Jump to content

Jason

Free Account+
  • Posts

    2,345
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by Jason

  1. 9 hours ago, jeffthehat said:

    The problem is that it is undectable. Algorithms can identify language that is consistent with AI-generated output. But they're prone to false positives (e.g., OpenAI's text classifier page shows sections of Don Quixote being flagged as AI-generated). And how can you stop a writer from using an LLM to generate the broad outline of a plot or a character? It will still be used even if it is banned in name. The productivity gains are too significant to ignore on an individual level. Those who aren't using it are going to lose out to those who are. The WGA's stance seems rose-tinted and doesn't help them from a negotating standpoint, I think.

     

    8 hours ago, keysersoze123 said:

    Again how can one guarantee that LLM was not used. Its not just ChatGPT. There will be many others which will pop up. Facebook has Llama which is open source. Its possible to compile and run it locally on your machine. There will be other models as well. 

     

    There seems to be a mistaken impression in that the WGA is advocating for a blanket ban on AI/LLMs. It isn't.

    From: WGA Would Allow Artificial Intelligence in Scriptwriting - Variety
     

    Quote

    The WGA proposal states simply that AI-generated material will not be considered “literary material” or “source material.”
     

    Those terms are key for assigning writing credits, which in turn have a big impact on residual compensation.


    “Literary material” is a fundamental term in the WGA’s minimum basic agreement — it is what a “writer” produces (including stories, treatments, screenplays, dialogue, sketches, etc.). If an AI program cannot produce “literary material,” then it cannot be considered a “writer” on a project.


    “Source material” refers to things like novels, plays and magazine articles, on which a screenplay may be based. If a screenplay is based on source material, then it is not considered an “original screenplay.” The writer may also get only a “screenplay by” credit, rather than a “written by” credit.


    A “written by” credit entitles the writer to the full residual for the project, while a “screenplay by” credit gets 75%.


    By declaring that ChatGPT cannot write “source material,” the guild would be saying that a writer could adapt an AI-written short story and still get full “written by” credit.

     

    So "undetectable" usage of AI-generated material by a writer isn't an issue for the WGA. What they are arguing against is the AI can't be treated as a writer by studios - that is, studios can't present a writer with material from AI and tell them to edit it, and then screw them out of a writing credit to save money.

    • Like 7
  2. On 9/19/2022 at 1:38 PM, Daxtreme said:

    I have 45 Hong Kong movies in my list lmaooooo

     

    I wonder which foreign movies I watch the most... :sparta:

     

    South Korea 2nd with 13


    I'd love to see your list. (PM if you don't feel like sharing to the thread?)

    Haven't quite found the time yet to get serious about it but I've been trying to watch more Cantonese-language films in the hope it might improve my Cantonese.
     

    • Like 1
  3. My dream home is just any home right now because affording a home in Toronto is a dream. T_T

    There are houses on my street smaller than the one pictured above that have sold for ~$6M. 

    (Past two years have seen year-over-year increases of 30%+. Market has gone completely batshit here.)

    • Sad 2
  4. 3 hours ago, Shanks said:

    This newest change to BOT forum's Profile, is this experimental or is it permanent? It is way better to have REPUTATION + Comments shown instead of number of posts -  easier to identify a person if he is regular troll or not.


    I imagine someone has mentioned this already, but at least when browsing on PC you can still see reptuation just by hovering your cursor over the person's name or avatar.

    Would be better to have it visible but yeah, it's a forum software thing as mentioned above.

  5. 30 minutes ago, Napoleon said:

    I support the changes the Academy is making. People need to be realistic about this. The ratings have been declining, the Oscars are becoming irrelevant. Without changes, it will soon have the winners announced via press release, because they won't have the money to fund a ceremony no one watches anymore. Movie fans have been mad about changes in the show that don't actually make the Oscars lose its purpose of honoring the best of the year in filmmaking.


    I don't think the changes in question are likely to compel general audiences to watch the awards. If anything, replacing actual awards with other content is likely to hurt, because the point of watching an awards show, is you know, to see awards. If you wanted comedy etc. you'd just go watch that instead.

