Gradually_Adjusting
@Gradually_Adjusting@lemmy.world
Priceless to you because I put you on the hype shit
- Comment on AGI is not coming! - Yanick Kilcher 1 week ago:
I’m less mentally organised than I was yesterday, so for that I apologise. I suspect the problem is that we’re both working from different ideas of the word intelligence. It’s not a word that has a single definition based on solid scientific grounds. The biggest problem in neuroscience might be that we don’t have a grand unified theory of what makes the mind do “intelligence”, whatever that is. I did mistake your position somewhat, but I think it comes down to the fact that neither of us has a fully viable theory of intelligence and there is too much we cannot be certain of.
I admit that I overreached when I conflated intelligence and consciousness. We are not at that point of theoretical surety, but it is a strong hunch that I will admit to having. I do feel I ought to be pointing out that LLMs do not create a model, they merely work from a model - and not a model of anything but word associations, at that. But I do not want to make this a confrontation, I am only explaining a book or two I have read as best I can, in light of the observations I’ve made about LLMs.
From your earlier comments about different degrees of intelligence (animals and such), I have tried to figure that into how I describe what intelligence is, and how degrees of intelligence differ. Rats also have a neocortex, and therefore likely use the self-same pattern of repeating units that we do (cortical columns). They have a smaller neocortex, and fewer columns. The complexity of behaviour does seem to vary in direct proportion to the number of cortical columns in a neocortex, from what I recall reading. Importantly, I think it is worth pointing out that complexity of behaviour is only an outward symptom of intelligence, but not likely the source. I put forward the “number of cortical columns” hypothesis, because it is the best one I know, but I also have to allow that other types of brains that do not have a neocortex can also display complex behaviours and we would need to make sense of that once we have a workable theory of how intelligence works in ourselves. It is too much to hope for all at once, I think.
So complex behaviour can be expressed by systems that do not closely mimic the mammalian neocortical pattern, but I can’t imagine anyone would argue that ours is the dominant paradigm (whether in terms of evolution or technology, for now), so in the interest of keeping a theoretically firm footing until we are more sure, I will confine my remarks about theories of intelligence to the mammalian neocortex until someone is able to provide a compelling theory that explains at least that type of intelligence for us. I have not devoted my career to understanding these things, so all I can do is await the final verdict and speculate idly with people inclined to do so. I hope only that the conversation can continue to be an enjoyment, because I know better than anyone I am not the final word on much of anything!
- Comment on AGI is not coming! - Yanick Kilcher 1 week ago:
I’ve watched a couple of these. You might find FreeTube useful for getting YT content without the ugly ads and algo stuff.
There are shortcomings that keep an LLM from approaching AGI in that way. They aren’t interacting (experiencing) with the world in a multisensory or realtime way, they are still responding to textual prompts within their frame of reference in a more discrete, turn-taking manner. They still require domain-specific instructions, too.
An AGI that is directly integrated with its sensorimotor apparatus in the same way we are would, for all intents and purposes, have a subjective sense of self that stems from the fact that it can move, learn, predict, and update in real time from its own fixed perspective.
Jeff Hawkins’ work still has me convinced that the fixed perspective to which we are all bound is the wellspring of subjectivity, and that any intermediary apparatus (such as an AI subsystem for recognizing pictures that feeds words about those pictures to an LLM that talks to another LLM etc, in order to generate a semblance of complex behaviour) renders the whole as a sort of Chinese room experiment, and the LLM remains a p-zombie. It may be outwardly facile at times, even enough to pass Turing tests and many other such standards of judging AI, but it would never be a true AGI because it would never have a general facility of intelligence.
I do hope you don’t find me churlish, I hasten to admit that these chimerae are interesting and likely to have important considerations as the technology ramifies throughout society and the economy, but I don’t find them to be AGI. It is a fundamental limitation of the LLM technology.
