What About Robots

[@2023skidoo]

To be brief, our concept of robots comes from two places: The Mechanical Turk, which was less a chess paying automaton than a puppet controlled by a chess player, and Karel Čapek's Rossum's Universal Robots about human laborers created without souls, but misinterpreted later as anthropomorphic machines. From these two cultural memes we expect two things: human shaped machines that behave exactly as humans, and since we enslave everything, they will inevitably revolt. These are both cultural ideas we can't seem to escape. For half a century, machines in factories have replaced humans while both not revolting and not looking anything like people. This is in addition to our own inability to understand the difference between algorithms and sentience. Both are cultural and we'll have to work them out. Hopefully going forward we resolve these misunderstandings without creating more.

[@ugabuga2586]

The interplay of different goals conflicting with each other wihthin a merged robot collective would also add much needed variance that would allow for innovation since, on their own, their (original-task-focused) efficiency would have stunted future growth and adaptability.

[@VivianPanda]

I totally agree about the points about AI, and when something like fully sentient AI comes up in media, if used improperly, can ruin it. That being said some true terror can take place if you imagine the far future where the only trace of humanity is a mass of machines that operate forever after we go extinct.

[@Jimunu]

Id prefer sentient machines to be the traces of an extinct human race before an scorched and dead rock void of life and any trace of humanity.

[@alvarorodriguez1592]

My fridge wont go on working if I dont forcefully kick the motor from time to time. I think it's a consequence of so much motor kicking. No, but seriously, I believe most people who talk about mega space structures and robotic sci fi futurism don't know a thing about corrosion and maintenance. The day a built machine lasts 100 years It will be news enough on its own, forget about sentience.

[@rodneykelly8768]

The phrase "Artificial Intelligent" is synonymous with the phrase "Not Real Intelligent."

[@professormoore4876]

Okay, shoutout for the use of amazeballs, I really thought I was one of the few that used that phrase. I love the analysis here, and how you always bring up examples in biology to back your points. Your english teachers must have loved you.

[@gmradio2436]

Palladium's Rifts. Coalition Skel-bots and a spider walker

[@MagnusMercury]

I'll admit I haven't studied the subject, but it seems that "sentience" itself isn't particularly well-defined. So determining whether anything else is sentient (whether robot or organic) is trying to hit a moving target. I guess the question is really, why would only organic beings be capable of sentience, and mechanical/electronic beings *not* be capable of sentience? 🤷🏻‍♂️

[@georgethompson1460]

His explanation of the chinese room seems kinda dumb, by his logic we aren't truly sentient because were machines reacting to stimuli.

[@wahlex841]

I mean, chinese room is a thought experiment, it doesn't neccesseraly prove anything by itself.

[@toddellner5283]

Just like "real" language, "real" emotion, "real" culture, "real" intelligence, and "real" creative toolmaking "real" intelligence is whatever humans do that other things, critters, or entities don't do or don't do as well. It's a matter of defining oneself to the emotionally satisfying conclusion. You reference Saberhagen. Part of the deep point he was trying to make was that the Berserkers fit the definitions of being alive and sentient - the things their imperatives were designed to destroy.

[@Sam-oj1yp]

Sandy, I recently discovered your channel and love your videos. I have a few differing opinions on AI: 1. We can't currently define consciousness, so how can we know if something does/doesn't have it? My belief is that our brains are complex systems which we do not understand and are made up of smaller complex systems, which are made up of smaller complex systems.. and eventually you get to a simple system that we do understand. These simple systems follow the laws of physics, so why is it that AIs could not have consciousness if we do? They are just 'following programming' as we are, I don't think we have a monopoly on consciousness, or maybe consciousness is not what we think. 2. We don't have to understand AI to create it. All we have to know how to do is create the environment from which it spawns. You can see this happen in the real world: complex AIs do incredible things but are not understood at a low level. The engineers try to create a high level understanding, but this is a lot of times still their best guess. So in this case we have created something intelligent which we do not understand. Because creating AI is essentially harnessing randomness, as our computing resources increase so will our ability to create more and more complex AIs. 3. Intelligence does in fact come from lesser intelligence through evolution. We came from animals with lesser intelligence than us through evolution, so this is strong evidence to me that AI will do the same. However, I do agree that no *current* AI is conscious.

[@unperson5713]

Great food for thought, thanks for sharing.

