• 1 Post
  • 85 Comments
Joined 7 days ago
cake
Cake day: February 14th, 2025

help-circle
  • There’s a number of problems with this.

    It would only work if enough communities and instances adhered rigidly to similar editorial decisions about what is “political”. I don’t think that’s achievable.

    Mods would have to do the heavy lifting in tagging/ untangling things.

    It’s not going to be as simple to implement as you think.

    Posts are already categorised by community. You can block the vast majority of political posts by blocking a few communities and users.

    At times if filtered out keywords like Musk and Trump. You don’t really need to implement a feature for this. Just get posters to include a tag in the title of their post.

    Looking at your other comments, you’re not going to be deterred by these criticisms, so my suggestion would be to find the git repo and create an issue to get some proper feedback from other contributors.




  • I’m absolutely in the “don’t self-host email” camp. That said, I think it could be done reliably if you wanted to use someone else’s SMTP server and let them worry about deliverability. As in, have your mx records on your domain route to your MTA and dovecot, but set your DKIM and SPF records to match a third party SMTP server. You could use mxroute as an SMTP server very cheaply. There are others like the email API type services. I still can’t think of why I’d want to self host with all this drama but just an idea I’ve heard.


  • null_dot@lemmy.dbzer0.comtoPrivacy@lemmy.worldPlease, don't!
    link
    fedilink
    English
    arrow-up
    12
    ·
    15 hours ago

    Sorry chief you might have embarrassed yourself a little here. No big thing. We’ve all done it (especially me).

    Check out huggingface.

    There’s heaps of models you can run locally. Some are hundreds of Gb in size but can be run on desktop level hardware without issue.

    I have no idea about how LLMs work really so this is supposition, but suppose they need to review a gargantuan amount of text in order to compile a statistical model that can look up the likelihood of whatever word appearing next in a sentence.

    So if you read the sentence “a b c d” 12 times you don’t need to store it 12 times to know that “d” is the most likely word to follow “a b c”.

    I suspect I might regret engaging in this supposition because I’m probably about to be inundated with techbro’s telling me how wrong I am. Whatever. Have at me edge lords.