• 2 Posts
  • 15 Comments
Joined 2 months ago
cake
Cake day: June 4th, 2025

help-circle
  • Thanks, I almost didn’t post because it was an essay of a comment lol, glad you found it insightful

    As for Wolfram Alpha, I’m definitely not an expert but I’d guess the reason it was good at math was that it would simply translate your problem from natural language into commands that could be sent to a math engine that would do the actual calculation.

    So basically act like a language translator but for typed out math to a programming language for some advanced calculation program (like wolfram Mathematica)

    Again, this is just speculation because I’m a bit too tired to look into it rn, but it seems plausible since we had basic language translators online back then (I think…) and I’d imagine parsing written math is probably easier than natural language translation


  • Engineer here with a CS minor in case you care about ethos: We are not remotely close to AGI.

    I loathe python irrationally (and I guess I’m masochist who likes to reinvent the wheel programming wise lol) so I’ve written my own neural nets from scratch a few times.

    Most common models are trained by gradient descent, but this only works when you have a specific response in mind for certain inputs. You use the difference between the desired outcome and actual outcome to calculate a change in weights that would minimize that error.

    This has two major preventative issues for AGI: input size limits, and determinism.

    The weight matrices are set for a certain number of inputs. Unfortunately you can’t just add a new unit of input and assume the weights will be nearly the same. Instead you have to retrain the entire network. (This problem is called transfer learning if you want to learn more)

    This input constraint is preventative of AGI because it means a network trained like this cannot have an input larger than a certain size. Problematic since the illusion of memory that LLMs like ChatGPT have comes from the fact they run the entire conversation through the net. Also just problematic from a size and training time perspective as increasing the input size exponentially increases basically everything else.

    Point is, current models are only able to simulate memory by literally holding onto all the information and processing all of it for each new word which means there is a limit to its memory unless you retrain the entire net to know the answers you want. (And it’s slow af) Doesn’t sound like a mind to me…

    Now determinism is the real problem for AGI from a cognitive standpoint. The neural nets you’ve probably used are not thinking… at all. They literally are just a complicated predictive algorithm like linear regression. I’m dead serious. It’s basically regression just in a very high dimensional vector space.

    ChatGPT does not think about its answer. It doesn’t have any sort of object identification or thought delineation because it doesn’t have thoughts. You train it on a bunch of text and have it attempt to predict the next word. If it’s off, you do some math to figure out what weight modifications would have lead it to a better answer.

    All these models do is what they were trained to do. Now they were trained to be able to predict human responses so yeah it sounds pretty human. They were trained to reproduce answers on stack overflow and Reddit etc. so they can answer those questions relatively well. And hey it is kind of cool that they can even answer some questions they weren’t trained on because it’s similar enough to the questions they weren’t trained on… but it’s not thinking. It isn’t doing anything. The program is just multiplying numbers that were previously set by an input to find the most likely next word.

    This is why LLMs can’t do math. Because they don’t actually see the numbers, they don’t know what numbers are. They don’t know anything at all because they’re incapable of thought. Instead there are simply patterns in which certain numbers show up and the model gets trained on some of them but you can get it to make incredibly simple math mistakes by phrasing the math slightly differently or just by surrounding it with different words because the model was never trained for that scenario.

    Models can only “know” as much as what was fed into them and hey sometimes those patterns extend, but a lot of the time they don’t. And you can’t just say “you were wrong” because the model isn’t transient (capable of changing from inputs alone). You have to train it with the correct response in mind to get it to “learn” which again takes time and really isn’t learning or intelligence at all.

    Now there are some more exotic neural networks architectures that could surpass these limitations.

    Currently I’m experimenting with Spiking Neural Nets which are much more capable of transfer learning and more closely model biological neurons along with other cool features like being good with temporal changes in input.

    However, there are significant obstacles with these networks and not as much research because they only run well on specialized hardware (because they are meant to mimic biological neurons who run simultaneously) and you kind of have to train them slowly.

    You can do some tricks to use gradient descent but doing so brings back the problems of typical ANNs (though this is still possibly useful for speeding up ANNs by converting them to SNNs and then building the neuromorphic hardware for them).

    SNNs with time based learning rules (typically some form of STDP which mimics Hebbian learning as per biological neurons) are basically the only kinds of neural nets that are even remotely capable of having thoughts and learning (changing weights) in real time. Capable as in “this could have discrete time dependent waves of continuous self modifying spike patterns which could theoretically be thoughts” not as in “we can make something that thinks.”

