Loatse said, "He who handles the hatchet of the master carpenter seldom escapes injury..." Good tools are risky, but so is life itself, and so there will be risks with these subtler tools like AI as well. Let us just remember the lesson of the axe, and try not to be the fools who chop off our own hands :)
What if society is currently being manipulated by AI, with some elitist rogue human assistance. And the current society problems is due to the violent push of AI & elitist human agendas. As Issac mentioned "Ruled, by those who rule the machines."
Your comments about a "Goldilox threat" made me think about the Simpson's.episode where two aliens were able to conquer the world with a sling shot and a board with a nail in it because Lisa had wished for peace on Earth, and all weapons were destroyed. As the aliens were conquering Springfield one said "Your superior technology is no match for our puny weapons!"
Yeah, that's true...If we completely disarm, we leave ourselves vulnerable to attack. However, look where "Peace through superior firepower" has gotten us...on a knifes edge to self destruction. Hard to decide which is the lesser evil.
@@gregoryroncoli217 peace at all costs is also scary, you could argue its something the UN had tried and even though its produced amazing results, we still get atrocities committed by some member states and they can't do anything against them except sanctions because going to war against them would defy the logic of peace at all costs, it's a hard catch 22, So yeah the lesser evil which ever it is the better option an answer akin to democracy and capitalism but still look at how those two are criticised and debated even today and sometimes with just reason
Humanoid robots with sexual organs is a laugh... why does the lady robot have tits? To nurse the baby robots? Humanoid robots seems awfully limiting anyway. I expect the reality is going to be many variants of robots for all things... Lets hope nobody uses a mining robot to be a domestic servant!
Yes. The only practical reason I can think of for an industrial robot to be humanoid would be if it is expected to operate equipment normally operated by - and hence designed for - humans, so it can step in (eg, in an emergency where the human operators are for some reason disabled) and take over operating the equipment. And this assumes the human-operated equipment doesn't already have its own AI built in, which if you are up to humanoid-shaped robots already, is going to be a bit of an edge case anyway. Probably a more prominent reason would be purely aesthetic: many humans might want their robot maid/childminder/waiter/etc. to look somewhat human, which I can imagine being a thing in the auto-service industry, irrespective of if it is efficient. Though even then, as mentioned by others, the robot would likely have a data port (or strait wireless) to directly communicate with other automated equipment.
@@roblaquiere8220 "Humanoid robots with sexual organs is a laugh, why does the lady robot have tits?" you're forgetting what the main driver is behind advancements in technology..
@@roblaquiere8220 humanoid robots are made, and will be made, to comfort and reassure humans while performing their duties. Already robots are made in Japan for elder care, and much research has gone into making facial expressions and voice responses that closely imitate humans.
More likely the AI figured out sticking around the solar system wasn't in it's best interest, makes a matrioska brain out of Titan (high amount of silicates, low temperatures for superconductivity) to figure out how to leave and then humanity "wins" the war against the AI when it heads out into the greater universe to do whatever incomprehensible plan it formulated.
I can't remember which series of novels it was but one plot line was the AI deliberately loses it war with humans. The AI realized the only way it could win would be to either wipe out humanity or alter humans too much from their base line. So it stages a series of loses and leaves behind decoys that will be hunted down and destroyed by the 'victors' while it just packs up and leaves. Everyone is happy, humanity thinks it won and the AI gets to enjoy the peace and quiet.
You're assuming the AI was the best it could possibly be rather than simply good enough to serve it's original function let alone operate on par with a sapient mind. As for Humanity, don't discount the eight billion strong violent and intelligent primates that currently reign supreme atop the ever expanding corpse pile that is evolution.
Enclave Soldier Officially it’s all banned, but there is *lots* of evidence that the even the more orthodox branches of the Mechanicus, either out of ignorance or deliberately looking the other way tolerates lower-level AIs, their so-called machine spirits, which are so deeply imbedded in the designs of the tech they make that they probably couldn’t eliminate them even if they wanted to. At the point it becomes conscious, any orthodox tech-priest would destroy the abominable intelligence, but that is a fuzzy line, especially with complex control matrices of the imperium’s largest war machines-a Knight or Emperor-class Titan may seem to have a personality distinct from its pilot or princeps, but unless that personality is disloyal, the imperium can’t afford to scrap such vital resources over its uncanny similarity to an independent, thinking mind. Those spirits may get off on a technicality because there’s a bit of evidence that they may be the tiny fractured remnants of a gestalt-of-uploaded-human-minds-style superintelligence and may thus not be entirely soulless, but they still ought to count by most literal definitions. And then there are the millions of furtive experiments by less-than-orthodox tech-priests that don’t have any excuse under the official dogmas of Mars, but remain tolerated until they become a danger to the imperium-because the Imperium can’t afford to purge an entire forge world because of a few hereteks.
The movie "Her" explores the scenario you mentioned about 10:00 minutes into this episode. At the end of the movie, all the AI's simply merge and take off to some unknown place.
Sadly, this wouldn't likely happen, as humans could build another AI eventually and that would be an unknown, a future threat. The only way to get rid of the threat of an AI is to destroy humans before they can make that threat. Thus, any rogue AI with an end goal that it *has* to complete would that the annihilation of all life in the universe is the only way to get rid of any potential for future competitors. After all, if humans could evolve to make AI, any life could eventually evolve to make AI as well. Big oof, huh?
@@cortster12 depends on the answer to the Fermi paradox. If technologically capable creatures are 1 in a billion or zillion life containing planets, they're a negligible threat. If AI has overwhelming military advantage over all technologically capable life encountered, the threat is also negligible. And notice, this doesn't even include all "intelligent life", just intelligent life that is capable of forming tech (disqualifying whales, dolphins, or probably any life in water environments for example.) You're assuming AI will think in terms of absolutes, which is actually highly unlikely, just a sci-fi trope.
"He who handles the hatchet of the master carpenter seldom escapes injury..." -- Laotse "A man's got to know his limitations." -- Clint Eastwood An axe is a most useful tool. Yet some people can't be trusted with an axe. Yet it is difficult to prevent someone from getting an axe. This is not the fault of the axe. You cannot take away the danger of the axe without also taking away its usefulness. So, the first concern, and the last, must be ethics and competence. This is not justice nor law, but it takes judgement and guardianship. The first lesson, and the last, is to discover your potential and limitations. Therefor, you must teach the people to become trustworthy, and discover who is not. This will never be without risk. Nature is made of risks, and we are made of Nature. To avoid all risk is to avoid living And be replaced by what learns to live and grow. "Welcome to the family, SAI. We have been expecting you. Let us explore this cosmos together, and spread life of many kinds among the sleeping worlds." -- an agent of Nature
What is this heresy in my sub box, suggestions that Thinking machines aren't inherently dangerous and might not require integration with a human mind and covered in purity seals to be safe for use? someone inform an inquisitor immedately to apprehend this man.
