• 0 Posts
  • 89 Comments
Joined 1 year ago
cake
Cake day: June 23rd, 2023

help-circle

  • I’m cynically viewing this as not a positive. I assume this is so they can make pages 2, 3 and so on as spammy as page 1.

    Not at first, obviously. You don’t boil that frog on high heat.
    You throw out a second page with a cute little text ad off to the side, then 1 or 2 at the top, then a mid-page ad. Maybe some suggested content.

    Instead of having to scroll through a page’s worth of ads to get to semi-relevant results with a gem hidden in them, it’ll be a pages worth of ads for your semi-relevant results per page, and maybe what you were looking for 4 or 5 pages in.

    Google used to be good. They ‘know’ what people are looking for. So they’ll probably hire someone familiar with gambling to figure out a minimum dispersion of relevant results on the pages, to keep people using the service and scrolling past ads. … I used to remember this. Variable-ratio reward schedule?





  • When I’ve run out of words to express my horror at the turn of events, or find out that yet another world altering problem is the result of a decision made well before I was born that put profits or power over morals, I fall back to “I just need a time machine and a shotgun.”

    Honestly, I probably wouldn’t just go blasting through the prime directive, but I’d definitely tell Gerald Ford what pardoning Nixon unleashed. And if I didn’t come back to swim in the thriving coral beaches in President Gore’s exclusive economic zone, I’m gonna mentor teenage Ronald Reagan to be a race car driver, and teach 1970’s Trump how stock options work with the hopefulness, enthusiasm and knowledge of a WSB YOLO investor.









  • So - I don’t think Firefox would be generating captions for PDFs on PDF creation.

    But of the major ways that PDF’s do get created - converted from text editors or design software, I know that Microsoft Word automatically suggests captions when the document creator adds an image (but does not automatically apply captions), and I believe that some design software does, as well.

    I think that, functionally, both suggesting captions at time of document creation, or at time of document read are prone to the same issues - that the software may not be smart enough to properly identify the object, and if it is, that it is not necessarily smart enough to explain it in context.
    By way of example, a screenshot of a computer program will have the automatic suggestion of “A graphical user interface” (or similar), but depending on the context and usage, it could be “A virus installer disguised as ___ video game installer.” Or “The ___ video game installer.” Between the document creator and the creation software or screen reader, only the document creator would really know the context for the image.

    Which is all to say that I think that Mozilla has the right idea with auto-tagging, but it will always fail on context. The only way to actually address the issue is to deal with it within the document creation software.
    But I wouldn’t be opposed to ML on those that can auto-suggest things or even critique how content authors write their descriptions.


  • Oh, man. I remember coming into awareness of his movie a bit late, and while I think I watched it, I don’t think I paid attention to it. But catching bits and pieces of Super-Size Me prompted me to watch both Food, Inc, and Fast Food Nation within about a two-week span of each other, and since I saw those videos (early 2009), I’ve not eaten a burger from a fast food place, since.

    Heck, my wife decided not to eat beef a few years ago, and it was really easy to just write it off, since I’d already been removing dubious meat sources from my life. (Or trying to, anyway.)

    Not singularly life changing for me, but definitely added some weight behind decisions.



  • Or… one is bad and the other one is way worse.

    The thing that always amuses me about this is that Iran was a burgeoning liberal democracy until the CIA and MI6 toppled it in 1953, installing a previously overthrown autocrat (overthrown by said forces of democracy), who ruled until 1979, when he was overthrown by religious hard liners, who really only had mass support because the autocrat was too authoritarian.

    And the reason the U.S. and Britain overthrew their democracy? They nationalized their oil industry to give profits back to their people, which entailed taking over refineries and wells ‘owned’ by British Petroleum.
    The U.S. created their own boogeyman in the area because they wanted to give a corporation near-free access to Iran’s oil. Which in turn lead to the oil crisis and instability in the region.
    The U.S. has really got to stop trying to put out fires while covered in crude oil.


  • Ironically, a similarity to a real, live person without an agenda is not a legal problem unless there’s an implied endorsement from the person. (Which I think was one of the goals here.)
    But characters in movies and books are subject to copyright and are considered the intellectual property of the rights holders.

    So like, if I wrote a book about Wolverine and used other Marvel X-men, Marvel could sue the shit out of me. Or if I used AI to create Hugh Jackman (as Wolverine) to endorse my bandaid product line, I could also be sued by both Jackman and Marvel.