    Truth be told, most of the forces driving down viewership are outside the Academy's control. ABC should have a simulcast (with ads of course) be made available, because there's a whole generation of people out there now that don't have cable and sure as hell aren't going to get it for ABC's sake.

    As for the funding of the ceremony, it's a pretty small portion of the Academy's expenses. For example, in 2020 they spent $24 million total on all expenses specifically related to the Academy Awards; the revenue from their contract with ABC was a total of $128 million. They could easily continue to fund the Awards ceremony itself even with a dramatic reduction in revenue from the Awards; the largest chunk of the Academy's expenses ($50 million a year) is actually used to fund the Academy Grants program.

    They've also been running a surplus of around $30 to $40 million dollars a year for the past several years, that they've been using to build an investment portfolio, now worth over $500 million.

    AMPAS is not going to be in a position where they can't continue to support the Awards themselves. The risk would be having to make cuts to other expenses that I'm sure are very important to them, such as the Grants program.

    (Edit: all figures are from the publicly available tax filings of AMPAS; as a non-profit it is required that they be made publicly available.)

    • Like 4
  6. 1 hour ago, The Panda said:

    Most Valuable Poster

    Blankments

     

    Cap

     

    Charlie Jatinder

     

    Jason

     

    Lorddemaux

     

    The Tracking Team

     

      Reveal hidden contents

     


    Thanks everyone! It's an honour to be included among the above company.

    The BOT community is great, and the mods were very supportive. I'm glad to have been able to be of some help.

     

    1 hour ago, Eric Says Trans Rights said:

    Tracking Team, we'll get it next time. 

     

    I'm optimistic that you-know-what doesn't have any more major tricks up its sleeve (fingers crossed), so should be wide open for you next year!

    • Like 6
  7. As one of the #1's 

    10 minutes ago, Cap said:

    #49

    17 POINTS, 3 LISTS, Av PTS 5.66, HR 1

     

    Shizuku / Seiji

    from

    Whisper Of The Heart

     

    7204744cafee770c92416b35b734577a9995e3ca

     

    Screen Shot 2022-03-01 at 12.46.16 AM.png

    Prompt: Shizuku gets into a heated argument with someone.  Shizuku begins threatening them, so Seiji, picks up Shizuku and carries/drags them out of the room before anyone gets hurt.

     

    Fascinating data from the List on this pairing.  They received TWO number one placements, BUT both of those placements were on Lists with only 20 Ships.  So this could've been even higher.

     

    AO3 TAG


    I totally could have had a list of 40 too. Didn't think I had that watched that many until the list reveal and I'm like oops...oops...oops.

    • Sad 1
  8. This was quite some time ago - pre-pandemic - but I still feel worth it's worth noting that back when I was playing Derby (2016-17) I looked at a lot of forecasts, and the Box Office Pro forecasts were, on average, the closest to actuals among publicly available forecasts.

    (The closest possible forecast over time was actually generated by a weighted average of the forecasts from different sites, but that's because there's a portion of the error in forecasting that is random - and therefore expected to average out to zero in a sufficiently large sample. Of course, that doesn't work for generating a de novo forecast. In a similar vein, an average of all Derby forecasts was almost always better than any individual forecast.)

     

    Granted, I haven't done this type of detailed mathematical analysis for a while (and definitely not since the pandemic made everything go to shit) but I still feel confident that anyone thinking the Box Office Pro forecasts aren't of least comparable accuracy to the others available is selectively remembering specific examples that weren't as accurate, rather than doing any kind of comprehensive comparison. (And apparently not too many forecasters were even trying in the pandemic era!)

    (cc: @Shawn)

    • Like 8
  9. 1 hour ago, Cap said:

    Y'ALL WE GONNA LEAVE THIS OFF THE LIST?  RICK AND EVIE?

     

    I'm sorry, this is definitely one I would've put on my list if I'd remembered.

    I am the wooooorst at remembering what movies I've seen without a prompt. Straight up when I made my list I combed through everyone else's lists and FYC's for reminders. Oh, and I looked at my top animated list since I had that ready already.

    So I left a ton of deserving couples off my list because I forgot T_T.

    • Like 1
    • Thanks 1
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use and Guidelines. Feel free to read our Privacy Policy as well.