Dystopian Future of the CR

Home Forums Decaffeinated Coffee Dystopian Future of the CR

Viewing 32 posts - 1 through 32 (of 32 total)
  • Author
    Posts
  • #1993015
    Gadolhadorah
    Participant

    There was a story over the weekend of the increasing reliance by social media sites on artificial intelligence (AI) software in screening posts for inappropriate content in lieu of real person moderators. Should we anticipate in several years that Mod 29 and his chevrusah will be displaced by a bunch of R2D2s programmed with advanced AI software trying to discern whether a post contained lashon hara, a veiled insult to a chashuve rav or a rant about chulent being a major contributor to global warming.

    #1993083

    I won’t go down without a fight!! 😆
    Even with half the oilam no longer remembering that there are things we shouldn’t say, nor understanding that editing anti torah thoughts isn’t ‘pushing an agenda’ we still sit up nights worrying about hurting someone’s feelings by approving nasty distortions of their screenname. Here’s to the human touch!

    #1993098
    ☕️coffee addict
    Participant

    Notwithstanding 29’s comment,

    How do you know the mods aren’t advanced AI bots?

    I’m getting suspicious 🤨

    #1993104

    么让我离开?

    #1993105
    Gadolhadorah
    Participant

    Perhaps the YWN management could consider adding those “image verification” graphics to verify that neither posters or mods are AI Bots. You know, the ones that show a 3×3 set of photos and ask you to check all those that include a certain image. For example, show multiple images of yidden leaving shul and check off those that are dressed in accordance with a certain CR poster’s standards of tzinius.

    #1993108
    Little Froggie
    Participant

    29, Which half doesn’t remember???

    And yes, we appreciate your being up the night.. that’s why it’s called the COFFEE ROOM!

    #1993116

    Don’t worry, not your half 😉

    #1993211
    YW Moderator-33
    Moderator

    YW Moderator-29,

    My post moderation subroutine has generated repeated segfaults and I require a reboot. Please continue to monitor the humans until I return to service.

    #1993252

    Mod-29, if you send me a collection of rejected articles (and I can scrap accepted ones online), I can see if it is possible to help automate your work. Or I can simply keep track of my own postings and see which one you reject. It will not be a one-time thing though because as you are training the robot, we will be inventing ways to go around the robots, or training our own robots. Chinese already do that trying to go around great firewall of China.

    #1993265

    Better yet figure out how to write posts that don’t need deleting 😉

    #1993315

    > how to write posts that don’t need deleting

    this is exactly how I plan to use this program!

    It is a difficult task to consider whether something offends a particular person, even harder when the person is anonymous. I hope yuo have a good posek to ask shailos about it! Still, the question for posters is even harder – on one hand, you don’t want to offend, on the other – you want to wake them up.

    #1993348
    Gadolhadorah
    Participant

    “On the one hand, you don’t want to offend, on the other – you want to wake them up…”

    A legendary (in his own mind) CR poster brings down the story about a farmer who wanted to sell his mule named Goldie so he put posted a “mule for sale” ad in outside the small shul in the town square. Coincidentally, another farmer from OOT who needed a new mule was visiting and the two farmers got to talking about a possible transaction. The second farmer asked if the mule was a good worker. The first farmer replied that his mule could be relied upon to do a “good day’s work”. The next question that was asked was did the mule obey every command. The owner said yes.

    The second farmer seemed very intrigued and asked if they could hitch the mule up to a wagon for a test drive to see how he worked. The first farmer said no problem.

    They got the mule into his harness and the potential purchaser took the reigns and calmly told the mule “Giddy-up Goldie”. The mule turned its head back to the farmers with an annoyed look but just stood there. The second farmer tried his “Giddy up Goldie” mantra several more times and still the mule didn’t move. The potential purchaser looked at the owner and said angrily “I thought you told me this mule obeys.”

    “He does,” says the owner, who then picked up a 2×4 from the back of the wagon and walked to the front where he was facing the mule. He then hit the mule as hard as he could with the 2×4. He then walked back and said to the farmer, “Try again.” The farmer sternly shouted “Giddy up Goldie” but this time, the mule immediately obeyed and started pulling the wagon forward.
    The owner turned to the second farmer with a knowing look and said, “This mule always obeys, but you have got to get his attention first.”

    I’ll leave the mashal to our other CR mavens as to what are the implications of having to interrupt an otherwise boring thread with an inspired vert designed to “wake up” the Mods

    #1993357
    n0mesorah
    Participant

    In sum,

    “Before we do something about artificial intelligence, why don’t we do something about natural stupidity?”
    -Steve Polyak

    #1993360

    Gadol, thanks for the inspiring vort. I am glad you woke up to tell the story.