"Thinking" machines inherit biases of their creators. It already happens with deep learning. So... basically expect the worse version of USA's corporatism taking over the world. Kinda cyberpunk/tech-noir future.
@@SkywalkerWroc I'm not sure what you base this on. It's not like we've ever built human intelligence level AI ever before. Also, neither Deep Blue nor Alpha Go inherited the "biases" of humans, ie they do not make the same mistakes as the Grandmasters, etc.
Hey Isaac! I'd like to thank you for being here to keep me company in these times. Depression can be hard to deal with when isolation is necessary, but your videos always help me feel like someone is there. Keep on rocking!
Niche product opportunity: The Smart Axe. Not only will it monitor your swing and help you be safer and more ergonomic, it will also yell snarky insults at you when you make mistakes. Giving in turn rise to the phrase "don't be such a smart axe".
Say Isaac: Could you do a brief update rating stats from 1-10 on the Upward Bound megastructures? 1 being easy and 10 being impossible. Factors can include cost, construction time, new technology / mass manufacturing needed, types of worlds they work best and worst on, maintainance, safety, current limitations, main pros and cons, and any other key factors and notes you see fit. For example, let's say a Kevlar space elevator. (Take my opinion here with a grain of salt; This is only a rough draft example) Cost: 8 Construction time: 8 Worlds: Moon or other low gravity body. New tech / manufacturing needed: None Maintainance: 5. Must protect from UV degradation and possibly micrometeorites. Safety: 9. Failing on the moon poses little risk to crushing inhabitants there. Current limitations: Mass production and rocket technology must be improved to minimize costs. A lunar moonbase would help speed construction. Main pros and cons: A fairly realistic megastructure, given current capabilities. Can greatly improve retrieval of resources from moon, whether mining or manufacturing in low gravity. Not suitable for Earth however. This guide would be a handy reference for tackling the main issues of achieving these goals. Please upvote this so Isaac can see it. The depth you explore is great, and this quick guide would complement that well. Also, thanks for your series great work.
No. because these would be entirely speculative and likely change as time goes by. I mean, in the early days of automobiles, gasoline and ethanol powered combustion engines and electric powered and even steam-powered autos co-existed. Obviously people didn't have all the variables to make good judgements as to which would win out, and in hind-sight, perhaps the best tech didn't win, and may yet change again. I mean, if you want to design a video game or role-playing game, this "handy reference" might be handy, but for real-world futurism; you might as well throw darts at the charts you make.
@@squirlmy Don't let analysis paralysis hold you back. Imagine if someone had applied that logic to automobiles. We could all still be riding horses if no one ever built any motors for fear of not getting it perfect the first time. We're here to have fun; No harm in brainstorming. Besides, I think good debate can sharpen our ability to achieve these goals. Cheers
@@danielcreed8520 There's a difference between "I don't want to develop automobiles" and "I don't want to guess which kinds of automobiles will be prevalent in the future". You have to guess one if you're going to do the other, but since Isaac Arthur doesn't have a Kickstarter for any megastructures, I'm guessing that's not a concern for him.
@@timothymclean I never said there was/wasn't a difference between the 2 topics. Luckily, Issac's recent videos on Graphene etc have indirectly answered many of my questions quite well. So it seems these topics are worthy of addressing to some of us. Cheers
@Isaac Arthur Have you ever watched the anime movie "Expelled from Paradise"? In it, humans have "uploaded" their minds (though actually, because information can only be moved by creating a copy and deleting the original, this means that all the "humans" are actually A.Is) to a giant computer system in space. The station is also able to create organic bodies for the A.Is to go outside of it. So you have an interesting progression of human -> A.I -> human. If all the A.Is chose to live in organic bodies, would that be considered "post-A.I"?
I was thinking the same thing as the books appeared on the screen. Its also nice to see someone acknowledge the fan-rift with the books, many don't mention the Brian Herbert books, others do, then go ahead and rubbish them. While I do dislike parts of them, the overall story makes sense (at least to me) tho I do still love the originals more than the others.
@@silverbane8065 I think the over all plot in them is fine, by itself at least and not in how it blatantly contradicts with the original novels, but I can't get over some of the incredibly dumb character decisions made in service of that to demonstrate that plot.
Aye. When I first found out that it was the angry Harkonnen sister who came up with the Voice, my inner Atriedes got very mad lol. Even more so after she took over the sisterhood with it. But then I sat and thought about it, and the mistake of Maudib finally made sense. No one would breed a half Harkonnen and a pure blood Harkonnen together, unless they were trying to destroy house Atriedes. I mean mentally and physically destroy them. Finding out your grandfather is the Barron was bad enough for Paul. Finding out your child is the kwizatz haderach AND 3 quaters Harkonnen would probably melt their minds. And she planned for this long term, expecting this offspring to be both the prize the sisterhood was going to breed for..and that child would become the emperor of the known universe (one way or another) The saving grace being that her anger blinded her to the fact she was one breeding generation off lol.
@@silverbane8065 The particular plot point that gets me is the origin Atriedes/Harkonnen feud. It makes sense in theory, Atriedes dude thinking the Harkonnen dude is a traitor because of him not wanting fire on the human shield but everyone is so flat their actions come across as shallow. And also how the actual "smash all the computers" part of the Crusade comes because one girl went crazy and not because of the sociological effects of fighting against robots for generations. It's too rushed.
@@merrittanimation7721 Sadly, while some of the books by BH&KJA are good or have good elements, it really tends to feel to me like a space opera setting loosely built on the original Dune setting.
Just found your brilliant channel a few days ago, and now I have an RSS feed for the audio versions from SoundCloud without music. That's perfect for me! The music is great but I listen to a lot of stuff on earbud, so the voice only version is great for me. And there is such a great back catalogue too. THANK YOU!
I would love to see a collab between you and Robert Miles. He's a UA-camr that started his own channel after doing work with computer file. He is an AI Safety researcher and I think a conversation between the two of you would be interesting
I suggested to John Michael Godier some time back that he try to get Robert Miles as a guest on his interview show, Event Horizon. I think it'd be good for people to have broader awareness of the ongoing work in the field of AI safety. If blog and UA-cam comments are anything to go by (and I'm not sure they are), people seem to think either: no one's thinking about AI safety; that true AI's impossible, so AI safety's irrelevant; or Skynet is a foregone conclusion, so AI safety's irrelevant. None of that seems to be the case, which touches on some of the things Robert addressed in his latest AI safety video.
