The catarrhine who invented a perpetual motion machine, by dreaming at night and devouring its own dreams through the day.

  • 0 Posts
  • 522 Comments
Joined 10 months ago
cake
Cake day: January 12th, 2024

help-circle


  • When it comes to how people feel about AI translation, there is a definite distinction between utility and craft. Few object to using AI in the same way as a dictionary, to discern meaning. But translators, of course, do much more than that. As Dawson puts it: “These writers are artists in their own right.”

    That’s basically my experience.

    LLMs are useful for translation in three situations:

    • declension/conjugation table - faster than checking a dictionary
    • listing potential translations for a word or expression
    • a second row of spell/grammar-proofing, just to catch issues that you didn’t

    Past that, LLM-based translations are a sea of slop: they screw up with the tone and style, add stuff not present in the original, repeat sentences, remove critical bits, pick unsuitable synonyms, so goes on. All the bloody time.

    And if you’re handling dialogue, they will fuck it up even in shorter excerpts, by making all characters sound the same.




  • The drop is slowing down considerably:

    Month Users Change from previous month in %
    Mar 53687 N/A N/A
    Apr 51298 -2389 -4.5%
    May 48832 -2466 -4.8%
    Jun 48472 -360 -0.74%
    Jul 47297 -1175 -2.4%
    Aug 47876 +579 +1.2%
    Sep 47227 -649 -1.4%
    Oct 45037 -2190 -4.6%
    Nov 44837 -200 -0.44%

    And given that March was a peak, I’m tempted to interpret it as newbies not sticking around. I think that it’ll plateau around 40k users, then provided that the conditions remain the same it won’t increase or decrease.

    That’s why I say that it’s stable - the core userbase will likely stick around.

    That said, these numbers may particularly be bad, e.g. if anyone left Lemmy and went to Mbin and/or PieFed, then I think they would not be counted in those charts?

    They wouldn’t be counted but I don’t think that this introduces a lot of inaccuracy. Mbin has 1.7k MAUs, and PieFed has 104.

    The number of instances dropping is far more concerning IMO. It means that smaller instances have a hard time becoming sustainable.








  • For real. Companies being extra pushy with their product always makes me picture their decision makers saying:

    “What do you mean, «we’re being too pushy»? Those are customers! They are not human beings, nor deserve to be treated as such! This filth is stupid and un-human-like, it can’t even follow simple orders like «consume our product»! Here we don’t appeal to its reason, we smear advertisement on its snout until it needs to open the mouth to breath, and then we shove the product down its throat!”

    Is this accurate? Probably not. But it does feel like this, specially when they’re trying to force a product with limited use cases into everyone’s throats, even after plenty potential customers said “eeew no”. Such as machine text and image generation.




  • Bots are parasites: they only thrive if the host population is large enough to maintain them. Once the hosts are gone, the parasites are gone too.

    In other words: botters only bot a platform when they expect human beings to see and interact with the output of their bots. As such they can never become the majority: once they do, botting there becomes pointless.

    That applies even to repost bots - you could have other bots upvoting the repost, but you won’t do it unless you can sell the account to an advertiser, and the advertiser will only buy it if they can “reach” an “audience” (i.e. spam humans).