Login
You're viewing the discuss.systems public feed.

Replies

  • 💬 1🔄 0⭐ 1
  • Peter Alvaropalvaro
    Apr 9, 2026, 3:03 AM

    @va2lam all of these things could be true at the same time

    💬 1🔄 0⭐ 0
  • Karl Statzkstatz12
    Apr 9, 2026, 3:09 AM

    @palvaro @va2lam the best use case I have for LLMs is still generating coloring book pages for my 4 year old. "you want Luigi, Sonic, Mickey Mouse and a super kitty robbing a bank?" parents of yesteryear would be shit out of luck.

    I use LLMs like if I had an insane parrot as a lab assistant. I've genuinely deepened my knowledge based on things an LLM has suggested. the trick is I actually did the legwork and read the source and walked away better off.

    oh and $DAYJOB has an AI code reviewer that I am passive aggressively mean to. that is better/cheaper than therapy

    💬 1🔄 0⭐ 1
  • Patrick Lam :tinoflag:va2lam@mastodon.nz
    Apr 9, 2026, 3:49 AM

    @kstatz12 @palvaro yeah, you have to do the legwork and not believe. The therapy is cheap for you at least, in a world where I hear people get told to use tokens whether they want to or not.

    💬 1🔄 0⭐ 0
  • Karl Statzkstatz12
    Apr 9, 2026, 4:06 AM

    @va2lam @palvaro yeah, I've never been told to "use AI" for something thankfully. I feel like every vibe code post is some form of "only I use qtips in my ears right".

    my favorite is "I was an engineer at (insert mega corp) here's my course on AI" TikTok slop. the true fun is watching chatbots argue in comment sections

    💬 1🔄 0⭐ 1
  • 💬 1🔄 0⭐ 1
  • Karl Statzkstatz12
    Apr 9, 2026, 4:38 AM

    @va2lam @palvaro don't worry Anthropic's Mythos apparently cost 10 billion dollars to train. which is totally sustainable and proves this is a good way to do things

    💬 0🔄 0⭐ 2
  • zoolzool@icosahedron.website
    Apr 9, 2026, 4:25 AM

    @palvaro for me it just reads "Unavailable Due to the UK Online Safety Act" which is a message in itself :/

    💬 0🔄 0⭐ 0