[object Object]

Autocorrect hates me, I am sorry.

  • 0 Posts
  • 16 Comments
Joined 1 year ago
cake
Cake day: July 20th, 2025

help-circle

  • I dumped my phone company over their AI answering system.

    It was totally broken:

    • you can’t know when the AI is lying or will do something it says it will for you
    • it’s not designed to be helpful, it’s designed to keep you away from humans
    • this particular system required you to say a correct and available time you want a callback, and if you were incorrect it would hang up on you.

    I’ll bet the cost metrics looked great, 100% of calls being handled without a callback! Oh, the customers leaving is unrelated, talk to the retention department.

    Just utter fucking stupidity. No, not stupidity, complete disregard for the customer to serve some department heads self interests. There needs to be a word stronger than greed…




  • I cannot for the life of me understand AI mandates.

    Who doesn’t want to be able to get their work done easier? Nobody.

    Then it follows you shouldn’t have to tell your employees to use a tool if it makes them however much as productive as CEOs think that tech demo should make them.

    I think it’s pretty obvious that at the exec level AI is the ultimate hoodwinking machine. It’s made 90% of them look like the out of touch morons they are. The 10% or whatever that aren’t trying to force this down their staff’s throats in the hopes they can replace said staff are coming out smelling like roses.

    Don’t fit a square peg in a round hole.


  • My idempotent claim was too broad and wrong, up to aliasing that is reversible.

    I think I’ve been talking about this differently than you, when you get 8bit into, 255 in/out is correct.

    When generating data I do it in the natural [0,1] space, then map to whatever domain I’m quantizing for.

    In that case, for N bins you divide by N+1 (and round however you want via correction) so 256. I was probably wrong earlier around this, 256 is correct for quantizing that way.



  • The correct way to scale a number from one range to another is ((value - min) / (max - min)) * output_range + output_min and in this way the standard colour normalization should be (value - 0)/(255 - 0) * 1 + 0 = value / 255

    Now the question becomes whether an 8bit pixel value is a correct representation of a colour, or should it quantize to a half value. To answer this, I aggressively point to the above equation. It is fine. There are 256 well considered values, and any system with higher resolutions/precision should also map from 0 to 1 per channel.

    It should be obvious to the reader that 0.5/256 is not zero, and therefore not black. The 256 method is also not idempotent, applying it once destructively alters the data.

    And the argument that 256 has higher resolution: so would 354256, but there’s no sensible place that matters.









  • It’s glaringly obvious that Israel is just there to sabotage every diplomatic avenue the US might take.

    And the US is basically getting cuckolded as Israel gets everything it wants and the US just impotently watches and pays for it all.

    There’s really no excuse for Israel to be doing this, and they also assassinated negotiators earlier. They really do not want any peace treaty or the US to pull out.