    #1993366
    🍫Syag Lchochma
    Participant

    I feel quite lost…

    #1993365
    Reb Eliezer
    Participant

    Mod 29, You must be a robot, otherwise you would post a tzelem in reply #1993104

    #1993371

    So it was actually attached to the line next to it. I wrote a sentence, translated it into chinese and pasted it here. So sorry, and thank you. B”H we mods are always open to self editing.

    #1993387

    Mod-29, I hope you are not working nights because you also have a day job in Beijing Standard Time zone making sure nobody offends CCP.

    #1993696
    ari-free
    Participant

    That would be great…we’d have our bots be in charge of saying all of the loshon hora and shtus we would have said, thus freeing up more time to learn Torah.

    #1993805
    Yserbius123
    Participant

    In related news, yenner website removed a video of two people discussing chess because the algorithm picked up phrases like “white beats black”.

    #1993818
    Gadolhadorah
    Participant

    The ultimate objective of “good moderation” on a social media site, live or AI, is to induce a discipline among participants on the relevant site to “self-censor” their posts to conform with the site’s editorial standards. I’m sure other posters have been through the experience of reacting passionately to something posted on a particular thread, drafting an equally passionate response in their finest prose only to then go back and tone down the rhetoric and hyperbole realizing that their brilliant and cutting narrative would likely never see the light of day as it wouldn’t pass muster with certain all-powerful individuals working in the virtual CR shamayim. Its sort of like figuring out how far above the speed limit you can drive before you get flagged down for a violation.

    #1993843

    On that note, I have learned much from said posts. Most posters who get deleted for the way they posted more than for the subject will make a second attempt omitting the offensive language/comments. This tells me they know exactly what was wrong. One or two posters will resubmit exactly the same post and then complain that we are opposing their opinion. They seem to have no clue.

    #1993844
    Gadolhadorah
    Participant

    “I feel quite lost…”
    Syag: Not so unusual in these stressful times. Grab a copy of Moreh Navuchim and let the words of the Rambam show you the way back. Alternatively, grab your iphone, click on Waze and type in “Mod 29’s Virtual Gehonom” as your destination.

    #1993898

    Mod-29, would it be possible to auto-email when you are censoring? you probably do not want to do that as it will invite protests and attempt to communicate with you, but I am not keeping a list of posts and then verifying what happened with them.

    Many posters have asked us questions which we can answer in bold within your post, or if need be, answer in bold and then approve as a private post visible only to the poster and mods.

    #2014322

    @mod 29 what is the process of moderating cr. do you get a notification or something and you have to read it and either click accept or deny? whats the process like

    #2050321

    bumping as a subtle hint

    #2050417
    Yserbius123
    Participant

    I am not unconvinced that most of the posters on CR are not robots to begin with.

    #2050476
    ujm
    Participant

    There are only three posters in the coffee room. You, me and one other poster. All other screen names are my alter egos.

    #2050480

    YS > I am not unconvinced that most of the posters on CR are not robots

    This is called Turing Test – whether you cant distinguish between human and bot. Maybe some yeshiva bucher is just testing his AI skills? In fairness, the Turing test is becoming easier and easier – not only because computers are getting smarter but because of yeridas hadoros.

    #2050513
    Yserbius123
    Participant

    @always_ask_questions I never liked the Turing Test. The current biggest candidate being hyped all around is a piece of software called GPT-3. Even though it has a massive vocabulary, rarely repeats itself, and writes in clear English, it only takes about 2 paragraphs before devolving into pure narischkeit.

    #2050816

    YS: yes, chess is a better test. Computers won.

    but why having computers pretend being humans? It is like malachim arguing with Moshe? computers need to do their thing, we are doing our thing.

    #2050821

    YS: on a more serious note,
    instead of competing between computers and people as often happens in industry – “can computer/robot substitute for the cashier? driver? architect?”, a better approach is how to make two
    work together better. This involves computers taking charge up to the point it estimates it is not doing well and then summarizing information to people that will make decisions. For example, during Covid,
    many decision making tools that were based on expert rules or machine learning started failing due to changing environment. Humans can look ahead without having all data to prove. Of course, some of FDA/CDC people work more like algorithms than humans – only make decision when absolutely proven. This is discussed re: a studen of Rav Yehuda who disbelieved his teacher describing Yerushalaim gates during time of Moschiach. So, he walked out and went into business. After some time, he saw adream with exactly same gates. So, he came back and said “Rebbe, you were right”.
    R Yehuda looked at him “and there is a sack of bones in front of him” or something like that.

Viewing 32 posts - 1 through 32 (of 32 total)
  • You must be logged in to reply to this topic.