• 0 Posts
  • 177 Comments
Joined 3 years ago
cake
Cake day: June 14th, 2023

help-circle
  • Bullshit article, as usual from news talking about French politics.

    All non-left parties are fully supportive of the far right, as Macron proved by transforming France into a budding Trump-USA.

    The left parties are split into a lot of shit, between the “moderate left” that would rather see nazis win than actual leftists, the “extreme left” that is just moderate leftism but has been going through a smear campaign by Macron for a decade and is now labeled as a violent terrorist group (lol), and the other “extreme left” that are, this time, extreme, because they are just far-right fascism calling themselves leftists.




  • As I said, they exist, they just have no impact.

    A slippery slope is a had argument, but my point here was that we are visibly on the slope. We see more and more people tolerate AIs, including shit like slopcoding and slopart and the like. You can even see it on lemmy, which is probably one of the most AI-hostile places you’ll find. It’s more like starting to have heart problems after eating chips everyday, and thinking “heh, it’ll be fine, I’m not going to have a heart attack”. The opposite of the slippery slope, assuming that everything will be fine no matter what, is not more logical.

    And good for you, but you’re a tiny minority. Your point of view is not shared by most AI users, even on lemmy.

    Yeah, well you do you. It doesn’t take a long time looking at how the world is evolving to have a view of what’s coming.


  • I’m saying that “no one” runs a local, green-powered model, like yeah a handful of people probably do but they have no impact at all.

    And when you start with the mentality of the “it’s there so why not use it”, you are much more prone to let more happen. If when someone gets murdered, everyone reacts by just mourning and not trying to find out who did it, and to prevent further murders, things won’t get better.

    AIs need to be fought with the strongest force or they’ll win. That’s why you already see a split between “no AI here” places, and “oh yeah I’m just using AIs for this and this and that, but reasonably, and I also tried generating images but just for fun, and I made a slopsoftware just to be quicker, and I’m using it more because it frees my time, etc”.

    It’s a slippery slope argument, you might say, but we’re seeing it happen. People slowly accept AIs, just like they slowly accept capitalism, just like they accept fascism, and just like they accept all those brain-killing things that are hammered into their heads by everyone around. The only way to resist is to unequivocally reject it, not tolerate exceptions, and once the problem is solved, once AIs are green and powered by the wind, we can have fun running our local models and whatnot.

    TLDR: I get your point, but I think it’s a position that is very vulnerable to being corrupted, and the state of the world, including FOSS software and the like, seem to support my point of view.









  • So many critical bugs and security holes have been made from an oversight of the people handling the code.

    Now you want to tell me that instead of having people write code that tries to make sense, and then review it (sometimes a bit too late), you want to have an hallucination machine produce some code randomly, then have people “fix” it, then review it?

    This is just a recipe for disaster.

    AIs are not “AIs”, they’re just bullshit generators that everyone is falling for. Technical debt and lack of code reliability were the main problems of software dev, and AIs are sacrificing those two specifically, just in exchange for the illusion of speed.

    If you train monkeys to pile up bricks, it doesn’t make a house, it makes a disaster waiting to happen. And monkeys, unlike AIs, are actually intelligent and sentient, which would make them more reliable still.