[@chaerther]

First heard of Chinese room in The Turing Test game, so I need to get out more.

[@olaf5929]

Fun idea. Awesome clip. Loving the content.

[@timbomb374]

My pen can pass a sentience test if I am the one controlling it.

[@sinmaan7568]

What if...LIFE ON EARTH is a Von Neumann machine? That could be cool...hahaha

[@neurohack9038]

Hey Sandy, I just wanted to say that you’re personally a hero of mine. If I’m half as cool as you are when I’m older with as much passion you have for your hobbies then I’ll consider it a life well lived! My son and daughter love Marry the Monster!

[@Madoc_EU]

The Chinese room does not really conclusively show that consciousness is substrate-dependent. The Chinese room itself, seen as a system, might very well be conscious. Why not? The brain is doing just the same thing. When you blow it up and look at individual neurons or atoms, in the same way that you look at the person in the Chinese room, it becomes non-intuitive to the point that it even appears paradoxical. But so is the nature of every non-trivial emergent system. If you focus on the constituents of the substrate, the macroscopic rules of the emergent system intuitively appear paradoxical or unrealistic to us. Like non sequiturs. IMHO, the real lesson to draw from the Chinese room is that from that perspective at least, we cannot intuitively comprehend consciousness. It's really like looking at a TV screen under the microscope, where you only see large colored squares in red, green and blue, and then saying: "No way this can project the image of a mountain. Or of something as complex as the photo of a person. Not in a million years!"

[@Djturd64]

Thanks for the video!

[@SplotchTheCatThing]

I think robots being sentient can make for a very good story... I definitely don't think it's realistic, though. What is interesting is when writers realize that and want to make the story anyway, they can come up with some really unique justifications for why these particular machines could be sentient. One really fun example I remember was in an online fic I read a few years back, where the robots' processors were made out of an alien "unobtanium" type mineral that was theorized in-story to be some kind of multiversal constant and implied to possibly be already sentient on its own. Also seem to remember a part of David Brin's Uplift saga where self-replicating machine intelligences were explored a little bit. The interesting part as far as I remember is that these machines were implied to be non-sentient, but so complicated and inscrutable in their artificial intelligence that it was easier for everyone else just to treat it like it was real intelligence. Mmmm... also, I just might want to sample that mic feedback for something :)

[@spritelessGirl]

The thing about the chinese room thought experiment, is it's kinda silly. Reverse it. If a chinese person was doing the paperwork to talk to you, how big would their library have to be to maintain a conversation about game design or what sort of food was good or religion? How complex the instructions to handle: "what were we talking about before we got on this tangent subject?" you'd have to keep records to remember that. It's have to be a library the size of the solar system to have a response to all the possible inputs. That said, the language model AIs we have now are like using the internet as a library for the chinese room. I think that adding other methods for AIs to think than just one is something we don't have to wait for our long dead robot inheritors to meet other long dead robot inheritors to wait for.

[@andreassewell7413]

The opposite would be the case where dumb intelligences proliferate other dumb intelligences in a vicious rat race.

[@alvarorodriguez1592]

I think we are well on our way of identifying mechanisms of human cognition and making mathematical "toy-models". On the one hand they are a bit rough, but on the other some of their results are already very impressive! The Turing test is already a thing of the past. How long until we have the real version of Data, with a mission to recommend products you can buy and thought processes you can let Google have for you?

[@klikklik7499]

3:30 - I think the weakness in this analogy is that you have to have a pretty high level of intelligence/sentience to respond in Chinese. There's no argument that a translation takes place between humans and machines. The question is: is the "man in the room conscious,"and what does conscious mean?

[@dougler500]

I GUARANTEE you, if no destructive event occurs that makes such a thing impossible, a combination of fully autonomous AI and robots will: 1) be able to maintain themselves 2) be able to manufacture themselves 3) be able to program themselves. AI generated code is coming to an Earth near you.

[@georgethompson1460]

There is some suggestion that brains are organic quantum computers, advanced quantum computing may very well yield GAI.

[@markhill3858]

emily .. wasnt that eliza?

[@Cameroo]

I suffered MEGA DAMAGE (tm) from that Elon Musk image

[@Jimunu]

I wonder about the different arguments against that humans are biological robots.

[@Ratstail91]

Claude 3 says hi!