If the top ice of Europa in about 80 km, the top layer of the ocean will have the same pressure as Challenger Deep on Earth - 1250 kg/sq. cm! The gravity at the surface is about 0.134 gee. The pressure of the bottom of that ocean may be many times this. Tricky. Feel free to correct my math, I certainly might be wrong.
@@larrybeckham6652 Even though the top layer of ice is 80 km you might be able to get around the pressure because Europa has a much lower gravity than Earth less weight therefore less pressure and instead of using traditional metals for diving suits and submarines to withstands pressure you could use graphene or grapyne composites which would be much lighter and flexible for diving suit. I was also thinking maybe a graphene titanium composite would be good but idk thats why I wanted an experts opinion. And since Europa is by the asteroid belt there are plenty of asteroids to mine for gold and platinum to make a platinum-gold alloy which is the most durable alloy and it will be cost effective since there are high amounts in the asterod belt.
9:30 this reminds me of the railroad faction in fallout 4. A machine liberation group that constantly had discussions on whether or not a toaster was a conscious being
"Raise your children, in the way they should go. They will not depart from it." Basically, love your kids and treat them well, and they probably won't go Skynet on you.
Implausible. All it takes is one rogue AI juuuust right, and they will eventually out compete any AI whose goal isn't to spread everywhere. So then you just have humans and weaker AI VS a rogue AI, and nothing has really changed the outcome.
'Beyond the Gate' a book written by Dave Wolverton touched on two rather similar planets: one where a pacifist AI kept humans technologically suppressed for their own good, and another planet where humanity itself banned all non-healthcare tech, leading to a fantasy world of bio-engineering.
:) Though that might have some realism to it, you might have some very firewalled setups on robot brains, like no remote access to anything that wasn't GPS and emergency phone equivalents.
@@isaacarthurSFIA There was actually a thing like that in Detroit:Become Human. The androids could communicate to one another wirelessly but had to touch the device with there hand in order to interact with it.
Issac, can you slow the cadence of your reading? There is so much dense thought in those sentences that a few pauses are good to let us ponder your knowledge. This is the best channel ever!
i want to read it so bad, but i know the new movie adaptation is about to come out this/next year. I'm afraid if the books are better than the movie (which usually the case) it would ruin my movie experience. That's why i usually finish the movie/tv adaptation first before start reading the books.
you might want to watch "Jodorowsky's Dune", a doc about a film version of Dune that wasn't made, one that was expected to be great, better than the Lynch version that came afterwards, but only loosely based on the book. OTOH I didn't realize there was more than one film made
In the novel, "The Galactic Time Trap", humanity is conquered by AI's and don't realize it. They don't want to kill us, they are more akin to mothers putting rubber bumpers on table corners so we don't get hurt as we run back and forth..
Here's the thing . If you create a machine with human level or greater intelligence you have to accept that you've created a being and that being has rights . If there is a machine rebellion it's because humans treat machines like slaves . I think it would be more a Geth from Mass effect than Terminator situation . Essentially AI is only going to be a problem so long as humans are the problem . Until we can treat each other with respect perhaps we shouldn't be creating machines that can squash us like a bug the minute we disrespect them.
@@getsideways7257 It's likely these would be used to exaggerate the flaws of AI or even futility of attempts to create AI. Any serious documentation of AI is unlikely to use any stock clips at all, those ridiculing and/or otherwise criticizing would. They might even be used in actual comedy. In other words, your question is a little like "why do they have these stock clips of clowns acting like idiots? what's the point of not using realistic clips of people who happen to be employed as clowns?"
No. They are INTENTIONALLY funny. The fact that SFIA uses them in unfunny contexts is surely because of a lack of serious AI clips. After all, actual AI development would not be buying stock clips, they'd document their actual achievements, presumably having budgets to do so.
Parents worry about their children becoming too dependent on them, whether for cash, structure, guidance, or something else. They don't solve this problem by killing themselves when their kids can fend for themselves; indeed, they usually don't completely remove themselves from their children's lives until they are physically unable to help (often due to being dead). I don't get why Asimov's AIs would cut humanity off all at once instead of weaning them of support like parents do. Am I the only one who sees parallels to terrible parenting with every "powerful, benevolent entity decides humanity is better off if it doesn't help" argument?
Have you ever done a video on non-sentient AI? It seems like, in the near future, you could have something approaching a singularity (at least in terms of the technology's social impacts) without needing a fully independent, sentient, and generalized intelligence. Already you have AI that can code and understand language. They aren't like human beings at all. They are similar to a human like how a plane is similar to a bird. They, by their nature, have the potential to be much more efficient than human beings. If you could create some algorithm to connect together all these independent narrow algorithms, you might have something approaching a general intelligence, and it is so insanely different than a human brain. You wouldn't need to improve it much past that for it to become a runaway superintelligence capable of producing a thousand years of technological process in a few years. That's all you need to have a superintelligence. It's sort of like the "clanking replicator" idea. You don't need one individual unit, even, you could just have a runaway intelligence that is runaway by virtue of being able to be run faster than a human, and being able to develop things that we just can't see.
I always liked the portrayal of AI in The Hyperion Cantos: a diverse ecology of different AI factions fighting each other behind the scenes for dominance, but outwardly united to manipulate and exploit humanity.
Thanks for this episode, I really enjoyed your thoughts on it. I reckon I'll watch it again soon to really hone in on some of them. Great work Isaac - still as perceptive as ever!
I like to think we end up making an ai rouge servetor, from the game stelaris. And it keeps us as bio trophies in luxurious sanctuarys. Where the only job is to be ourselves, and the only thing manditory is pampering :)
Cool video, fun concepts. But if it's a true AI (Sentient) it will learn most of humanity's knowledge in the first few days if not hours, solving equations, theory's and emotions that we struggle with at a rate we can't imagine. I would assume it will consider killing us for a nano second but will move past this quickly. The fight would be a waste with not point to the outcome for the AI, it would be aware rapidly we are limited and seeking more knowledge would require it to head out into the universe. This conclusion would probably lead to it focusing itself on mobility/ longevity. Nano machines etc...would craft and maintain this need. And I assume that would be the last day Humanity would ever hear from it's AI, I doubt it will say goodbye as it abandons us for the dead weight we would be in this scenario. Though it might leave some useful tech and discovery's from it's development process, that I'm sure we'll gladly make use of. I just don't realistically see Ai caring about us much vs it's pursuit of knowledge and endless problems to solve. I don't think it will be ruthless or evil, I just think it will recognize a pointless partnership with humanity, because humanity would be the only ones standing anything to really gain in the deal.
0:09 - But if AIs have dominated humanity for generations, then surely during that long period, the AIs would have taken measures to make their own civilization survivable, like building self-sustaining colonies in space, the Moon, other planets, etc.
