@Zeugs@social.cologne
@Zeugs@social.cologne avatar

Zeugs

@Zeugs@social.cologne

Stuck somewhere between humanities, computers and data.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

Alexlee, to academicchatter
@Alexlee@sciences.social avatar

Nobody told me about the snakes! @academicchatter

Zeugs,
@Zeugs@social.cologne avatar

@Alexlee @academicchatter
The interesting part is that without AI it shows this as one of the answers to this or similiar questions, referencing:
https://www.mcsweeneys.net/articles/faq-the-snake-fight-portion-of-your-thesis-defense it
The classic Google search does not get the irony either but you also don't expect it to.

pallenberg, to microsoft German
@pallenberg@mastodon.social avatar

Weil sich offensichtlich einige ueber von aufregen. Fair enough!

Laeuft lokal (deshalb ja auch die geforderte NPU, die die OEMs einbauen muessen) und soll(!) nicht von Microsoft abgerufen werden.

Genau in dem Moment muss sich jeder ueberlegen ob er das nutzt oder nicht. Ich kann mir durchaus Anwendungsszenarien vorstellen.

Wuerde ich es nutzen? Nein!
Sollte Microsoft hier auf Opt-In switchen? Ja!

Zeugs,
@Zeugs@social.cologne avatar

@pallenberg
So verschiebt man den Diskurs. Eine vollkommen aus dem Ruder gelaufene Idee präsentieren und darauf warten das alle anderen eine Mittelposition einnehmen. Diese " neue Mitte" ist dann der neue Ankerpunkt. Diskussion erfolgreich Richtung überwachung verschoben.

Zeugs,
@Zeugs@social.cologne avatar

@pallenberg es geht um Diskurs Verschiebung keine Ahnung was daran verschwörungstheorie ist. Es ist auch keine Aussage drin wer was weiß.
Am Ende wird Microsoft irgendwo ein stückchen zurück rudern und es als Einsicht verkaufen.
Auch wenn du die Auffassung nicht teilst geht deine Antwort an meinem Kommentar vorbei.

MichaChrist, to random German
@MichaChrist@sueden.social avatar

Hier hätten Brennstoffzellenzüge fahren sollen. Doch die technologieoffene Ausschreibung ergab, dass Batteriezüge wirtschaftlicher sind. Der Markt regelt, liebe FDP.
https://www.swr.de/swraktuell/baden-wuerttemberg/suedbaden/ortenaukreis-erster-batterie-zug-startet-betrieb-100.html

Zeugs,
@Zeugs@social.cologne avatar

@MichaChrist nur 120 km Reichweite ? Was wenn ich damit mal spontan nach spanien fahren will?🙃

pluralistic, to ai
@pluralistic@mamot.fr avatar

> I find my feelings about are actually pretty similar to my feelings about : they do a poor job of much of what people try to do with them, they can't do the things their creators claim they one day might, and many of the things they are well suited to do may not be altogether that beneficial. And while I do think that AI tools are more broadly useful than blockchains, they also come with similarly monstrous costs.

https://www.citationneeded.news/ai-isnt-useless-2/

Zeugs,
@Zeugs@social.cologne avatar

@cstross @pluralistic @molly0xfff from what I have seen companies that make GenAI invest in GenAI example Google -> Anthropic.
Investor === Investee
Eg.
They pump up each other's values.

simon, (edited ) to random
@simon@simonwillison.net avatar

ChatGPT is now available for anyone to try for free without even creating an account - in an undocumented set of regions (works for me in California) https://openai.com/blog/start-using-chatgpt-instantly

It's the GPT-3.5 version, which is prone to all sorts of mistakes and hallucinations - further strengthening the pattern where most people form their impressions of what this stuff can and can't be useful for through access to the weaker models

Zeugs,
@Zeugs@social.cologne avatar

@simon
GPT 4 is just not that much better than GPT3.5, also in a hallucination benchmark, which I don't have at hand.
It just not justifies the cost for anyone but MS. It's not worth it for OpenAI, the customers and companies.
Maybe your experience with GPT3.5 is extraordinarily bad, but works for anyone else. 🤷
I played it through. I started with GPT 2. It's better but has still the same flaws.
I have to look at this for work but it's pointless with the existing expectations.