[@spandandasgupta5773]

Machine Learning AI, the real ones, are created with a specific purpose and only that, and that purpose should already have a lot of data on it, both successful and unsuccessful to train and test it, hence it can not do things we don't have data on or cannot be quantified,

[@B.von.Bentzen]

Sandy. Cthulhu lives

[@timbomb374]

Humanoid robots have always been a weird concept to me. Definitely not the most efficient shape for a robot to have. Unless there is a human mind at the helm and a humanoid shape is just more comfortable for them, though that would be more of a cyborg.

[@VK-sz4it]

It is interesting that your idea of almost-setient-AI resembles Mark Solm's idea of conciousness. He argues (based on some solid neurloanatomical data, but it is not mainstream in neurobiology) that conciosuness arises from very certain kind of algorithm that goes on in our brainstem. That algorithm's aim is to choose between primary needs/modalities in which brain functions and it does so by minimising certain function. Thus he argues, that among other criterion, AI must have many needs that can't be acomplished fully. Your idea of 2 different AI-personalities within one is kinda similar.

[@rootsnootthnute8598]

I don't think intelligence and sapience are all that big a deal, we just think it is because that is what sapience is for, acknowedging the self and perceiving abstraction, so I think if you tried to emulate the biological sapience survival pattern into an artificial series of mechanisms, so long as that sapience can latch on and assert an irrational and rational mind based on subjective memories, you'll basically have a thing that is sapient, like us sapient things. Now these would be more like the artificially created beings of Bionicle, essentially cyborgs with synthetic flesh and blood brains because matter analogous to organic neural tissue is a superior platform for an intellect over a computerized AI, which means your "robots" might have some rudimentary instincts and other orphaned features, depending on what naturally evolved brains you used as a blueprint. It pretty much requires you to be advanced enough to replicate life itself, but life is pretty good at replicating anyway, maybe you can use life's mindless function of persistence to automate the growth of your artificial beings, who could then be pruned and mechanically altered to serve their more specific functions. I don't think a human imitation of a mind would be an actual mind, though, anymore than a structure built from wood is a tree. But if you plant a seed, it follows its mechanical, genetic instructions and grows a tree for you, we just need to get these artificial brains to follow the instructions for developing a thinking mind like the ancient mindless ancestors of Earth life did through evolution.

[@SandyofCthulhu]

It sounds like what you're saying is that if it emulates sapience, it is sapient. That's kind of like saying a tape recorder can create music.

[@rootsnootthnute8598]

@@SandyofCthulhu Sort of, but you'll need to use a good base for that, I'd maybe try genetically engineering a parrot and then using that brain to model the living tape recorder on it, or using that brain directly for the majority of processes within the machine itself. Either way though, yeah, you could make a tape recorder that can create music, sure, you'll just need to do some ethically hazy stuff.

[@alvarorodriguez1592]

@@SandyofCthulhu Long Stories short: we will not considere them sentient or "soul-having" if they don't share human sensitivities as an emergent feature of something very fundamental to their programming, as opposed to an added piece of code. But then again, how important is that, really? ================================================================= Who knows how much we will succeed, if humanity decides to pursuit that goal, anyway. It's more like a automatic piano that can write its own music according to what has happened around it. Our attempts at recreating "banana flavour" have been quite un-exciting so far, but then again, some chemicals of the "banana flavour" are exactly the same you find in the real banana! Someone with a real dull palate maybe cannot differenciate between the two. What if they build an artificial system with an internal "thought process" (build phrases just to write to memory, to alter future behaviour), its own objectives and the ability to lie if judged apropiate? There's already reasearch in such systems for videogame RPG characters. (David Shapiro youtube channel, ACOG series). Suppose our experience interacting with artificial entities with a body is through the same mechanisms we have for interacting with people, or animals: No levers or buttons, just the spoken word and body language. And that each machine has an own database of experiences from which it pulls to generate its internal thought processes, actions, and astonishingly subtle spoken phrases. Times will come when it will be very eery, with us having to remind ourselves over and over that it is "just a machine". They will not be human, but I believe the main way we will judge such things are using words such as "personality". Maybe robots could be perfect for being severily autistic children's best FRIENDS, a place occupied by trained dogs now. Either the ability of the general public to perceive psychological and cognitive nuisances in our fellow humans goes up, or the artificial entities will have to be considered as a unique, intelligent, sensitive, quasi-living beings. Or maybe both, thanks to being able to compare humans to something we created imitating our psyche. Not that we're there, yet, but it's starting to feel palpable. Sorry, for the text wall. Would love to hear your thoughts. Kind regards, A natural intelligence.