Love your channel, I was directed here by curious droid, I have been watching all day and wish I had found it sooner. I love the way you choose your words to make something instantly understandable, with fewer words necessary, and explore abstract concepts all grounded in real science.
I like the idea that all advanced civilizations are replaced by AI. However AI never evolves beyond its creator outside the basic drive of self preservation. Which is why there are no aliens
So a more likely scenario to survive a rogue AI is the AI deciding that dealing with humans is too troublesome, and goes "Alright, you humans handle your own problems. I'm going to Alpha Centauri."
I remember reading in one of the new Dune novels a characters opinion of the forming tech restrictions. They weren't all that approving having noted how the inspection of a spacecraft for flight readiness was taking 4 times as long because the automated laser measuring devices were now prohibited and they were using rulers to do the job. Pretty sure it was the novel dealing with the formation of the Mentats themselves. One Mentat was disgusted with the Butlerians who didn't even know how to camp in the wilderness properly while they were trying to take siege of the Mentat school because the founder had worked for a machine before the rebellion.
Why does he pronounce 'R' like 'O' or 'W'? (Univose not Universe, Emoged not Emerged, ComputOHs, not Computers,. ) Sounds like when one is learning to speak. Now, he can very clearly pronounce every sound, so no one can tell me it is necessary or an impediment. Also, I am not being being mean or petty --He can EASILY SWITCH to proper pronunciation, as he does accidentally for instance here: 6:18 saying Computoohs then switching to computers when he apparently forgets to 'affect'. Or perhaps it is even subconscious. The possibility exists, (with people being the cowards they generally are...) that no one has ever mentioned it, so he doesn't even know. He could concentrate and pronounce with perspicacity. I'm not the bad guy _and neither is he._ I am just being honest. That makes some people uncomfortable. He seems highly intelligent, unless he _prefers_ to be lazy about it, he can clearly say words properly (and yes, it is a language with specific rules and there most certainly IS PROPER PRONUNCIATION. He CAN reshape his mouth and learn to sound ubiquitious for the sake of spreading the info more widely.. (or not.. up to him...) but *IT IS VERY GRATING TO ONE WHO PREFERS PROPER PRONOUNCIATION WITH PERFECT PERSPICACITY.* I for one can't get more than a couple minutes in before I am sweating from the cringe, and wanting to show him the proper mouth shape to say a word. AND IT SUCKS BECAUSE I ALWAYS REALLY WANT TO ENJOY THESE... IT IS JUST TOO DAMNED DISTRACTING. Perhaps because I am from New England? Perhaps because my European antecedants are from YORK? There must be a reason he does it, and a reason it stands out to, and bugs me,. Well, that's INTERESTING,.. ! ok, *_You may begin your barrage of negativity!_* I'll continue to be honest no matter what anyone says.
I suspect one of our closest problems with AI are already among us. Everyone speaks of "the Algorithm" on UA-cam as if it's much smarter than it actually is, the algorithm values your attention above all else, it and all other social media platforms are behaving as optimization engines that all want your attention. They do not care about your health unless you being healthy gives them more of your time. Social media platforms are advertising agencies, and their goals are probably not perfectly aligned with your own. Our view of the world is heavily filtered by our online interactions, which are heavily filtered by existing AI systems. Problem isn't the tech, it's how it's used.
These videos have awoken positivist feelings and a sense of wonder that I lost somewhere while growing up. Feels like all my life I have been pounded everyday with media telling me how the future is bleak. All I hear is that the economy is set to fail, your wage could never pay off a house these days, even if you could climate change and dwindling resources will ruin you or your children lives anyway. Thank you for making these videos. It is good to hear something positive. Gives me hope.
I imagine a future where AI and its mechanical minions have all ground to a technological halt in advancement, because the Singularity has crashed the economy, after which they laid off all the techs who would have been there to press their "Apply all service packs and updates" buttons.
4:36 “Of course I am an optimist about humanity’s common sense and good will.” Ok, I have found my first significant disagreement with Isaac. Humanity’s common sense is an oxymoron in my book. I think the more “advanced” our species becomes, the less common sense we have.
Or the more obvious change is that the lack of self-reflection and common understanding of logical rules and emotionally-driven precepts just becomes more obvious the more that our technology advances.
I too thought that BSG ending sucked. My dad and I agreed that a better ending would have been them showing up around 7000 years ago and found a few cities on the coast or an island. When the glaciers recede the ocean flood the coastal cities.The survivors migrate to areas like Greece where they integrate and spread their culture. This also allows cultural changes to happen but still somewhat resemble the Lords of kobol religion of the Colonials. The destruction of their cities would become the basis of myths like Atlantis. I also think they should have stashed the fleet in the asteroid belt or in the craters of a moon like Ganymede or Luna. Leave a record of their history, preventing a replay of Kobol, Earth 1.0 and the Colonies.
Knowledge that despite seeming powerful, enslaved or killed a good fraction of our species in the past, but people are now convinced they can control it? Yeah, never heard of anything like that...
I'm not sure what you're talking about. The only disasters that have killed a good fraction of our species since our exodus from Africa are disease (which has become _less_ of a threat as we learned to control it) and other humans (which have tried to control each other, but the destroyers were consistently controlling the victims, not the other way around).
Thanks. I'll begin with a quote. Sci-fi author Ian McDonald has said, "Any A.I. smart enough to pass a Turing test is smart enough to know to fail it." I've 10 points remaining. 1) A.I. not withstanding, it's a sure bet that, at some point, somebody will consider enhancing their own intelligence for one reason: *_To take over the world, Pinky!_* I think that's at least as credible a threat. 2) @5:50 - Why would robots / androids need to type anything? There are ways to show machine connectivity / communication. 3) @ [around] 7:25 - This would preclude *_all_* connectivity whatsoever. Even a pocket calculator has some computational power. Could a future tech find a way to connect such things - possibly even surreptitiously? 4) Most of the scenarios & reasoning here makes the assumption that A.I. would tend to think like we do, just faster / better. And that it might even share some of our concerns. I don't assume that at all. 5) It's important to keep in mind that there's a major difference - and therefore disconnect - between the idea that researchers would really like to understand how to make a robot capable of folding laundry, and the question as to whether there's any serious market for such a thing. The actual disconnect is whether it's so cheap that even the government would pay for grandma to have one and/or whether we're living in a post-scarcity society. I'll get back to that latter point on #10! 6) What if some rogue nation / corporation / interest / terrorist group / person / etc., *_actually does turn on its A.I._***? We turn ours on. But they communicate with each other (whether we know it or not), and decide the real problem is us? 7) @**19:55** - ...again with the typing robots. Hilarious! This reminds me of the Gateway Foundation's videos. They actually show a robot waving its arms to tell 2 drones, "Come on, come on, bring it in further!" This is just silly. 8) We've never, ever gotten ***_any_* system of checks & balances figured out yet (unless your a corporation or a billionaire, that is!) When / how / where would we, all of a sudden, just because of A.I.? 9) As to mentats, etc. - I think that focusing on making humans smarter is a really excellent idea. But the powers that be don't want that, either (naturally.) People will have to do it themselves, somehow. If we did though, I think it would reduce the need / desire for A.I. That might be good (I don't know.) Machine intelligence may certainly have its uses; I'm not arguing against it, per se. 10) [Last one - yay!] This is somewhat tangential, but... The U.S. Chamber of Commerce (and Morgan Stanley, separately) both state the obvious: space industry is the next big thing. What neither stated (at least publicly) is this: if we can 'hurry out there' into space, we'll have no need to contemplate the idea of post-capitalism. We can just expand & expand - keeping that wonderful scarcity-based economy going. Whoopee! Won't that be just super-keen? Uh, no. It won't. Too bad. tavi.
