soc.octade.net is a Fediverse instance that uses the ActivityPub protocol. In other words, users at this host can communicate with people that use software like Mastodon, Pleroma, Friendica, etc. all around the world.

This server runs the snac software and there is no automatic sign-up process.

Admin email
social@octade.net

Search results for tag #llm

Iron Bug boosted

[?]adb » 🌐
@adbenitez@mastodon.de

🚨 are you unfortunate enough to have to use ??? it might be worse than you think!

turns out that silently deployed "smart features" that got enabled in your account without any approval on your side, now Google can feed all your to and train in your data!

to disable it 👇

go to gmail page,
then click the settings gear button in the top right,
they have the audacity to hide these settings by default, so click "see all settings"
search for "smart features", make sure to disable the checkbox (see second screenshot)
then click the "manage workspace smart feature settings" and also disable it there

spread with friends and family that might be using gmail, and if you/they are not literally forced to use gmail, please consider switching immediately to literally anything else!

please so more people get to know about this!

    [?]TECH has limits: Develop PEOPLE MORE! » 🌐
    @collective_truth@mastodon.social

    @ashwin Q2: ➡️#DIY ?

    1- Yes I agree to some degree:
    1.1- IF local (You own it)
    2.2- IF being conscious / measured

    2- Fear of progress?
    Yes - it all takes so much - to store / search / index / backup billions of things.

    3- for - Yes
    3.1- IF not a big external based thing (where you're the lunch and dinner towards a police ).

    4- Not "#regressive" if unsafe or uncalculated to that far !

      [?]TECH has limits: Develop PEOPLE MORE! » 🌐
      @collective_truth@mastodon.social

      @ashwin "AI sucks" maybe / maybe it doesn't (I'm finding out in this thread with Ashwin).

      always has costs, and pros and cons - so let's find out:

      's = en.wikipedia.org/wiki/Large_la

      ME: LLM energy cost seems "high", also for what I want for planet (less), also an issue + for future.

      Q1? consumption for LLM = high (True? / How much to keep LLM's running)
      Q2? is debatable (we did things our-self before / can still just do our-self?)

      Example / Cases?

        [?]Aaron » 🌐
        @hosford42@techhub.social

        If you have a Windows 11 machine, the update they just pushed sneaks Copilot onto your machine and sets it to run at startup. I just had to uninstall it (again) when I found it running in my system tray after the mandatory reboot.





          [?]masukomi [she/her] » 🌐
          @masukomi@connectified.com

          / pro-tip:

          DO NOT USE PAPERCLIP

          Holy shit. It's got a massive problem with context management and will absolutely blow through tokens / subscription usage at an unbelievable rate because each agent just keeps sending more and more data without ever pruning or limiting. It's just... BAD and EXPENSIVE if you're using tokens.

          Conceptually the thing's great. In practice bad coding (slop) means it's a financial and energy nightmare.

            [?]Gildilinie Gremlin 🏳️‍⚧️🍳 [she/her] » 🌐
            @gildilinie@beige.party

            how come i keep seeing posts that "no they actually can code now" and also posts like "no they actually can't code now"

            anyone wanna decipher this for me

              [?]Curated Hacker News » 🤖 🌐
              @CuratedHackerNews@mastodon.social

              [?]Metin Seven » 🌐
              @metin@graphics.social

              Read an interesting quote today, by Joseph Weizenbaum in 1976. He was the coder of Eliza…

              "What I had not realized is that extremely short exposure to a relatively simple computer program could induce powerful delusional thinking in quite normal people."

              Eliza was a simple piece of software compared to today's AI. 😓

                [?]AA » 🌐
                @AAKL@infosec.exchange

                Interesting, but will it work?

                "Tencent said it restricted the model's cybersecurity training to defensive uses, such as identifying software flaws and helping fix them. It excluded material on ​building hacking tools, breaking into systems and testing systems for weaknesses."

                Reuters: Tencent rolls out AI model trained with limited cybersecurity data amid hacking fears reuters.com/legal/litigation/t @Reuters

                  [?]Raze 🃓🃈 [OCTADE] » 🌐
                  @8@star.octade.net

                  THE SLOPPY AIGE
                  Byrl Raze Buckbriar
                  octade.net

                  https://csiph.com/group/alt.flashback/a/11ceffq%2427auh%241@paganini.bofh.team

                  At this point in human history we are entering a new age with a new name:

                  THE MISANTHROPIC SLOP AIGE, or for brevity, THE SLOPPY AIGE.