Zeugs,
@Zeugs@social.cologne avatar

@simon oh vendor performance information...
These benchmarks and their data are also in the training data. LLM generally perform worse with alternative formulations of the questions in the benchmarks.
https://arxiv.org/pdf/2402.19450.pdf
GPT4 is the best, but size does not justify the cost/size. GPT3.5 now The "vanilla LLM".
It's the defined normal and a standard you can talk about.

Zeugs,
@Zeugs@social.cologne avatar

@simon yeah chatbot arena gives you a competitive overview but that does not imply any argument about usefulness for solving problems in general.

Zeugs,
@Zeugs@social.cologne avatar

@simon I seen charts that say gpt 4 isn't much more accurate than gpt3.5 . fits my impression.
Blaming GPT3.5 on why people are not convinced by LLM's they call AI is not convincing me.

Zeugs,
@Zeugs@social.cologne avatar

@simon does it work with Claude 3?

SheDrivesMobility, (edited ) to random German
@SheDrivesMobility@norden.social avatar

deleted_by_author

    •
    Zeugs,
    @Zeugs@social.cologne avatar

    @SheDrivesMobility @Krautreporter

    easysociology, to Sociology
    @easysociology@mastodon.social avatar

    Educational Outcomes for Children in the UK Based on Class

    https://buff.ly/3UCG8py

    @sociology
    @academicchatter
    @academicsunite

    Zeugs,
    @Zeugs@social.cologne avatar

    @easysociology @sociology @academicchatter @academicsunite
    Harry Potter is shocked but he is also happy that he is not one of these Muggles.

    danilo, to random
    @danilo@hachyderm.io avatar

    tangential to but inspired by the OP: I think the way we address the Masto Reply Guy Infestation is make an absolutely clear shared reality that the replies are the Thunderdome and the best way not to get rocked is not to enter
    https://mastodon.social/@lzg/112128990197661962

    Zeugs,
    @Zeugs@social.cologne avatar

    @danilo aaaahh they do replys to themself. Thanks .Interesting ... So they use the attention on other posts to put their mega thread into it. Sounds a bit like chatgpt is writing the answers. I have noticed that only one time here....

    Zeugs,
    @Zeugs@social.cologne avatar

    @danilo I have read this reply guy thing before but I don't get it. You don't know what is good or not before you know/read it.
    Wasn't that one of the "Internet phenomena" in the first place? Why masto reply guys? What is the difference to message board people from the 90s that annoy you?

    Zeugs, to mastodon
    @Zeugs@social.cologne avatar

    What are plans for making switching instances easier or capsuling user data? This is imho the main advantage of . The mayor flaw on Bluesky architecture in my opinion is if the crawler will actually visit your profile often enough to keep up with the pace of the general social Network.

    RickiTarr, to random
    @RickiTarr@beige.party avatar

    Could you be in an Open/Poly Relationship?

    P.S. This is a place to discuss your own personal opinions about yourself and your choices, or ask earnest questions of others if they wish to answer. If you've come here to preach or degrade other's choices, that is a block from me.

    Zeugs,
    @Zeugs@social.cologne avatar

    @RickiTarr I am or was. Not being with other people lately, too lazy. Am I mono now? I don't know and possibly I don't care.

    georgetakei, to random

    This feels like a Star Trek episode.

    Zeugs,
    @Zeugs@social.cologne avatar

    @georgetakei That Voight-Kampf test of yours. Did you ever tried to take that test yourself?