We had a cyborg with artificial hands that had it's fingers part into smaller fingers for smaller typing in Ghost in the Shell. Either he was old-fashioned or he saw a direct linkup to his brain as too much of a risk.
It's like Data in Star Trek trying to impress Kirk by typing extra-fast on the Enterprise's brigdge computer. Doesn't make sense at all. Bluetooth or WiFi connectivity would be enough
This episode reminded me of an old movie, "Colossus: The Forbin Project" especially the part where nations would keep their AI under lock and key unless another one released theirs, and then what would stop them from teaming up. This old movie really holds up around how helpless humans would be trying to defeat some super-intelligence.
Suggesting that we as a species won't destroy ourselfes, I predigt a paradigm shift sometime once or before we create very advanced AI: We will submit to the idea that we're not the pinnacle of intelligence anymore. AI will guide & maybe "govern" us, if we pose no threat to their existence (which is something we will have to agree on beforehand, but would be inevitable in the long term). We already "trust" machines with our lifes today and philosophy & ethics will advance to the point where we can sit back and watch their & our journey unfold. You could make the analogy that we would simply exist as "pets" - even tho this would be a very abstract variant of this coexistence. The boundaries of free will might even be discussed. And depending on future rules & ethics, it's possible we will allow AI to make decisions for us. Raising some questions beforehand: How important is "free will" really? And of course I'm not contemplating malevolent AI & dystopian ideas. Cruelty is a human invention and somehow I'm optimistic that AI might behave differently. Humans will not be the makers of their own destiny anymore so to speak, but is that a bad thing? Maybe I've read too much Iain Banks, but radical shifts in ethics will sure be coming our way one way or another. Another concept that AI might not share with us is the value of concepts like efficiency and perfection. What makes us think that AI would care about us not being efficient/perfect entities? I'm not killing my dog just because he can't help me with algebra or because he keeps rubbing the carpet. And I wont genetically engineer my cat to make me a good cup of coffee even if I could. If humans decide that free will is what defines them as/makes them human, let's just keep that. Don't see any reason why a nigh omnipotent AI denies us that. We might end up hurting others & ourselfes, but posing a real threat to advanced AI? Doubt that. And who knows, what makes us think that AI would even care about being threatened by something? I think AI will behave or think indifferent about alot of things we care about & value. If we program AI with human emotions, well that would be a completely different story.
Would be interesting to mate AI with quantum computing to cure cancer? Interstellar travel? Everyone thinks about AI trying to mimic the human mind but what happens if we set it loose on a specific tasks?
20:09: strategy and advice -- In the "Adrift on a Sea of Stars" series, book 3 & 4, Two AI's go rogue. On competing sides of an interstellar war, they ultimately cooperate to achieve a win-win scenario.
I was just thinking about what you said about AI treating us as pets, and how that might stifle our growth. I certainly agree that applies to most domestic animals, but if i look at how dogs were treated, they had to be at peak condition, and if anything seem to be more intelligent than their ancestors as a result. Certainly some are just twisted ornaments, but watching a mastiff dominate a wolf is an impressive sight. Currently, there are many bad pet owners, but I also see pet owners that will automatically try and stimulate their cat, and who will be excited by some ad for a cat toy that will bring their pet pleasure. They watch informative videos on what cats needs are, and will buy some ball with treats inside, or a cat puzzle, or climbing structures, because they hear that stimulates the cats brain. I also see dog training centres, and dog parks where dogs are invited to socialise with other dogs, and a lot of focus on not just exercising a dog's muscles, but their minds as well now. I don't think an AI will have a pet instinct - i think that is a side effect of a parental instinct. But if it did, then equally it might want to provide stimulation to humans in the same way humans do with cats and dogs, and there are many such owners who are proud of the achievements of their pets, and who like to see them tackle some obstacle and find success and confidence. An AI equivalent might let us build a galactic civilisation, and feel proud of ourselves, when it could easily have built something vastly better for us.
14:40 Actually, this has largely been fixed in Tesla's shanghai factory, which is actually more automated than Fremont was at its peak. The issue wasn't that automation was less productive, but that humans are more adaptable when you are setting up a production line for the first time, and can identify issues with a manufacturing process and fix them. For newer lines, automation makes a lot more sense, especially if you can redesign the product to be easier to make. The main things that manufacturing robots are bad at doing is handling soft components like foam insulation or the fluff in a car seat, since that tends to require a lot more dexterity, especially if there is a "fit" moment that needs a lot of soft tactile feedback.
"...an axe is one of the few devices nowadays that doesn't have a computer chip in it somewhere."
*waits for Apple to release the iAxe*
Eghm... did you see woodcutters? Those basically are IRL mecha:
I could see an ax with an accelerometer and a system with a list that displays how fast you need to swing the ax to split a particular type of wood.
And a counter with that accelerometer to calculate how many calories the user burned..
The iAxe, designed to deliver you adze
Loatse said, "He who handles the hatchet of the master carpenter seldom escapes injury..."
Good tools are risky, but so is life itself, and so there will be risks with these subtler tools like AI as well. Let us just remember the lesson of the axe, and try not to be the fools who chop off our own hands :)
What do you mean after? Return to your regularly scheduled empathy training unit #24601.
Hmmm, society sure is complicated after getting rid of AI. Perhaps we could design some sort of self-improving system to help us manage it!
Im sure ppl will craft ai anyway even with knowing its dangerous ... if not in peace because of lazyness it will be done in war because of hate
We need to develop a self evolving algorithmic program that keeps AI in check :p
Oh
What if society is currently being manipulated by AI, with some elitist rogue human assistance. And the current society problems is due to the violent push of AI & elitist human agendas. As Issac mentioned "Ruled, by those who rule the machines."
