cultural reviewer and dabbler in stylistic premonitions

  • 1.19K Posts
  • 1.67K Comments
Joined 4 years ago
cake
Cake day: January 17th, 2022

help-circle
  • FYI, the day after you published this blog post, a spam blog posted… their AI reimplementation of it 🤦

    details:

    here is a snapshot of (maybe?) the “original” slop post borrowing from your title; i first saw it reposted on this slightly-more-credible-looking (at least if you haven’t seen it in previous search results and already realized it is spam) page:

    screenshot of dev dot to spam site

    i tried to archive that page with the repost of it, to avoid directly linking to spam from this comment, but it crashes archive.org’s browser:

    archive.org screenshot showing error message saying their browser crashed

    i also was curious to see if this spam is in search engines, so i searched for AI reimplementation, and… well, the good news is that your blog post is the first hit and the above-linked spam blog is pretty far down in the results list.

    The bad news is that the second hit is to yet another piece of slop/spam evidently also “inspired” by your post:

    duckduckgo screenshot





  • Nice post. Relatedly, see also malus.sh and this talk by the people that made it (both of which I posted in this lemmy community here).

    A couple of minor corrections to your text:

    Blanchard’s account is that he never looked at the existing source code directly.

    Blanchard doesn’t say that he never looked at the existing code; on the contrary, he has been the maintainer (and primary contributor) to it for over a decade so he is probably the person who is most familiar with the pre-Claude version’s implementation details. Rather, he says that he didn’t prompt Claude with the source code while reimplementing it. iirc he does not acknowledge that it is extremely likely that multiple prior versions of it were included in Claude’s training corpus (which is non-public, so this can only be conclusively verified easily by Anthropic).

    The GPL’s conditions are triggered only by distribution. If you distribute modified code, or offer it as a networked service, you must make the source available under the same terms.

    The GPL does not require you to offer GPL-licensed source code when using the program to provide a network service; because it is solely a copyright license, the GPL’s obligations are only triggered by distribution. (It’s the AGPL which goes beyond copyright and imposes these obligations on people running a program as a network service…)









  • Nice, thanks.

    It would certainly be nice to be able to pre-download language pair models without selecting to and from and then actually initiating a translation using the model i don’t have yet.

    re: getting uBlock externally, i also see the attraction of that approach but unfortunately Debian’s package was last updated in October (from 1.62 to 1.67) while AMO has a release from January (1.69) :/

    imo it would be better to bundle UBO and ship its updates along with browser updates.

    are there plans to distribute Konform via flathub?