• 0 Posts
  • 359 Comments
Joined 1 year ago
cake
Cake day: May 18th, 2025

help-circle

  • I can’t really differentiate among any of the models cause if one doesn’t work I’ll just toggle to a different model on duckduckgo. They tend to work very well for my purposes, which is locating sources of info that a regular search doesn’t turn up (or the results would take a lot of time to parse through).

    Their shortcomings aren’t a big deal for this because if one returns an incorrect result or hallucinates I can just correct it or hop models. And the ability to do fuzzy searches using plain language descriptions engines have never been good at, so in that case the LLMs are straight up better. But yeah otherwise a lot of their role for me is filling in for how bad search engines have gotten.

    I only code a little bit to solve my own problems so I’ve left it to people who’ve actually investigated to have an opinion on that angle. But last week I needed to add a quick function to one of my things so I tried one of the models, and it knocked it out immediately with no need for corrections. Less fun than solving the problem myself but woroed in a pinch since I was busy.

    I need to look into local models again now that they’re getting good and lightweight. But I’m not sure I’d be able to get much utility since my sole use case right now is search and I’m running an old thinkpad. One of these days I’ll tinker with it.






  • I haven’t seen the whole movie, but from a poker perspective it’s a pretty terrible scene (doesn’t mean it can’t be enjoyed dramatically)

    The only way that kinda runout happens is if someone was deliberately stacking the deck to cooler everyone and give Bond the win. Otherwise the cards are entirely random in which case a runout like that is not only astronomically rare, but also entirely out of Bond’s control. So maybe that aligns with the theme of the movie, idk, but it’s basically just charasterizing Bond as disgustingly lucky rather than skilled.

    Skill expression in poker comes from things a player can control, like bet sizing and hand selection. If the filmmakers wanted to show flashy skill, they could have made Bond induce a fold from the villain with a well-timed bluff, or have him make a risky hero call for all his money.


  • good post

    This user specifically, though, was called out on like a weekly basis, wasting dozens of hours of comrade’s labor trying to help them learn, and invariably each time crashing out and blocking the comrade engaging with them. They got reported many times before this latest crashout.

    This is the kind of person that needs to be excised from a community quickly, without wasting effort that could be spent teaching others. Especially online where there are limited stakes and a single person can disproportionately influence the culture of a site by being always online.

    We have finite time and there are too many shitheads to carefully explain to the ones that won’t listen. Organize the advanced, educate the intermediate, and isolate the backwards. Sometimes people in the party are backward enough that they need to be purged.




  • Actors are a hyper-complex subclass (speaking of the US, probably also true elsewhere). You’ve got the usual layers of complexity that most workers in the US have, like imperialism, racialization, etc, and the relationship between higher income and having marginal access to pre-defined systems of capital.

    But also, their labor is wholly superstructual, both in the sense of what it produces (which is usually a mistake to put too much weight into) and its relation to the superstructural reproduction of the dominant ideology. This local worker-owner dynamic has a unique relationship with the larger worker-owner dynamic of the nation/world. This might be a contributing factor for its level of unionization compared to other sectors.

    There are also the gradients between being in the reserve army of labor or otherwise employed*, versus selling their labor (dominant variable for the majority of working actors), versus selling their image (dominant variable for highly successful actors). The nature and duration of those experiences are going to color someone’s outlook, in addition to whatever immediate material interests they give them.

    And there’s all sorts of selection happening regarding who gets to enter the actor subclass and where on the above scale they land, from lookism and the usual social classes, to basic resources like fucking rent in LA, to networking access, and even explicit political ones like McCarthyism.

    As a whole, the section of working actors as well as reserve army of actors can probably be grouped with rest of the US proletariat for general purposes. Every sector is going to have its particularities, though.

    *And the class character here can have a strong influence, like how a lot of struggling or formerly struggling actors have experienced the service industry, which may leave them with an echo of radical class politics even if their immediate interests change.


  • A lot of good comments in this thread but this one is especially important, imo

    • Categories are abstract tools that can describe aspects of the world, but don’t materially exist.

    • The aspects of the world that we try to understand through categorization are dynamic, thus the categories and signifiers we use must also be dynamic.

    • When using categories we should always remember that they are, by necessity, reductive; we are deliberately emphasizing variables we deem operative or useful, at the expense of others. Those unimportant variables may rise to importance or even dominance in certain situations or at different scales.

    • The words we use in everyday discussion and especially in introducing people to new concepts are just shorthand for (hopefully) much more complex concepts. A static two-class model is easier to teach, just like we might assume frictionless vacuums to teach intro to physics. But to actually use the science in the real world needs more than freshman-level concepts. Problematically, the same words get used at every level (e.g. proletariat and bourgeoisie) so there isn’t like a semantic impetus to move into more complex understandings. A lot of people get stuck at that point.

    • We have to be very careful not to flip symbol and referent (falsely imbuing class signifiers with class character) or forget the core essence or purpose of our categories (using class character as a moral classification rather than a political-economic one).


  • The rightwing doesn’t “support” lonely men, it reasserts the need for dominance and power that is instrumental to their feeling of alienation in the first place.

    Yeah it sucks that that damage is being psychologically engraved in them by patriarchal culture. Yeah it would be great if there were more social resources to help them offramp and deconstruct it. But as it is, for our own safety and well-being, it can’t be our responsibility to do outreach for men who won’t demonstrate some willingness to work on themselves.

    I dunno. We could probably benefit from seeding communities with low-stakes social activities, fuckin commie-sponsored bowling or something. But there’s so much fucking damage to overcome.




  • Asking directly for knowledge from a tool that relies on statistical weights is a misuse of the tool, as I stated. That example was specifically to draw into question whether the commenter had actually investigated the things she was speaking on, because the arguments seemed severely outdated (in addition to being non sequitur).

    As I said in my comments, if I needed that kind of info I’d probably look to YouTube or forum posts. It’s entirely irrelevant to this conversation.