• 0 Posts
  • 9 Comments
Joined 1 year ago
cake
Cake day: February 5th, 2025

help-circle




  • I’m a bit scared of the commitment, I work full time and barely have energy for projects

    But mainly I would either start my own deserted community, or take over some that I’d be unqualified to manage

    I don’t think it’s the kind of socialisation I seek anymore, if I had the time I’d rather play a game of tribal wars and be part of a big organised tribe. But IRL stuff is probably what would interest me the most, unions or activism.




  • Tbh it could just be that the dems are doing a China move: do nothing and win

    Right now trump can either reference Biden, knowing people are sick of that, or he simply doesn’t have a definite target, “the radical left”. An emerging leader of the centre would give him a target. So it makes sense they stay quiet.

    I’m not a us citizen tho, I know that there is a lot of political activism, especially around the DSA, and the recent big protests are good signs. What I expect is that whoever ends up being the democratic leader will have a very hard time being a centrist. The demo-socialists have a big chance, but corporate media is also going to play its role in supporting a rethoric of “now that we’ve seen crazy let’s go for someone respectable, not another crazy”.

    The opposition, in the wide sense, is basically Sanders and the DSA, in my view, because they are popular and they are winning some local elections, including NY. But I still doubt they are going to be the face of the dem strategy for 2028, let alone the midterms.


  • I think generators have some kind of inherent style that we somehow learn to recognise

    Like sure they have learned on thousands of styles for each type of image, and you have some control of the style through prompt, but one issue with the transformer decoder model (the principles of which back almost all genAI at this point) is that at each generation step it gets the stuff generated so far as input.

    This feedback loop might induce repeated choices even on different prompts in the later stages of the generation. This is not apparent on images because they are seen all at once, but it is pretty evident on Suno (at least v3): later parts of different songs might share sounds. At least in my experiments making it generate EDM. I’m now able to spot the synth it often ends up creating.

    In terms of pictures and videos, that might be a reason generated stuff are consistently uncanny across image types.