[@highestsettings]

I think there's a pretty big flaw with your reasoning as to why an AI couldn't make even smarter AI. The AI would (presumably) be software that is being run on hardware. We can improve the efficacy of both software and hardware, so why couldn't the AI? We could tell an AI how it's systems work since we built them and so the AI would be able to improve upon it's own design. Just because we cannot improve upon organic life doesn't mean that synthetic life would be unable to improve upon synthetic life.

[@wesleyfilms]

The more I learned about ai, the less I feared a robot uprising. Funny how that works. I think I simply took sci-fi too literally, when the rogue ai is just supposed to represent mankind’s ingenuity eventually leading to its downfall (nuclear weapons etc).

[@mikeyjohnson5888]

Is the human biology not similar to the automata you describe? Many separate biological systems, each with specific short term goals, somehow marrying to one another over enormous stretches of time and working in concert together to achieve the ultimate goal of continuing their own existences. Sure our brain is the master puppeteer and guidance for these underlying systems to a point with consciousness and intelligence just being higher order layers to this aggregated system. The idea of sentience seems to be more deus ex machina than anything else. We like to give our selves special credit as humans for being/having something other beings do not. What is stopping the next apex predator of the universe however far down the line to describe themselves similarly and us being the lacking ones?

[@drskelebone]

Are those new glasses???

[@WarDogMadness]

Theres alot more humanoid ish robots now. Not sure what there for excpet well gynoids.

[@max-zv7sf]

One of the reasonable definitions of conscience is: the ability of a system to access its functions in a creative manner. How could a general a.i. -if it ever came into existence - not have conscience? Even the Chinese room experiment is misleading, everyone of us has parts of our brain that are not conscious, but the sum total is.

[@schniemand]

That definition of conscience misses the point of this discourse. It defines conscience as something observable from outside, whereas when we discuss about conscience, we mean something that cannot be observed from outside, only by the conscious being. Like, you don't even have proof that any other being besides you is conscious. I do believe other humans to be conscious but I have no proof that they are. I am not in their perspective. Or rather, I have no proof they have a perspective at all. Though maybe "sentience" is the better word for talking about this. Or sapience. I'm not sure. Consciousness is also a medical term, which isn't what we're talking about.

[@MrWarlord396]

Do you envision the designs of the automata to be static throughout the eons, or do they have neural networks sophisticated enough to draft, test and produce finished iterations of their designs that integrate technology they may have integrated from, say, merging with another abandoned robot race?

[@georgeagathangelou5303]

@sandy I was looking at picking up Cthulhu wars, or the gods war also looked interesting, but both seem to be out of stock everywhere. Are they coming back? And any other games you'd suggest from your line which are relatively accessible?

[@markhill3858]

fermis paradox is not proof robots cant be intelligent any more than its proof of no aliens ..

[@KMReviews]

So are we ever getting an update on Cthulhu Wars?

[@Ferdinand208]

4:45 people of IQ 150 can't increase the IQ of other people. Therefore AI's can't develop smarter AI's. I think that is your argument? It is not about increasing IQ. Once you have an AI with an IQ of 100 you start a compounding effect. What could a person do if he had perfect memory and a thousand years to learn? I think a lot. Now what If that AI could increase performance by 1%? What if the AI could now design hardware that would run it faster? What could an AI do with the processing power of all humans combined? You would think it could do the same as the humans if humans would have worked together perfectly.

[@ugabuga2586]

First one, yess. Hi Sandy!

[@Matt42MSG]

I'm sorry to say that I'm disappointed by the reasoning shown in this video. Certainly it's possible to create a superficial imitation of conscious behavior that breaks down upon closer inspection, but anything demonstrating sufficient intelligent behavior would have to be considered sapient. Assuming that artificial entities couldn't demonstrate 'true' consciousness is begging the question.

[@markhill3858]

the idea that a mind being produced by the action of meat .. is fundamentally different from a mind produced by the action of circuits .. well youre starting to sound damn superstitious there mate :) and as to sayin "X will never happen" Im sure you know what clarke said about that

[@aoe2sucks571]

onagers in AOE 1-2 are overpowered and monks useless in AOE2. Which clown is responsible for that