[–] 1 point 1 month ago (4 children)

I think their point is that for a business a captive audience is a good thing. Kind of like how some grocery stores have noticed that when checkout takes longer, people are more likely to buy the drinks and snacks they put for sale near the checkout line. Likewise a charging station could have an overpriced restraunt selling low quality burgers for twice the price of the restraunts 2 minutes away by car, and customers will eat there because, they don't want to sit around for 30 minutes, and then drive down the road and stop for another 10.

Unless I'm missing something and fast chargers charge by the watt instead of the hour, in which case that's pretty big BS.

  • source
  • parent
  • context
  • [–] 2 points 1 month ago

    Well technically yes, but defending them from what. IMO I don't get the point of gratuitous punishment.

    IE I see 2 possibilities.

    1. Someone can be rehabilitated. IE we think it's reasonable that after the punishment we consider them able to be low enough risk that more or less they can contribute to society and function as a normal person again.

    2. Someone can't be rehabilitated. In which case punishment is completely meaningless, and we need to keep them away from any potential victims for the rest of their lives. Of which I do understand but not agree with the concept of shortening it.

    But in the case of 2, I flat out don't get the concept of intentionally adding suffering in that point for any purpose other than keeping them from getting near innocents. If we've given up on changing them, then we've negated any potential purpose to the suffering, and thus it's just gratuitous and... makes those involved in it's execution the monsters.

  • source
  • parent
  • context
  • [–] 2 points 1 month ago

    I agree with the overall concept, LLMs seem extremely useful, I'd say our direction for them is just idiotic. Local devices trained for what they are actually doing, without sending everything back so you can confidently train it on your styles of things etc... I see as a very practical idea. But of course the US's super focused giant corporate controlled everything goes to a data center of some huge dangerous tech company is beyond ridiculous, and of course just the build out scale of investing literal trillions of dollars into it.

  • source
  • parent
  • context
  • [–] 9 points 1 month ago

    I don't know to me the level it worked, is what makes it really silly to me that so few things can handle local transcription today without cloud services. I felt like it was 99% accurate running on a crappy win XP system with like 256 MB ram. Seems to me that it's crazy that 20 years later we're "well we need to hook you up to a supercomputer to do that".

  • source
  • parent
  • context
  • [–] 6 points 1 month ago (2 children)

    Well absolutely in the judeism base of it. Old testament was constantly "god prepared this perfect place just for you" "whoops looks like there's already a thriving civilization of happy people there, they must be evil why else would they have taken the land I promised to you".

  • source
  • parent
  • context
  • [–] 22 points 1 month ago

    I could see that, but I think the key difference.

    Adblocker is an active block of something from outside.

    AI block is telling firefox not to use any of it's own.

    Kind of a do this, versus don't.

    "Do protect me from attackers."

    "Don't attack me"

  • source
  • parent
  • context
  • [–] 4 points 1 month ago* (last edited 1 month ago)

    We've also decided to add a "pinky promise" layer to this agreement.

    You know what, considering that the world see's AI on the same level of nukes. Maybe the iran war can be solved this way. Just give Iran a "morally binding" request to not build nukes. Declare that the new victory and end the iran war.

  • source
  • [–] -3 points 1 month ago (1 child)

    I think that's about step 1... if the republicans were gone... more democrats would be challanged from the left. Either a real left wing party could be established. Or democrats would be able to consistantly be forced in the right direction. The biggest problem is the constant if we don't 100% unify quickly we risk the facists taking over and creating messes that will take 10 years of good policy to reverse even if we get a perfect candidate in, is the giant problem.

  • source
  • parent
  • context
  • [–] 9 points 1 month ago

    They do and don't want the ideas and the misinformation. but they want it in a subtle enough way that it isn't overtly obvious. Hence why musk had to roll mechahitler back.

    "they like my ideas, they just have a problem with the word nazi".

  • source
  • parent
  • context
  • [–] 14 points 1 month ago (3 children)

    It's the same problem Elon Musk has, in short gen AI is too complicated, in order to have any actual value it has to have access to a good data set of facts. If you try to teach them to lie about certain topics, it just goes full mecha hitler.

  • source
  • view more: ‹ prev next ›