He / They

  • 28 Posts
  • 1.4K Comments
Joined 1 year ago
cake
Cake day: June 16th, 2023

help-circle
  • One of the worst and most predictable outcomes of Marxist-Leninism is the meshing together of the post-revolution counter-revolutionary force, with the single-party state.

    Marx never intended for there to be a central government wielding authority to combat counter-revolutionary backlash, this was supposed to be a continuation of the proletariat revolutionary movement, that served both to prevent society from sinking back into top-down control during peacetime complacency, and to keep the productive, positive, unifying spirit generated in a freedom-seeking revolutionary movement from dying out once the revolution was complete, and seeing the proletariat become divided.

    When Lenin (and drawing on Lenin, Mao) murdered the whole “stateless” part of Communism, that counter-revolutionary force became a tool for justifying suppression of the proletariat by the State, because the state was (not actually, of course) the revolutionary force. This also insulates the state against ever dissolving into the actualized stateless society that MLs still claim to totally be moving towards, because anyone seeking to dissolve the state to that end is, in the state’s eyes, indistinguishable from other counter-revolutionary forces seeking to dissolve the state to return to Capitalism/ Feudalism/ Monarchy/ etc.

    *After all, how can the single-party State know if the proletariat are actually good Communists, or whether they’re actually counter-revolutionaries? Only by maintaining constant surveillance and vigilance (and continuing the State)!






  • One thing I’d push back on in the article is:

    That cost-per-user doesn’t decrease as you add more customers. You need more servers. More GPUs.

    This is assuming constant use, which is not the case. If I have a server handling LLM prompt requests, and for illustrative purposes each request uses 100% of the single discrete GPU in it, and I only have 1 customer, but that one customer only uses it 5% of the day (which would actually be pretty high in real terms), I can still add additional customers without needing to buy additional servers. The question is whether the given revenue of a single server outweighs its cost to run.

    And when it comes to training, that is an upfront cost, that you could (if you get a model to where you want it) stop having to pay whenever you want. I’m pretty surprised they haven’t been really leaning into training models for medical diagnoses, because once you have a model that can e.g. spot a type of tumor with n% accuracy beyond a human, you don’t really have to refine it further if you don’t want to (after all, it’s not like the humans can choose to do it better themselves at that point, like they can with writing prompts).





  • I think that it just makes it super easy to condemn anyone you dislike. Anyone can be matched to one or more of the behavioral indicators for NPD, individually, because many of them are just regular human behaviors. Selfishness, callousness, anger, egotism, not taking criticism well, etc, are all normal things to experience, and everyone does them, so if you want to condemn someone you can just subjectively and selectively label instances of that behavior as being an indicator of NPD, ignoring any absent or incongruous patterns or criteria.