cultural reviewer and dabbler in stylistic premonitions

  • 168 Posts
  • 648 Comments
Joined 3 年前
cake
Cake day: 2022年1月17日

help-circle



  • They aren’t pro corpo Ai.

    They’re very much against the mass scraping/ddos ai companies are doing.

    All of the self-hostable LLMs and image generators (or at least, all of the ones capable of the quality people have come to expect for the last few years) people are using today are trained on massive scraped datasets far beyond the reach of hobbyists. There are many so-called “open source” models which are free to modify (eg, by fine-tuning) and to redistribute, but the data used for the initial training (which hobbyists are allowed to build upon) cannot be published because doing so would obviously be large-scale copyright infringement.

    Also, even with the data (which in many cases also needs to be labeled/annotated using human labor), the cost of training such a model from scratch is astronomical.

    As a pirate myself, I totally understand how, after reading that Meta’s training data included 82TB of pirated books they torrented, one’s first thought might be “🤤” … but to imagine that this makes Meta our ally in the fight against copyright is some temporarily-embarrassed-millionaire kind of thinking.


  • This article buries the lede so much that many readers probably miss it completely: the important takeaway here, which is clearer in The Register’s version of the story, is that ChatGPT cannot actually play chess:

    “Despite being given a baseline board layout to identify pieces, ChatGPT confused rooks for bishops, missed pawn forks, and repeatedly lost track of where pieces were."

    To actually use an LLM as a chess engine without the kind of manual intervention that this person did, you would need to combine it with some other software to automate continuing to ask it for a different next move every time it suggests an invalid one. And, if you did that, it would still mostly lose, even to much older chess engines than Atari’s Video Chess.

    edit: i see now that numerous people have done this; you can find many websites where you can “play chess against chatgpt” (which actually means: with chatgpt and also some other mechanism to enforce the rules). and if you know how to play chess you should easily win :)








  • Arthur Besse@lemmy.mltoScience Memes@mander.xyzfaen
    link
    fedilink
    English
    arrow-up
    14
    arrow-down
    2
    ·
    22 天前

    Due to the Norwegian language conflict there have been various competing forms of written Norwegian over time, two of which have been officially recognized as equally valid by the Norwegian parliament since 1885. Both apparently changed their spelling of “slut” to “sludd” in the 21st century, Bokmål in 2005 and Nynorsk in 2012, presumably in an effort to encourage English speakers to make jokes about Swedes and Danes instead of them.












  • The network never went down.

    You say that but, everything I ever posted on identica (and also on Evan’s later OStatus site Status.Net, which i was a paying customer of) went 404 just a few years later. 😢

    When StatusNet shut down I was offered a MySQL dump, which is better than nothing for personal archival but not actually useful for setting up a new instance due to OStatus having DNS-based identity and lacking any concept for migrating to a new domain.

    https://identi.ca/evan/note/6EZ4Jzp5RQaUsx5QzJtL4A notes that Evan’s own first post is “still visible on Identi.ca today, although the URL format changed a few years ago, and the redirect plugin stopped working a few years after that.” … but for whatever reason he decided that most accounts (those inactive over a year, iiuc, which I was because I had moved to using StatusNet instead of identica) weren’t worthy of migrating to his new pump.io architecture at all.

    Here is some reporting about it from 2013: https://lwn.net/Articles/544347/

    As an added bonus, to the extent that I can find some of my posts on archive.org, links in them were all automatically replaced (it was the style at the time) with redirects via Evan’s URL shortening service ur1.ca which is also now long-dead.

    screenshot of Roy Batty (Rutger Hauer) in the 1982 film Blade Runner, during his "Tears in rain" monologue. (no text)

    imo the deletion of most of the content in the proto-fediverse (PubSubHubbubiverse? 😂) was an enormous loss; I and many other people had years of great discussions on these sites which I wish we could revisit today.

    🪦

    The fact that ActivityPub now is still a thing where people must (be a sysadmin or) pick someone else’s domain to marry their online identity to is even more sad. ActivityPub desperately needs to become content addressable and decouple identity from other responsibilities. This experiment (which i learned of via this post) from six years ago seemed like a huge step in the right direction, but I don’t know if anyone is really working on solving these problems currently. 😢



















OSZAR »