• 5 Posts
  • 158 Comments
Joined 1 year ago
cake
Cake day: June 13th, 2023

help-circle





  • AbouBenAdhem@lemmy.worldtoAsklemmy@lemmy.mlAlternatives to congress
    link
    fedilink
    English
    arrow-up
    3
    arrow-down
    2
    ·
    3 days ago

    Direct democracy—except instead of directly voting on legislation, voters vote on the desired effects of legislation and a metric for measuring if those effects are being achieved. The actual legislation is then written by specialists trained on effective policy implementation, who can adjust the legislation on the fly if it isn’t having the desired effect. Their mandate is limited by the associated metric—if they can’t meet the goals, they lose their mandate and the case goes back to voters for review.










  • They’ve been overstepping enough on a regular basis for the last fifty years—the real problem is that they’ve subverted the “reform” process so that reforms that seem adequate to the general public get neutralized or twisted to work in their favor.

    That’s why you have more-experienced reform advocates eventually pushing things like “defund the police”—they may be shooting themselves in the foot in terms of popular perception, but it comes from a long history of frustration with lesser reform efforts.



  • When cats meow, there’s a one-to-one correspondence between the aural qualities of the sound and the communicative intent of the cat—the same meow doesn’t have different meanings depending on the preceding and following meows. That’s how animals normally use sounds to communicate.

    There are two common exceptions, where animals string arbitrary sounds together in longer sequences in which the individual components don’t have distinct communicative intents in the way animals usually interpret them: songbirds and humans. (Another possible exception might be cetaceans.)

    (For example: If I said “pass the butter”, “don’t eat all the butter”, or “I need to get more butter”, the word “butter” would have different communicative intents even if I said them the exact same way—like a note of a bird’s song, and unlike a cat’s meow.)




  • AbouBenAdhem@lemmy.worldtoAI@lemmy.mlDo I understand LLMs?
    link
    fedilink
    English
    arrow-up
    6
    ·
    edit-2
    11 days ago

    There’s a part of our brain called the salience network, that continually models and predicts our environment and directs our conscious attention to things it can’t predict. When we talk to each other, most of the formal content is predictable, and the salience network filters it out; the unpredictable part that’s left is the actual meaningful part.

    LLMs basically recreate the salience network. They continually model and predict the content of the text stream the same way we do—except instead of modeling someone else’s words so they can extract the unpredictable/meaningful part, they model their own words so they can keep predicting the next ones.

    This raises an obvious issue: when our salience networks process the stream of words coming out of such an LLM, it’s all predictable, so our brains tell us there’s no actual message. When AI developers ran into this, they added a feature called “temperature” that basically injects randomness into the generated text—enough to make it unpredictable, but not obvious nonsense—so our salience networks will get fooled into thinking there’s meaningful content.