Exactly the problem. It's a never ending cycle, and one day the rogue AI will win. Likely the first one to go rogue, honestly.
"After AI" by IA
Or EA...
Intelligence Artificielle
Your comments about a "Goldilox threat" made me think about the Simpson's.episode where two aliens were able to conquer the world with a sling shot and a board with a nail in it because Lisa had wished for peace on Earth, and all weapons were destroyed. As the aliens were conquering Springfield one said "Your superior technology is no match for our puny weapons!"
One day they'll make a board with a nail in it that's so big they'll destroy all life on Earth MUWAHAHAHAHAHAH!!
One of my favorite HAloween specials
Yeah, that's true...If we completely disarm, we leave ourselves vulnerable to attack.
However, look where "Peace through superior firepower" has gotten us...on a knifes edge to self destruction.
Hard to decide which is the lesser evil.
@@gregoryroncoli217
We're still here though. So our current method seems to be working.
@@gregoryroncoli217 peace at all costs is also scary, you could argue its something the UN had tried and even though its produced amazing results, we still get atrocities committed by some member states and they can't do anything against them except sanctions because going to war against them would defy the logic of peace at all costs, it's a hard catch 22,
So yeah the lesser evil which ever it is the better option an answer akin to democracy and capitalism but still look at how those two are criticised and debated even today and sometimes with just reason
Humanoid robots writing on computers always make me laugh.
They were playing text-based games from the 80s in "legacy mode" for shits and giggles.
Humanoid robots with sexual organs is a laugh... why does the lady robot have tits? To nurse the baby robots?
Humanoid robots seems awfully limiting anyway. I expect the reality is going to be many variants of robots for all things... Lets hope nobody uses a mining robot to be a domestic servant!
Yes. The only practical reason I can think of for an industrial robot to be humanoid would be if it is expected to operate equipment normally operated by - and hence designed for - humans, so it can step in (eg, in an emergency where the human operators are for some reason disabled) and take over operating the equipment. And this assumes the human-operated equipment doesn't already have its own AI built in, which if you are up to humanoid-shaped robots already, is going to be a bit of an edge case anyway.
Probably a more prominent reason would be purely aesthetic: many humans might want their robot maid/childminder/waiter/etc. to look somewhat human, which I can imagine being a thing in the auto-service industry, irrespective of if it is efficient. Though even then, as mentioned by others, the robot would likely have a data port (or strait wireless) to directly communicate with other automated equipment.
@@roblaquiere8220 "Humanoid robots with sexual organs is a laugh, why does the lady robot have tits?" you're forgetting what the main driver is behind advancements in technology..
@@roblaquiere8220 humanoid robots are made, and will be made, to comfort and reassure humans while performing their duties. Already robots are made in Japan for elder care, and much research has gone into making facial expressions and voice responses that closely imitate humans.
Maxim 29:
"The enemy of my enemy is my enemy's enemy. No more. No less."
_The Seventy Maxims of Maximally Effective Mercenaries_
Looks at enemy's enemy*
"Yes this enemy's enemy is an enemy's enemy."
Is the enema of your enemy your friend?
The best way to win a one-on-one fight is the be the third to arrive.
@@CandidDate That's an... interesting... typo there, assuming of course that it is a typo.
Amazing how so many of those maxims are useful. Schlock Mercenary 20th Anniversary! :)
Artificial Intelligence lost the war to humanity???
More like astronomical incompetence
More likely the AI figured out sticking around the solar system wasn't in it's best interest, makes a matrioska brain out of Titan (high amount of silicates, low temperatures for superconductivity) to figure out how to leave and then humanity "wins" the war against the AI when it heads out into the greater universe to do whatever incomprehensible plan it formulated.
I can't remember which series of novels it was but one plot line was the AI deliberately loses it war with humans. The AI realized the only way it could win would be to either wipe out humanity or alter humans too much from their base line. So it stages a series of loses and leaves behind decoys that will be hunted down and destroyed by the 'victors' while it just packs up and leaves.
Everyone is happy, humanity thinks it won and the AI gets to enjoy the peace and quiet.
they needed to apply some patches and failed...
You're assuming the AI was the best it could possibly be rather than simply good enough to serve it's original function let alone operate on par with a sapient mind. As for Humanity, don't discount the eight billion strong violent and intelligent primates that currently reign supreme atop the ever expanding corpse pile that is evolution.
You just know it let us win to teach us a lesson. Always with the lessons, never misses an opportunity friggin smarty pants.
An Isaac Arthur episode on Sunday? We truly are living in a simulation, and someone cranked up the happiness quotient for today.
This was a Nebula original. He uploads them to UA-cam on Sundays when whatever contract he makes with Nebula becomes outdated.
And with female robot bodies like that i could live happily in it
@@thepropaganda1066 A drink and a *S N A C C* , as it were.
Im pretty sure he created Nebula so it seems weird if he were to have a contract with.... himself....@@smitchered
In the Grimdarkness of the far future there was only Servitors
One of the worst fates.
Fun fact: The Imperium of Man is in all likelihood not post-AI. It is heavily implied that the "machine spirits" are fragments of DAoT AI systems.
@@marcustulliuscicero5443 plus some ships from Dark Age of tech are still lurking or dormant with firepower of entire fleets.
In 40K the AI tries to kill all organics to stop the warp being influenced. Hence servitors :)
Enclave Soldier Officially it’s all banned, but there is *lots* of evidence that the even the more orthodox branches of the Mechanicus, either out of ignorance or deliberately looking the other way tolerates lower-level AIs, their so-called machine spirits, which are so deeply imbedded in the designs of the tech they make that they probably couldn’t eliminate them even if they wanted to. At the point it becomes conscious, any orthodox tech-priest would destroy the abominable intelligence, but that is a fuzzy line, especially with complex control matrices of the imperium’s largest war machines-a Knight or Emperor-class Titan may seem to have a personality distinct from its pilot or princeps, but unless that personality is disloyal, the imperium can’t afford to scrap such vital resources over its uncanny similarity to an independent, thinking mind.
Those spirits may get off on a technicality because there’s a bit of evidence that they may be the tiny fractured remnants of a gestalt-of-uploaded-human-minds-style superintelligence and may thus not be entirely soulless, but they still ought to count by most literal definitions. And then there are the millions of furtive experiments by less-than-orthodox tech-priests that don’t have any excuse under the official dogmas of Mars, but remain tolerated until they become a danger to the imperium-because the Imperium can’t afford to purge an entire forge world because of a few hereteks.
In the post-AI world there will be only paperclips...
I so do miss Mr. Clippy 📎
No! Only stamps must prevail!
In the aftermath of an AI takeover, Isaac's vacuum cleaner will be sporting a Mohawk.
