Post B85tlXvcpb7lDpTz4i by glyph@mastodon.social
 (DIR) More posts by glyph@mastodon.social
 (DIR) Post #B8206MH9ApkgFt15vM by glyph@mastodon.social
       1 likes, 0 repeats
       
       Somewhat alarmed that the LLM detritus seems to no longer have a toggle to be disabled in the iOS 27 and macOS betas. I hope, but do not expect, that it will come back in the final releases.If I want to still be able to set a timer when I'm cooking with my voice, I probably need to accept a bunch of nonconsensual slop ("writing tools") being shoved in my face in every app.
       
 (DIR) Post #B8206Mh1ccS5Y97lGC by glyph@mastodon.social
       1 likes, 0 repeats
       
       Protesting LLMs by refusing to use any software that includes them feels like attempting to protest the introduction of tetraethyl lead into gasoline by refusing to breathe until everyone stops putting it in their cars. So I am drawing my personal moral lines in such a way that I will probably accept this.But please don't mistake this for excitement about huffing a bunch of vaporized lead.
       
 (DIR) Post #B820KbI4969MsREnRo by lumi@snug.moe
       0 likes, 0 repeats
       
       @glyph i hate how there is this duopoly with phone operating systems and neither option really gives you much freedom. we can't just make an ios distro without the apple spyware and llm stuff. making one for android is ... somewhat possible, for now, but google really wants to make sure we do not have that ability anymorewe should have 1000s of competing phone operating systems, such that they are forced to interoperate and work together. not this shitty duopoly
       
 (DIR) Post #B820v9ZSAYTZcYng9I by glyph@mastodon.social
       1 likes, 0 repeats
       
       @lumi in android, we do have 1000s, with all the budget handset manufacturers, and it is a miserable slog to make software in that environment, which is why nobody does. stochastic competition won't solve this problem, what we need is regulation. a UN agency that develops software in the global public interest, mandates that apple/google accept it, and regulates their developer programs.
       
 (DIR) Post #B820v9ndJpooKXR0Fc by glyph@mastodon.social
       1 likes, 0 repeats
       
       @lumi I mean this is absolute fantasy but so's the competitive environment where 1000s of independent OS providers could exist, so, as long as we're fantasizing, let's dream big
       
 (DIR) Post #B821Cp2jEJWVARv1Iu by lumi@snug.moe
       0 likes, 0 repeats
       
       @glyph i think government regulation could also work here, yeahthough it would have to be done in a way where anyone could make their own from scratch (as in, devices should not be locked down, remote attestation tech should be banned, etc)also abolish IP law and put software freedom into law (as in, if you distribute software, you must make sure someone else can build it themselves, modify it and run it with the same features even if modified)i know, lofty goals, but software freedom and device neutrality should be non-negotiable
       
 (DIR) Post #B822FBMtIQRIwQ2SgK by moses_izumi@fe.disroot.org
       1 likes, 0 repeats
       
       @lumi @glyph You'd love Maemo Leste's NLnet page.Unlike postmarketOS's entry they directly call out Apple and Google's business practices:https://nlnet.nl/project/MaemoLeste/
       
 (DIR) Post #B85tlWIEuY1c9O312u by glyph@mastodon.social
       0 likes, 0 repeats
       
       @dianea I frequently think "the metaphors I am coming up with are way too bleak. surely someone with direct experience with ionizing radiation / lead / biohazards would chastise me for catastrophizing this" and then people with experience in the fields keep showing up and telling me that I'm not pessimistic *enough*
       
 (DIR) Post #B85tlWYtubLuz3qK12 by matt@toot.cafe
       0 likes, 0 repeats
       
       @glyph So from what you've heard and read so far, do you think there's a point where the mental damage, or whatever you choose to call it, from usage of LLMs is irreversible? If one hasn't yet reached that point, does merely stopping usage of LLMs reverse the damage, or does one have to do something more to restore what they had?
       
 (DIR) Post #B85tlWn53sh9h2Te7M by glyph@mastodon.social
       0 likes, 0 repeats
       
       @matt I haven't seen any evidence to indicate that. It's tough to know, because there are hardly any studies of the harmful effects at all, let alone big, rigorous, methodologically sound studies that have large populations involved, which is what you'd need to start making claims like "irreversible".
       
 (DIR) Post #B85tlWzUJkcUJWHYSO by glyph@mastodon.social
       0 likes, 0 repeats
       
       @matt From what I understand, most modern psychology & neuroscience seems to indicate that adults have far greater neuroplasticity than was previously thought, so it seems unlikely that LLMs could do anything that would permanently negate that. I know people who have largely recovered from severe physical TBIs, so getting better from AI psychosis or vibe-sickness seems like it's probably not all that difficult.
       
 (DIR) Post #B85tlXDfT1xj1UusYi by glyph@mastodon.social
       0 likes, 0 repeats
       
       @matt If you want me to give you a gestalt of the anecdotes I've heard, I'd say that there are 4 parts of the damage:1. addiction to the constant validation & stimulation that chatbots provide2. a general sort of hypomania; over-optimism, ambitiously taking on other projects3. this might be the same as 2, but "poor decision making". particularly skipping steps in assessing risks4. skill loss, from delegating whatever intellectual activities you were previously doing to the bots
       
 (DIR) Post #B85tlXSYZfs7lfsllY by glyph@mastodon.social
       0 likes, 0 repeats
       
       @matt 1 isn't legible as damage *until* you stop, but once you stop, it can produce a bit of withdrawal / depression, which the user has to fight through. But particularly if they're committed I haven't heard about this being particularly bad, just … noticeable.2 & 3 seem to go away fast. without the constant validation, critical thinking returns. this can take a little while too.4 is the big one that takes a long time to recover from, the only one that might plausibly be permanent
       
 (DIR) Post #B85tlXh5hdUwUkgNQ8 by glyph@mastodon.social
       0 likes, 0 repeats
       
       @matt but even the skill loss is weird. Like I am "rusty" at a lot of things, and it doesn't feel odd or distressing to me at all that when I pick up the violin every few years, I'm out of practice, because I haven't been practicing. It makes sense. I mean I wish I could hold on to that skill over years of inactivity but nothing about it is incongruent with my self image, just annoying.
       
 (DIR) Post #B85tlXvcpb7lDpTz4i by glyph@mastodon.social
       1 likes, 0 repeats
       
       @matt heavy LLM users who forget stuff, who talk about being "made dumber", feel distressed during the recovery period because they feel like they _have_ been practicing their craft; whether it's writing, coding, art, whatever, they typically won't have realized just how much they've delegated thinking to the bot, and so it is quite distressing to reach for a cognitive skill that you think is very much active, only to find that you're 18 months out of practice, because you actually *are*
       
 (DIR) Post #B85tlutHSTumRDYw76 by glyph@mastodon.social
       1 likes, 0 repeats
       
       @matt people who are deep into dramatic AI psychosis have a similar issue with 2, because even though they have stopped using the bot, they are still holding on to a bunch of ideas, many of which they may not even realize or remember *came* from the bot.