Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)I
Posts
5
Comments
190
Joined
4 wk. ago

  • LLMs are AI - always have been. The term “artificial intelligence” has always been broad in computer science: it covers anything that performs a cognitive task normally requiring human intelligence. A chess engine from 1999 is AI. A spam filter is AI. An LLM is AI. Narrow AI, sure, but still AI.

    The confusion comes from people equating “AI” with sci-fi AGI (human-level general intelligence, HAL/JARVIS/Skynet/etc.). That’s a specific subset, not the whole category. When companies say “AI-powered” they’re not claiming AGI - they’re saying the product uses machine learning or pattern recognition in some way. Marketing inflates the language, yes, but the underlying tech is real and fits the definition.

    If/when we reach actual AGI, it will be a civilization-level shift - far beyond today’s spell-checker-that-sometimes-hallucinates. People will look back and say “we had AI for years,” but they’ll mean narrow tools, not the thing that can invent new science or run a company autonomously. The goalposts aren’t moving; the hype is just using the broad term loosely.

  • "Don't recommend this channel" feature exists and works.

  • There are multiple ways to be "unconscious." Head trauma, sleep, general anesthesia, fainting, coma - for example.

    The experience varies wildly: from absolute nothing under general anesthesia to extremely vivid stuff during sleep.

  • This is one of the things LLMs are actually pretty good at. Just don't blindly trust its output.

  • It’s not dependent on constant software updates.

    I bet every modern fighter jet is. "Dependent" might not be the best word, but if you can make your existing jet better just by optimizing the software, then of course they should.

    It's probably true to say that F-35 is objectively better than a Gripen, but it's way more expensive too. More Gripens might actually be better than fewer F-35s. My understanding is they're more focused on electronic warfare.

  • Even broke the warranty seal over the USB port.

  • Firstly, it's not just about the price of the drones being shot down - it's also about the price of whatever they were going to hit if you didn't stop them.

    And secondly, that's entirely beside the point anyway. The Dutch need stealth fighters just as much as anyone else would.

  • The interesting thing about general anesthesia is that it's quite unlike dreaming. It's like you're just instantly teleported into another place and time without any sense of time having passed in between. You were effectively dead for a few hours from everyone else's perspective, but for you there was no gap at all. It's not like there's a blank section in the film - rather like someone entirely cut out that part and you just jumped instantly to the next act.

    I can't help but wonder if something similar happens when you actually die. By definition you cannot experience being dead, so what if your consciousness just jumps over the being-dead part and continues from whatever is next? Even if there's a million-year-long queue before you get to respawn, that would still happen instantly from your subjective experience. Perhaps death is only for your physical body, but your consciousness can only continue to have experiences wherever there are experiences to be had.

    I think this idea is called quantum immortality.

  • I think the consciousness they're talking about here is the subjective sense of something happening - that it feels like something to be. The fact of experience itself. Unconsciousness in the medical sense doesn't necessarily mean the end of experience.

  • Either you're talking confidently about something you couldn't possibly know, or you're risking the rest of your life in prison for leaking top-secret military info. Which is it?

  • It was Dutch F-35s that shot down the Russian drones over Poland. It could've just as well been a Russian fighter jet they scrambled to intercept.

    Yeah, they do need a stealth jet. Stealth is what lets you fire your missiles before the enemy even knows you're there.

  • they imply that anti-depressants aren’t useful

    The title says they're just as effective as excercise. The only way to intrepret this as saying medication isn't useful is if you think excercise isn't useful either.

  • General intelligence refers to human level intelligence where it's not only limited to one task like playing chess or generating language. General intelligence exists - just not artificial one.

  • I've yet to figure out a good way to carry a bit driver with me but if I had one I'd use it a ton. Just like I do flashlight now.

  • Radiator bleed-screw tool on my keyring.

    I'm a plumber by training.

  • Just create it, and if there's demand, people will show up. There are a few communities meant for advertising new ones, but simply posting there yourself will also make it appear in the "All" feed for people browsing that.

    There's only so much you can do to grow it. Lemmy is still a pretty small platform with a limited audience, and I'm getting the sense that cage fighting isn't exactly close to most people's hearts around here.

  • Saying that it’s good at one thing and bad at others.

    But that's exactly the difference between narrow AI and a generally intelligent one. A narrow AI can be "superhuman" at one specific task - like generating natural-sounding language - but that doesn't automatically carry over to other tasks.

    People give LLMs endless shit for getting things wrong, but they should actually get credit for how often they get it right too. That's a pure side effect of their training - not something they were ever designed to do.

    It's like cruise control that's also kinda decent at driving in general. You might be okay letting it take the wheel as long as you keep supervising - but never forget it's still just cruise control, not a full autopilot.

  • It's a Large Language Model designed to generate natural-sounding language based on statistical probabilities and patterns - not knowledge or understanding. It doesn't "lie" and it doesn't have the capability to explain itself. It just talks.

    That speech being coherent is by design; the accuracy of the content is not.

    This isn't the model failing. It's just being used for something it was never intended for.

  • I love this.