- Comment on Grok Claims It Was Briefly Suspended From X After Accusing Israel of Genocide 1 week ago:
A lot of us didn’t have a cool philosophy teacher who explained p-zombies and it shows
- Comment on AGI is not coming! - Yanick Kilcher 1 week ago:
For a snappy reply all I can say is that I did qualify that a “conventional” LLM likely cannot become intelligent. I’d like to see examples of LLMs paired with sensorimotor systems, if you know of any. Although I have been often inclined to describe human intelligence as merely a bag of tricks that, taken together, give the impression of a coherent whole, we have a rather well developed bag of tricks that can’t easily be teased apart. Merely interfacing a Boston Dynamics robo-dog with the OpenAI API may have some amusing applications, but nothing could compel me to admit it as an AGI.
- Comment on AGI is not coming! - Yanick Kilcher 1 week ago:
The argument is best made by Jeff Hawkins in his Thousand Brains book. I’ll try to be convincing and brief at the same time, but you will have to be satisfied with shooting the messenger if I fail in either respect. The basic thrust of Hawkins’ argument is that you can only build a true AGI once you have a theoretical framework that explains the activity of the brain with reference to its higher cognitive functions, and that such a framework necessarily must stem from doing the hard work of sorting out how the neocortex actually goes about its business.
We know that the neocortex is the source of our higher cognitive functions, and that it is the main area of interest to the development of AGI. A major part of Hawkins’ theory states that because the neocortex is arranged into many small columns (cortical columns), it is chiefly the number of them that differs between creatures of different intelligence level, and it forms essentially a basic repeating unit across the whole of the neocortex to model and make predictions about the world based on sensory data. He holds that these columns vote amongst each other in realtime about what is being perceived, constantly piping up and shushing each other and changing their models based on updated data almost like a rowdy room full of parliamentarians trying to come to a consensus view, and that it is this ongoing internal hierarchy of models and perceptions that makes up our intelligence, as it were.
The reason I ventured to argue that sensorimotor integration is necessary for an AI to be an AGI is because I got that idea from him as well; in order to gather meaningful sensory data, you have to be able to move about your environment to make sense of your inputs. Merely receiving one piece of sensory data fails to make any particular impression, and you can test this for yourself by having a friend place an unknown object against your skin without moving it, and having you try to guess based on that one data point. Then, have them move the object and see how quickly you gather enough information to make a solid prediction - and if you were wrong, your brain will hastily rewire its models to update based on that finding. An AGI would similarly fail to make any useful contributions unless it has the ability to move about its environment (asterisk - that includes a virtual environment) in order to continually learn and make predictions. The sort of thing we cannot possibly expect from any conventional LLM, at least as far as I’ve heard so far.
I’d better stop there and see if you care to tolerate more of this sort of blather. I hope I’ve given you something to sink your teeth into, at any rate.
- Comment on AGI is not coming! - Yanick Kilcher 1 week ago:
AGI can’t come from these LLMs because they are non-sensing, stationary, and fundamentally not thinking at all.
AGI might be coming down the pipe, but not from these LLM vendors. I hope a player like Numenta, or any other nonprofit, open-source initiative manages to create AGI so that it can be a positive force in the world, rather than a corporate upward wealth transfer like most tech.
- Comment on Google says it's working on a fix for Gemini's self-loathing 'I am a failure' comments 1 week ago:
Oh no, no. No! They’re finally creating an accurate LLM! Fuck
- Comment on Palantir gets $10 billion contract from U.S. Army 2 weeks ago:
Knew a dude who worked there during their startup days. They told me Peter Thiel would wear double bluetooth earbuds way back when it was still considered immensely douchey to have even one.
What I’m trying to say is Thiel has always been an unbelievably worthless person. Fuck him and fuck Palantir.
- Comment on Smaller rally/racing games I recommend you try! 5 weeks ago:
Wishlisted Art of Rally! Looks right up my street. Most realistic rally and racing games don’t hold my attention as much, I usually need a fictional element to keep it escapist.