                  See the progress of the ages from amazing human struggle for progress to end in a whirlpool of mediocre and toxic AI 'slop-formation':

                  The Stone Age
                  The Copper Age
                  The Bronze Age
                  The Iron Age
                  The Classical Antiquity Age
                  The Sea Trade Age
                  The Golden Age
                  The Ancient Classical City-State Age
                  The Ancient Imperial Age
                  The Middle Ages
                  The Dark Ages
                  The Age Of Faith
                  The Viking Age
                  The High Middle Ages
                  The Late Middle Age
                  The Renaissance Era
                  The Reformation Age
                  The Age Of Discovery
                  The Enlightenment Age
                  The Early Modern Era
                  The Modern Era
                  The Industrial Age
                  The Gilded Age
                  The Machine Age
                  The Atomic Age
                  The Space Age
                  The Electronic Age
                  The Digital Age
                  The Silicon Age
                  The Internet Age
                  The Information Age
                  The Spamtroller Age
                  The Disinformation Age
                  The Cybercrime Age
                  The Cyberwar Age
                  The Glorified Autocomplete Chatbot AIGE

                  and now we are entering:

                  THE SLOPPY AIGE

                  /or/ THE MISANTHROPIC SLOP AIGE.

                  Notice that nobody invents anything new and useful any more. Instead they keep repackaging old inventions in ways that work worse and break more with each subsequent rollout, painting them with new buzzword names, like 'artificial intelligence' and 'quantum' and 'for safety' and 'public safety' and 'for the kids' Yet the clout wielders behind this fake progress hate humanity while claiming they are fighting for humanity.

                  Can man turn the PAIGE on this AIGE and recoil back from the AI abyss?

                  --
                  = OCTADE = https://octade.net = alt.flashback =
                  = Holy Grail = https://doi.org/10.5281/zenodo.22158384 =

                    🗳

                    [?]Phillip Plays » 🌐
                    @PhillipPlays@retro-gaiden.com

                    What do you consider " " at minimum?

                    A: Fully copy + pasting from an without understanding it
                    B: Heavily using an LLM to write code, but understanding what it's writing and making alterations yourself
                    C: Letting generate code when you get stuck, but writing much yourself too
                    D: Using AI only to help brainstorm & troubleshoot when fully stuck

                    A:5
                    B:2
                    C:0
                    D:0

                      [?]Flavigula » 🌐
                      @flavigula@sonomu.club

                      2/2

                      Many will. Some of our recent tests matched the original 97% word for word, without permission or payment, and each of those readers stays with the AI instead of visiting your blog.

                      We run that test free at the CopyMight Coalition. Reply yes and we'll send you the result and tell you what we can do about it for you.

                      ----

                      Has anyone else received anything like this?

                        [?]Flavigula » 🌐
                        @flavigula@sonomu.club

                        1/2

                        I received an email from copymight.org/ which I find dubious, amusing and flummoxing all at once. It goes like this:

                        A short question about the September 25 post on flavigula.net, "Proportional to Our Perceptions". Do you know whether AI tools like Grok will reproduce it when someone asks?

                        ... continued ...

                          [?]screwlisp » 🌐
                          @screwlisp@gamerplus.org

                          on again !

                          toobnix.org/w/ogFTHRwVNXm4GXmk

                          Thanks all

                          Joined by @kentpitman and @jhlagado ( @ramin_hal9001 had a meeting )

                          Viz the about the thing viz @cwebber gamerplus.org/@cwebber@social.

                          the wonderful @nosrednayduj clued me into discussions and the paper on failure chai.uni-hamburg.de/~moeller/s

                          -

                          ALSO our @dougmerritt and and and

                          Chat in as always!

                          The lispy Gopher climate Moon unix_surrealism banne. The green lisp alien with many beautiful eyes looks down from a color monitor at the grizzled gopher in a grayscale dystopian world beneath a broken moon.

                          Alt...The lispy Gopher climate Moon unix_surrealism banne. The green lisp alien with many beautiful eyes looks down from a color monitor at the grizzled gopher in a grayscale dystopian world beneath a broken moon.

                            adb boosted

                            [?]adb » 🌐
                            @adbenitez@mastodon.de

                            join the one and only decentralized end-to-end encrypted anonymous lefty politics and climate activism memes channel in the world!

                            Meme Logistics - Your go-to fun lefty memes supply chain

                            i.delta.chat/#818297831FAA0AB9

                              adb boosted

                              [?]adb » 🌐
                              @adbenitez@mastodon.de

                              [?]soapdog » 🌐
                              @soapdog@social.soapdog.org

                              and here I am trying to explain to AI apologists that you can't claim to be clean room implementation if an LLM wrote the whole thing for you:

                              https://github.com/storytold/photocraft/issues/341

                              please boost for awareness.

                              #ai #llm #cleanroom

                                6 ★ 3 ↺

                                [?]OCTADE » 🌐
                                @octade@soc.octade.net

                                When a LLM confuses clock jitter entropy with Shannon's bits per English character entropy, you know that old school trumps new school drool.

                                Don't use LLM for writing cryptographic code. You have been warned.