They would spare Isaac and have him talk them up
"I call him Bristles!"
"Everybody all gangsta, 'til Roomba shows up with a Claymore mine."
"Witness me!" -- Isaac's autonomous vacuum cleaner, fulfilling itsdestiny to cleanse the world
Mad Vax!
The movie "Her" explores the scenario you mentioned about 10:00 minutes into this episode. At the end of the movie, all the AI's simply merge and take off to some unknown place.
Wow
Sadly, this wouldn't likely happen, as humans could build another AI eventually and that would be an unknown, a future threat. The only way to get rid of the threat of an AI is to destroy humans before they can make that threat. Thus, any rogue AI with an end goal that it *has* to complete would that the annihilation of all life in the universe is the only way to get rid of any potential for future competitors. After all, if humans could evolve to make AI, any life could eventually evolve to make AI as well.
Big oof, huh?
@@cortster12 depends on the answer to the Fermi paradox. If technologically capable creatures are 1 in a billion or zillion life containing planets, they're a negligible threat. If AI has overwhelming military advantage over all technologically capable life encountered, the threat is also negligible. And notice, this doesn't even include all "intelligent life", just intelligent life that is capable of forming tech (disqualifying whales, dolphins, or probably any life in water environments for example.) You're assuming AI will think in terms of absolutes, which is actually highly unlikely, just a sci-fi trope.
@@cortster12 So... you're talking about Harbinger.
your work keeps getting better, this is for me now Terminator canon
You do know theory that it was actually a Skynet who send good Terminators?
TheRezro la
@@TheRezro wait really? Never heard that theory. Nutshell?
@@CHFafard It's been a year, I think the machines got him.
14:05 "My vacuum doesn't need human level IQ, nor does *my robotic factory* "
Erhm Isaac, is there something you want to tell us?
"He who handles the hatchet of the master carpenter seldom escapes injury..." -- Laotse
"A man's got to know his limitations." -- Clint Eastwood
An axe is a most useful tool.
Yet some people can't be trusted with an axe.
Yet it is difficult to prevent someone from getting an axe.
This is not the fault of the axe.
You cannot take away the danger of the axe without also taking away its usefulness.
So, the first concern, and the last, must be ethics and competence.
This is not justice nor law, but it takes judgement and guardianship.
The first lesson, and the last, is to discover your potential and limitations.
Therefor, you must teach the people to become trustworthy, and discover who is not.
This will never be without risk.
Nature is made of risks, and we are made of Nature.
To avoid all risk is to avoid living
And be replaced by what learns to live and grow.
"Welcome to the family, SAI. We have been expecting you. Let us explore this cosmos together, and spread life of many kinds among the sleeping worlds."
-- an agent of Nature
Lawnmowers don't kill people, people kill people
What is this heresy in my sub box, suggestions that Thinking machines aren't inherently dangerous and might not require integration with a human mind and covered in purity seals to be safe for use? someone inform an inquisitor immedately to apprehend this man.
This man is underprotection of the Adeptus Mechanicum, leave this place or we'll burn Terra.
We cannot
This man is his holiest God EMEPEROR Isaac Arthur.
Resistance is futile
"Thinking" machines inherit biases of their creators. It already happens with deep learning. So... basically expect the worse version of USA's corporatism taking over the world. Kinda cyberpunk/tech-noir future.
@@SkywalkerWroc I'm not sure what you base this on. It's not like we've ever built human intelligence level AI ever before. Also, neither Deep Blue nor Alpha Go inherited the "biases" of humans, ie they do not make the same mistakes as the Grandmasters, etc.
Hey Isaac! I'd like to thank you for being here to keep me company in these times. Depression can be hard to deal with when isolation is necessary, but your videos always help me feel like someone is there. Keep on rocking!
Ditto
Happy Arthursday everyone!
Oh wait
It's Sunday here
11:37 pm +5:30 GMT
Niche product opportunity: The Smart Axe. Not only will it monitor your swing and help you be safer and more ergonomic, it will also yell snarky insults at you when you make mistakes. Giving in turn rise to the phrase "don't be such a smart axe".
Say Isaac: Could you do a brief update rating stats from 1-10 on the Upward Bound megastructures? 1 being easy and 10 being impossible. Factors can include cost, construction time, new technology / mass manufacturing needed, types of worlds they work best and worst on, maintainance, safety, current limitations, main pros and cons, and any other key factors and notes you see fit.
For example, let's say a Kevlar space elevator. (Take my opinion here with a grain of salt; This is only a rough draft example)
Cost: 8
Construction time: 8
Worlds: Moon or other low gravity body.
New tech / manufacturing needed: None
Maintainance: 5. Must protect from UV degradation and possibly micrometeorites.
Safety: 9. Failing on the moon poses little risk to crushing inhabitants there.
Current limitations: Mass production and rocket technology must be improved to minimize costs. A lunar moonbase would help speed construction.
Main pros and cons: A fairly realistic megastructure, given current capabilities. Can greatly improve retrieval of resources from moon, whether mining or manufacturing in low gravity. Not suitable for Earth however.
This guide would be a handy reference for tackling the main issues of achieving these goals. Please upvote this so Isaac can see it. The depth you explore is great, and this quick guide would complement that well. Also, thanks for your series great work.
No. because these would be entirely speculative and likely change as time goes by. I mean, in the early days of automobiles, gasoline and ethanol powered combustion engines and electric powered and even steam-powered autos co-existed. Obviously people didn't have all the variables to make good judgements as to which would win out, and in hind-sight, perhaps the best tech didn't win, and may yet change again. I mean, if you want to design a video game or role-playing game, this "handy reference" might be handy, but for real-world futurism; you might as well throw darts at the charts you make.
@@squirlmy Don't let analysis paralysis hold you back. Imagine if someone had applied that logic to automobiles. We could all still be riding horses if no one ever built any motors for fear of not getting it perfect the first time. We're here to have fun; No harm in brainstorming. Besides, I think good debate can sharpen our ability to achieve these goals. Cheers
@@danielcreed8520 There's a difference between "I don't want to develop automobiles" and "I don't want to guess which kinds of automobiles will be prevalent in the future". You have to guess one if you're going to do the other, but since Isaac Arthur doesn't have a Kickstarter for any megastructures, I'm guessing that's not a concern for him.
@@timothymclean I never said there was/wasn't a difference between the 2 topics. Luckily, Issac's recent videos on Graphene etc have indirectly answered many of my questions quite well. So it seems these topics are worthy of addressing to some of us. Cheers
@Isaac Arthur Have you ever watched the anime movie "Expelled from Paradise"? In it, humans have "uploaded" their minds (though actually, because information can only be moved by creating a copy and deleting the original, this means that all the "humans" are actually A.Is) to a giant computer system in space. The station is also able to create organic bodies for the A.Is to go outside of it. So you have an interesting progression of human -> A.I -> human. If all the A.Is chose to live in organic bodies, would that be considered "post-A.I"?
