• 0 Posts
  • 80 Comments
Joined 2 years ago
cake
Cake day: July 11th, 2023

help-circle

  • Schadrach@lemmy.sdf.orgtoMicroblog Memes@lemmy.worldNever Oracle
    link
    fedilink
    English
    arrow-up
    6
    ·
    edit-2
    8 days ago

    Once had Optimum (the ISP) do this to a business I was working with. Just yeeted all their email accounts and their contents and all backups thereof one day. Declared it was because no user on the account was accessing their email via the webmail system so all the email was nuked for inactivity. IMAP and POP do not count, apparently.

    Short version, do not use Optimum unless they are the only option, and if they are the only option seriously consider moving to fix that.



  • See, she belongs firmly on one of the lists of people Trump prefers to appoint:

    1. Republicans with significant media presence, especially social media. Ideally ones who treat themselves like a brand.
    2. People plausibly accused of being Russian assets.
    3. Whoever wrote the relevant section of Project 2025.

    The last one is totally by chance of course, since Trump doesn’t support Project 2025 and knows nothing about it and all. /s



  • f you can’t see everyone as equal then you are beneath us and need to be dealt with in the most severe method possible.

    Does that mean you need to be dealt with in the most severe method possible, after all if you see them as beneath you then you clearly can’t see everyone as equal, by definition?


  • I actually know someone this happened to. Dad came home and midway through changing got a call from his father in law and went to go help him, leaving his concealed carry weapon on the nightstand. Toddler son got a hold of it and killed himself. On the day before his slightly older sister’s birthday.

    Kid was buried on the property, within sight of the front porch. Mom demanded that and then couldn’t handle being there so they moved in with her folks for the next couple of years, and their living room was practically a shrine to the kid.


  • In parallel to what Hawk wrote, AI image generation is similar. The idea is that through training you essentially produce an equation (really a bunch of weighted nodes, but functionally they boil down to a complicated equation) that can recognize a thing (say dogs), and can measure the likelihood any given image contains dogs.

    If you run this equation backwards, it can take any image and show you how to make it look more like dogs. Do this for other categories of things. Now you ask for a dog lying in front of a doghouse chewing on a bone, it generates some white noise (think “snow” on an old TV) and ask the math to make it look maximally like a dog, doghouse, bone and chewing at the same time, possibly repeating a few times until the results don’t get much more dog, doghouse, bone or chewing on another pass, and that’s your generated image.

    The reason they have trouble with things like hands is because we have pictures of all kinds of hands at all kinds of scales in all kinds of positions and the model doesn’t have actual hands to compare to, just thousands upon thousands of pictures that say they contain hands to try figure out what a hand even is from statistical analysis of examples.

    LLMs do something similar, but with words. They have a huge number of examples of writing, many of them tagged with descriptors, and are essentially piecing together an equation for what language looks like from statistical analysis of examples. The technique used for LLMs will never be anything more than a sufficiently advanced Chinese Room, not without serious alterations. That however doesn’t mean it can’t be useful.

    For example, one could hypothetically amass a bunch of anonymized medical imaging including confirmed diagnoses and a bunch of healthy imaging and train a machine learning model to identify signs of disease and put priority flags and notes about detected potential diseases on the images to help expedite treatment when needed. After it’s seen a few thousand times as many images as a real medical professional will see in their entire career it would even likely be more accurate than humans.








  • Schadrach@lemmy.sdf.orgtoFediverse memes@feddit.ukRule
    link
    fedilink
    English
    arrow-up
    1
    ·
    22 days ago

    Convince the admins of the original instance to hand over the community on that instance to new mods, open it back up and let the old mods have their tantrum over on .world? Literally nothing requires there only be one community of a given name across all instances.





  • The ADL considers Pepe a hate symbol, which I agree with is daft but that’s kind of key to their data and they are considered experts in the field by most. They scanned Steam with some automated tool looking for hate symbol images, came up with like a million hate symbols detected. If something contained more than one detected hate symbol, it got counted as however many hate symbols the tool detected (so for example Pepe saluting a swastika would count as a Pepe and a swastika).

    Almost 55% of those were Pepe. The next highest was the swastika at 9%. A literal majority of hate symbols they detected with that tool were Pepes, at more than 5 times the rate of the next most common symbol. It’s literally included to make the problem bigger in the hopes that most readers either won’t look that deep or won’t know what Pepe is.

    EDIT: Another fun one is if you go look at their hate symbol index, about an eighth of two digit numbers are either hate symbols or part of a hate symbol.


  • Schadrach@lemmy.sdf.orgtoTechnology@lemmy.worldItch.io games site taken down
    link
    fedilink
    English
    arrow-up
    12
    arrow-down
    3
    ·
    2 months ago

    I am a big GOG enjoyer myself, but when I need to use steam for anything, I have never encountered such content.

    You’ve never seen a Pepe meme on Steam? I’m not kidding there either - if you dig into that ADL link and follow it to the research, they have a list of top extremist and hateful symbols on Steam and the swastika is number 2 at 9 percent of detected symbols. #1, representing something like 55% of extremist and hateful symbols their automated detector found on Steam was Pepe.

    Perhaps there is such content in private or otherwise not very visible spaces (such as user profiles), where they will not get reported, but that is true for any site with user content. I call BS on this being an issue.

    If you dig into their research, it’s mostly private user groups and profiles. Game discussion pages are moderated by their respective devs or whoever the devs appoint but user groups are moderated by their owner/appointees and user profile pages aren’t really moderated at all unless you’re doing something actually illegal in the US.

    So unless you go looking at the user profile pages of white supremacists, or go searching for white supremacist user groups you won’t run into much of it.