    I think it’s obvious here that Sky was intended to represent Sam from Her, and is almost certainly trained on her voice data (which is copywritten). After a few days thought however, I’m less certain of the argument this could be seen as a false endorsement scheme, since Johansson isn’t mentioned anywhere. (Despite the character being solely played by her, and the numerous attempts to have Johansson work on the project in an official capacity.)


  • I’m saying practically any voice with the associated bubbly flirty personality is going to make you think of the movie Her in such a context.

    I don’t know about you, but even a flirty Joaquin Phoenix voice would never make me think of Her.

    But if they’d had a “voice actor” do a spot on impression of Paul Bettany, complete with the little pauses and other flourishes of his portrayals of Vision, I’d think they ripped off the character.

    I think you and I differ there.
    As best as I can figure, you’re stuck on the flimsy excuse from Altman that they hired a voice actor. I see a line of events that points to OpenAI/Altman making a conscious effort to glom onto the Her movie, and specifically, the Samantha character to drum up interest, create viral buzz to enrich further themselves (without compensating anyone involved in the movie), and to try to add a veneer of credibility to a fading trend.

    Slight turn.
    In another life I was a photographer, and one of the things that they do not mess around with is model releases. Any person that appears in your photos that distributed must absolutely have a legal agreement in place. Using someone’s likeness for commercial purposes without consent and/or compensation will get you fucked in triplicate.

    There’s also the moral part of it. Artists know that you don’t rip off artists. Inspired by, sure. But there’s a line, and you don’t cross it. It’s as simple as that.

    Okay, and finally, this is based less on facts I know, and more feelings I have about the situation -
    It’s fucking creepy, dude.

    Okay, so the movie Her - an entire society becomes obsessed with their AI companions and falls in love with them, causing tremendous grief and trauma. And that’s like, what they’re going to lengths to brand but not brand this latest version with? What kind of fucked up things are going on in their heads over there?
    It doesn’t make sense.

    The rundown, again. (Sorry, I like to establish context. Yay neurodivergence.).
    Altman is on record saying that Her is his favorite movie, and that it is a major inspiration to him. One of the reasons he got into this field. He spent 9 months trying to convince Johansson to work with him on this, and lend her name/voice to this latest iteration of ChatGPT. He’s been so focused on getting her to lend his name to this, that he continued asking her to join in on this even just a few days before the announcement, which was like, the 13th. And that’s after she’s already turned him down, so he was just ignoring her boundaries and trying to pressure her…
    On the 11th - two days before this announcement, Altman does a Reddit AMA (which he was doing as part of the 4o press junket) and says that he’d like to open up ChatGPT for personal NSFW usage.

    I mean, everyone is focusing on Her, but we probably should also to be thinking about the Lucy Liu Futurama episode, because… well, I’m just going to say it. I think he already fucked the robot. The line of events from A to B is transparent and fucking gross.
    Not the whole fucking a robot thing - people got needs - but that the likeness is obviously stolen from a non-consenting actress that I’m beginning to believe he’s obsessed with.

    So … yeah… I have all the problems with this. I view concerns over the usage of her voice as immaterial to the usage of the character, and I see an inherent difference between an LLM mimicking a random voice that happens to sound like someone, and this situation, where the voice was clearly created to represent the character, and by extension, the actor that played the character. I don’t think there’s a slippery slope here. Most judges are fairly smart, and will be able to articulate something I (a non legal) took as a given from the outset.


  • No matter what they make it sound like, there’s going to be similarities with the movie.

    I don’t follow.
    They literally disabled the ‘Sky’ voice Sunday night and now users can’t pick a voice that sounds like the character from Her.
    And, mind you, this is not a ‘huh, they sorta sound the same’ this is a ‘they sound very similar, and have the same personality’ situation, in addition to the fact that Sam Altman is on the record talking about being obsessed with the movie Her - which is circumstantial. What isn’t circumstantial is they literally referenced the movie’s name in their marketing materials. Sam tweeted a vague hint, and his colleagues confirmed it. It’s not speculative.

    There’s nothing wrong with leaning into these for advertising purposes.

    Actually, intellectual property theft is either wrong or merely only technically illegal, depending on where you stand on copyright, but it’s still wrong, either way. Then there’s trying to mislead the public into thinking that GPT-4o was endorsed in some way by those involved in the Her movie. A false endorsement is also illegal. So - wrong there, too.
    I’m sure an actual lawyer could find more wrong with it, but just those two things are actual, literal crimes.