that new usages/applications violate existing laws. This is the issue at the heart of Tremblay and Awad’s suit.

This site uses Akismet to reduce spam. Learn how your comment data is processed.