    Like these neural nets are good with sensory input and that’s about as far as we’ve gotten (hyperbole but not by that much). But these networks are still fascinating, and they do help us test theories about how the human brain works so eventually maybe we’ll make a real intelligent being with them, but that day isn’t even on the horizon currently

    In conclusion, we are not remotely close to AGI. Current models that seem to think are verifiably not thinking and are incapable of it from a structural standpoint. You cannot make an actual thinking machine using the current mainstream model architectures.

    The closest alternative that might be able to do this (as far as I’m aware) is relatively untested and difficult to prototype (trust me I’m trying). Furthermore the requirements of learning and thinking largely prohibit the use of gradient descent or similar algorithms meaning training must be done on a much more rigorous and time consuming basis that is not economically favorable. Ergo, we’re not even all that motivated to move towards AGI territory.

    Lying to say we are close to AGI when we aren’t at all close, however, is economically favorable which is why you get headlines like this.





  • Once upon a time this happened to me when I had classes. I kept saying I’d get up for the next one and never did (in fact I basically stopped going to class for a week).

    I really knew I should go and get out of bed but the more I thought about it, the harder it got.

    Eventually, I got medicated, and while sometimes my body tries to stop me from taking my meds, it’s easier to overcome that than it is to get out of bed when executive dysfunction hits.

    That’s why I keep my pills right next to my bed and typically I take them immediately when my alarm goes off.

    I also set out a protein shake next to them every night so I can eat “breakfast” without having to wait for them to kick in.

    Probably not the best solution for everyone, but if you want a suggestion, that’s mine.

    Also remember that you’re not alone and while some people might not see the struggle, that doesn’t make it any less valid. Be kind to yourself





  • This typically happens to me when I’m in the middle of something or when I’ve been trying to think my way through some problem for a long time.

    It’s not zen because it’s like I literally don’t exist, and when I do come back around, I think “shit I’m wasting time I need to get back to things” and then it happens repeatedly so i dont make any progress.


  • Unfortunately no, also I haven’t really watched much anime (I only know who three of the people on the chart are).

    However, in the non-anime category, the tv show “Hannibal” has two very intelligent characters play this same kind of game of cat and mouse.

    You know who the antagonist is, but the other characters… not so much. So you get the same kind of dynamic where people are talking with Light like he’s a normal person when you know he’s Kira.

    You also get some seriously fucked up mind games between the antagonist and protagonist and get those same moments of “well how is he going to get out of this?” For both of them.

    It’s also like a normal-ish crime TV show. Anyway, good stuff, idk if it’s what you’re looking for but you might like it regardless


  • Disclaimer because I’m about to start defending his intelligence: Light is a horrible person regardless.

    Light’s first slip up is his temper, causing him to act without thinking to kill Lind L Taylor on live TV. In all fairness, until that point he had killed hundreds of people already with zero resistance and if you want people to “know of my existence” what better way to do so than to kill someone on live TV and get away with it.

    The next “mistakes” he makes are in focusing on killing L instead of hiding from him. But really he would rather play a deadly game with L than anything else, and he does play that game intelligently and win despite many unforeseen obstacles.

    Point is, light doesn’t get caught because he isn’t intelligent, he gets caught because he’s more obsessed with power and domination than not getting caught or actually saving the world.

    I mean that’s literally why he’s evil: he’s not trying to fix the world.

    He himself mentions that society just tends to decay and—knowing that he won’t live forever—he must know that killing evil people and trying to keep people in line with fear will not work and will not last. However, he does it anyway because fixing the world is just the excuse he uses to justify his god complex.

    He’s not stupid, he’s just more entertained playing with death than he is doing anything else.





  • Technically, yes, you are correct because I don’t inhale at all lol

    I have bad lungs, so I’ve only had edibles and shit like that. The first one I tried was 15mg…? I know one of my stoner friends only took a half because a full one was “too much for right now” idk what that means since we basically just sat around and watched movies.

    Tbf, that time I think it might have made me slightly more likely to laugh at stupid shit, but it was also the first time I’d been with my old friends in a while, so maybe I was just in a more relaxed mood already. Didn’t really feel any other affects.

    All other times have been unremarkable. Also I really am not a fan of being able to faintly smell/taste THC for days afterwards.


  • AnarchoEngineer@lemmy.dbzer0.comtoComic Strips@lemmy.worldJuice
    link
    fedilink
    arrow-up
    9
    arrow-down
    1
    ·
    1 month ago

    Idk if I’ve completely missed the joke here, but seriously this is me with alcohol and weed and even stimulants.

    The only affect I get from alcohol is a headache or slight dizziness if I drink a significant amount. THC doesn’t seem to do anything at all, and stimulants typically just make me sleepy or anxious or nothing at all (I’m ADHD)