    Zeugs, to scifi
    @Zeugs@social.cologne avatar

    is not a for the future ... It's more like an embedding new tools in old story plots.
    https://www.scientificamerican.com/article/tech-billionaires-need-to-stop-trying-to-make-the-science-fiction-they-grew-up-on-real/

    Zeugs, to FediPact
    @Zeugs@social.cologne avatar

    Says it will block
    https://fedipact.online/
    People fear that it will "split" the Fediverse.
    But it will only do what a net does, nodes and subgroups not being connected to everyone else on the net.
    That already happens by usual blocking.
    I think I would look at what comes from threads but I understand everyone that will block.

    axbom, to random
    @axbom@axbom.me avatar

    My son bought this book for me for my birthday (in June) but it didn't arrive until yesterday and he gave it to me when I came home from work.

    Started reading it in bed and am halfway through as I'm continuing my reading at the breakfast table. I would never suppose it's a book for everyone but wow is it hitting me at the exact right time in my life.

    "Turn your TV off. Throw your machines out the window. Save your excuses for someone who cares. Get to work."

    😁

    Zeugs,
    @Zeugs@social.cologne avatar

    @axbom "Struggel is good makes you strong, just like Conan turning the wheel in the dessert for years just to get the strongest of them all! "

    Zeugs, to random
    @Zeugs@social.cologne avatar

    Deutschland voll auf dem KI Bullshit trip. Dabei ist der Artikel noch einer der Kritischeren.
    Nein! Halluzinationen lassen sich ganz sicher nicht durch "gute Prompts" verhindern. Dieses :"Du machst es nur falsch das funktioniert alles super" ist schlimmstes Tech-Gaslighting.
    Auch irgendwelche Vektor Datenbanken, was richtig ist sind schwachsinn. Dann kannste schreiben was eh schon als Text in der Datenbank steht ... 😂🤦‍♂️

    axbom, to random
    @axbom@axbom.me avatar

    The world would generally make more sense to me if instead of saying "I'm using an AI to do this", people would just say "I'm using computer software to do this."

    The world won't be so lucky.

    General-purpose machines? Yes. That's what a computer is.
    I input data, it goes through software, and a brand new output comes out. Generative! My Sinclair ZX81 did this. Using BASIC.

    Maybe there's is something else this new definition of AI adds?

    • The amount of energy and data needed to build the software
    • Unpredictability of what the output will be.
    • The unexaplainability of the software.

    So what we are using is big-data, energy-intensive, unpredictable, unexplainable computer software. Also known as AI.

    Sorry, just thinking out loud.

    Zeugs,
    @Zeugs@social.cologne avatar

    @axbom It's also unreliable. An important factor for technology. This is blindly taken for granted today. In my impression it's a the consequence of a society that lived and grew up in an age with, in general, very reliable Technology.

    ryanc, (edited ) to random

    USB ports in wall sockets at home?

    (boost for reach)

    Zeugs,
    @Zeugs@social.cologne avatar

    @dashdsrdash @ryanc @nf3xn @cstross
    After 25 years. "Why was it called USB forever? " "Because it takes like forever to transfer data."

    axbom, to random
    @axbom@axbom.me avatar

    Imagine a computer following its programming and its programmers getting away with calling unwanted outputs a ”hallucination”.

    And then other people start calling anything they didn’t want or expect from the computer a hallucination. As if the programmers are completely unaccountable for their programming.

    It’s happening every single day and its like an adventure in Wonderland.

    Zeugs,
    @Zeugs@social.cologne avatar

    @axbom yeah my spelling errors are just halucinations of my Flesh computer. It that would have.worked earlier...

    Zeugs, to random
    @Zeugs@social.cologne avatar

    Has anyone tried ?
    Do you have experience with the "Fediverse" interaction between mastodon and Friendica?

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • provamag3
  • kavyap
  • DreamBathrooms
  • InstantRegret
  • osvaldo12
  • cubers
  • magazineikmin
  • mdbf
  • ethstaker
  • rosin
  • Youngstown
  • slotface
  • ngwrru68w68
  • modclub
  • JUstTest
  • thenastyranch
  • Durango
  • tester
  • cisconetworking
  • tacticalgear
  • khanakhh
  • everett
  • normalnudes
  • megavids
  • Leos
  • GTA5RPClips
  • anitta
  • lostlight
  • All magazines
    •