• 0 Posts
  • 379 Comments
Joined 2 years ago
cake
Cake day: June 30th, 2023

help-circle










  • The main advantage is, if you keep a lot of “bookmarks” (which I don’t, but I like to be prepared) they are searchable not just by name, but also by tag.

    This means when you find that good stack overflow page about resolving cuda issues on a VM you can tag it as “linux” “python” “cuda” “VM” in addition to giving it a name. Then you can search for any of those and find it.

    The free version is pretty good and it lets you do one level of folders and as many bookmarks as you want. (Oh yeah. It supports folders so I have one for example that is called Unwatched YouTube, you might guess what I use it for). The paid version is probably better because it includes unlimited folder nesting AND it saves a snapshot of whatever webpage you’re bookmarking so you never lose anything. I would think this would be indespensible if you habitually save things like reddit or Lemmy threads so it doesn’t get deleted from under you.

    I don’t use raindrop.io as much as I would like, but I would assume it’s a far cry better than a browser with 200 tabs open.



  • Now hang on a minute lol. AI is just stolen garbage, but obviously we expect the “wiser folks” here in the agora of lemmy are going to give reasonable / acceptable answers? This is like the mcdonalds of philosophy here.

    So I see here two conjectures:

    1. AI is bad because it is creating a mashup of information (which may or may not be accurate) from sources it took without permission.
    2. People are within their rights to outsource the articulation of their opinions to experts (see also “wiser folks”. After all, this “division of labor” has always existed.

    1q.) So let’s say I take my 2 GPU workhorse PC and train basic language (not obviously line of reasoning, or guardrails, or other languages or anything like that) from a library of articles and professional documents I own or control. Then, by way of something like resource augmented generation (or similer idc) it gives me a well articulated argument of why AI is bad, is that reasonable? I would think this is a BETTER perspective than 2q below.

    2q.) In what way is mining the totally anonymous, unverifiable posts of literally any person with a keyboard on lemmy MORE valuable than a reasonable sounding argument from any generative AI or just pressing the middle button on your phone over and over? This sounds totally stupid. “Division of labor” has probably made all of us dumber. I (coincidentally) build language models as part of my job, somewhere in this thread is an AI expert who has read 1 newspaper article and is “training” on the information from other lemmy comments.

    At the same time we say “Holy shit AI bad, AI hallucinate, AI lies!” we are going to say it’s totally cool and reasonable to shout into the internet box where rando people can say anything they want, and that’s better?

    I mean I do like the smug argument and the smiley face, but the premise that “Gen AI sucks, hone your argument against AI using ask-fucking-lemmy” borders on content for c/selfawarewolves. It’s so ridiculous I practically expect you’re just trolling the thread.