Have you tried the EXO Rally Championship demo? I feel like it’s probably geared towards more serious rally fans who understand certain fundamentals of driving and tuning than I do, but it has a really fun setting and tone. The added complexity of the 360⁰ jets and the six wheels is a hoot. Most realistic rally and racing games don’t hold my attention as much, I usually need a fictional element to keep it escapist.
- Comment on Smaller rally/racing games I recommend you try! 5 weeks ago:
Make Way didn’t quite scratch my Split/Second itch, but it helped a little.
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
That may the problem, I don’t happen to have any of my FromSoft titles on PC, and the PS4 won’t talk to my 8BitDo.
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
That’s wild, I was just talking about using it with my actual old switch. I’ve got the Pro 2 (the one on the left) and it pairs between the switch and my computer pretty seamlessly. I’m so much more basic than I thought, reading these replies!
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
I don’t do competitive stuff other than couch multi, but it sounds intense.
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
Maybe I need to limber up mentally, jeez! I’m so set in my ways that when I try to imagine any of this, my eyes start to glaze over. Like I’m scooping what you’re pooping but I’m startled at how much my mind is resisting.
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
I like that idea about gyro aim engagement. I think the only place I’d use that is BotW, but I’d like to go back to that game someday.
- Comment on 8BitDo announces it's controllers now have Steam/SteamOS compatibility 1 month ago:
Antibody else find a use for the back buttons? I like the idea of them but I haven’t really found a great use yet.
- Comment on "I've Made The Equivalent Of Like 2 And A Half Undertales At This Point"- A Quick 'Deltarune' Chat With Toby Fox 1 month ago:
Toby is the best of us. I hope he gets hit by a bus made of praise.
- Comment on MrBeast scraps AI YouTube thumbnail generator days after announcing it: 'If creators don't want the tools, no worries' 1 month ago:
No allegations, but there’s still the Harlem shake poop video.
It’s on archive.org if you feel the need. I don’t recommend it.
- Comment on Reminder that you do not own digital games 1 month ago:
The suits found out what servers are. Worst shit that ever happened
- Comment on Reminder that you do not own digital games 1 month ago:
Agreed. If Steam goes under (not really likely), there’s still GOG & itchio.
- Comment on Reminder that you do not own digital games 1 month ago:
Piracy, in a word
- Comment on Reminder that you do not own digital games 1 month ago:
Multiplayer online games used to allow you to self-host so there was no obligatory centralised server. The game need not be deleted, that was purely a business decision.
- Comment on GOG News for Gaming/Linux/Steam Deck 2 months ago:
That’s hilarious, I felt weird not saying more. However it’s not quite 7 am and I’m rather hungover, so a comparatively lightweight read was just what the doctor ordered
- Comment on GOG News for Gaming/Linux/Steam Deck 2 months ago:
Thanks as always, what a pleasure
- Comment on France’s new laser rifle silently melts electronics at 500 meters — and Ukrainian infantry could really use it 2 months ago:
We could use a mulligan on the last 12 eons, give or take.
- Comment on The Return of the Pebble Smartwatch Is a Sign People Crave Something Unique 3 months ago:
Seriously how do you see the resurgence of the first, least feature rich smart watch and go write an article about people wanting “something unique”. Do they think what I’m really missing is a watch face with a quirky octopus?
- Comment on A court has rejected Elon Musk's request for a temporary injunction against OpenAI, OpenAI says Elon Musk "is finding out, facts matter - especially in court". 5 months ago:
Let them fight. With any luck they ruin each others lives.
- Comment on AI Markets Were Deceived To Believe In DeepSeek's Low Training Costs; They Are Actually 400 Times Higher Than The Reported Figure 6 months ago:
I fucking knew it.
- Comment on New Steam Console Powered by RDNA 4 Might Be in the Works at Valve 6 months ago:
What the fuck is that controller. God I hate it.
One thing I know about myself though, is my instincts suck and I’ll probably be eating those words before long.
- Comment on Kagi Introducing Fair Pricing 6 months ago:
Neologisms don’t get to tell me how they are pronounced. I decided what sounds best, and if others agree then that’s it.
Kagi rhymes with Maggie.