cultural reviewer and dabbler in stylistic premonitions

  • 166 Posts
  • 291 Comments
Joined 3 years ago
cake
Cake day: January 17th, 2022

help-circle























  • Arthur Besse@lemmy.mltoPrivacy@lemmy.mlI made a gpg Hat
    link
    fedilink
    English
    arrow-up
    6
    ·
    edit-2
    1 month ago

    were you careful to be sure to get the parts that have the key’s name and email address?

    It should be if there is chunks missing its unusable. At least thats my thinking, since gpg is usually a binary and ascii armor makes it human readable. As long as a person cannot guess the blacked out parts, there shouldnt be any data.

    you are mistaken. A PGP key is a binary structure which includes the metadata. PGP’s “ascii-armor” means base64-encoding that binary structure (and putting the BEGIN and END header lines around it). One can decode fragments of a base64-encoded string without having the whole thing. To confirm this, you can use a tool like xxd (or hexdump) - try pasting half of your ascii-armored key in to base64 -d | xxd (and hit enter and ctrl-D to terminate the input) and you will see the binary structure as hex and ascii - including the key metadata. i think either half will do, as PGP keys typically have their metadata in there at least twice.



  • 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.