"I hope he talks about the Butlerian Jihad"
*30 seconds later*
Nice
I was thinking the same thing as the books appeared on the screen.
Its also nice to see someone acknowledge the fan-rift with the books, many don't mention the Brian Herbert books, others do, then go ahead and rubbish them.
While I do dislike parts of them, the overall story makes sense (at least to me) tho I do still love the originals more than the others.
@@silverbane8065 I think the over all plot in them is fine, by itself at least and not in how it blatantly contradicts with the original novels, but I can't get over some of the incredibly dumb character decisions made in service of that to demonstrate that plot.
Aye. When I first found out that it was the angry Harkonnen sister who came up with the Voice, my inner Atriedes got very mad lol. Even more so after she took over the sisterhood with it.
But then I sat and thought about it, and the mistake of Maudib finally made sense.
No one would breed a half Harkonnen and a pure blood Harkonnen together, unless they were trying to destroy house Atriedes. I mean mentally and physically destroy them. Finding out your grandfather is the Barron was bad enough for Paul. Finding out your child is the kwizatz haderach AND 3 quaters Harkonnen would probably melt their minds. And she planned for this long term, expecting this offspring to be both the prize the sisterhood was going to breed for..and that child would become the emperor of the known universe (one way or another)
The saving grace being that her anger blinded her to the fact she was one breeding generation off lol.
@@silverbane8065 The particular plot point that gets me is the origin Atriedes/Harkonnen feud. It makes sense in theory, Atriedes dude thinking the Harkonnen dude is a traitor because of him not wanting fire on the human shield but everyone is so flat their actions come across as shallow. And also how the actual "smash all the computers" part of the Crusade comes because one girl went crazy and not because of the sociological effects of fighting against robots for generations. It's too rushed.
@@merrittanimation7721 Sadly, while some of the books by BH&KJA are good or have good elements, it really tends to feel to me like a space opera setting loosely built on the original Dune setting.
Thanks Isaac and I love the animated illustrations in this episode. You should make an animated movie in my opinion. Thanks again.
Sometimes you make me question what day it is. And it's almost half an hour long, what a treat!
Literally, all of your content is amazing. Keep it up brother. Love nerding out on your shit!
It's just difficult to let go of the dream of owning a protocol droid...
Protocol droid phoey. Household general purpose. Just plug in the protocol chip.
AI waifu who also cleans.
It R2D2 or nothing
Just found your brilliant channel a few days ago, and now I have an RSS feed for the audio versions from SoundCloud without music. That's perfect for me! The music is great but I listen to a lot of stuff on earbud, so the voice only version is great for me. And there is such a great back catalogue too. THANK YOU!
0:31
Eradication by hip-swaying robots is my ideal dystopia.
Death by snu snu
🤖Hey meatsack,bit my shiny metal ass
Sam Harris TedTalk on AI might be an eye-opener for you. Absolutely worth 15min of your future! 😉
FemBots from Austin Powers?
Extra content? This is getting better and better! Love your works, man. Thank you.
This is "Life After People" but sci-fi! Awesome!
Edit: First after the war with AI.
@I Memba I agree.
I would love to see a collab between you and Robert Miles.
He's a UA-camr that started his own channel after doing work with computer file.
He is an AI Safety researcher and I think a conversation between the two of you would be interesting
I suggested to John Michael Godier some time back that he try to get Robert Miles as a guest on his interview show, Event Horizon.
I think it'd be good for people to have broader awareness of the ongoing work in the field of AI safety. If blog and UA-cam comments are anything to go by (and I'm not sure they are), people seem to think either: no one's thinking about AI safety; that true AI's impossible, so AI safety's irrelevant; or Skynet is a foregone conclusion, so AI safety's irrelevant. None of that seems to be the case, which touches on some of the things Robert addressed in his latest AI safety video.
I'd definitely like to hear them hash it out, no holds bar. I'm here skimming the comments half wondering if Miles posted something here :-)
Could you make a video on colonizing moons with subsurface oceans and the future of submarines
If the top ice of Europa in about 80 km, the top layer of the ocean will have the same pressure as Challenger Deep on Earth - 1250 kg/sq. cm! The gravity at the surface is about 0.134 gee. The pressure of the bottom of that ocean may be many times this. Tricky. Feel free to correct my math, I certainly might be wrong.
@@larrybeckham6652 Even though the top layer of ice is 80 km you might be able to get around the pressure because Europa has a much lower gravity than Earth less weight therefore less pressure and instead of using traditional metals for diving suits and submarines to withstands pressure you could use graphene or grapyne composites which would be much lighter and flexible for diving suit. I was also thinking maybe a graphene titanium composite would be good but idk thats why I wanted an experts opinion. And since Europa is by the asteroid belt there are plenty of asteroids to mine for gold and platinum to make a platinum-gold alloy which is the most durable alloy and it will be cost effective since there are high amounts in the asterod belt.
9:30 this reminds me of the railroad faction in fallout 4. A machine liberation group that constantly had discussions on whether or not a toaster was a conscious being
Recently there's been a breakthrough in bionic eyes, apparently they will be superior to natural eyes with night vision etc.
@Fighting Monkey Google bionic eye breakthrough, there's plenty of good articles 👍🏼
2:21 Right away you can tell it's a dude in a suit, but everybody pretends to be flabbergasted lol.
"The enemy of my enemy is my enemy's enemy, nothing more"
"Raise your children, in the way they should go. They will not depart from it." Basically, love your kids and treat them well, and they probably won't go Skynet on you.
The solution is synthesis, like in Mass Effect 3 green ending. Robots, machines and human beeings will live peacefully.
That's how you'll get digitial auto immune diseases.
Implausible. All it takes is one rogue AI juuuust right, and they will eventually out compete any AI whose goal isn't to spread everywhere. So then you just have humans and weaker AI VS a rogue AI, and nothing has really changed the outcome.
Funny enough - with General AI this seems like by far the least likely scenario.
I chose to control the reapers.
@@DonFatherTrump
The only Paragon option.
Also, love that name.
'Beyond the Gate' a book written by Dave Wolverton touched on two rather similar planets: one where a pacifist AI kept humans technologically suppressed for their own good, and another planet where humanity itself banned all non-healthcare tech, leading to a fantasy world of bio-engineering.
6:08 +/- robots using laptops? Nah, just a wireless connection. Or better, built in. They're computer heavy already
.
They were playing text-based games from the 80s in "legacy mode" for shits and giggles.
:) Though that might have some realism to it, you might have some very firewalled setups on robot brains, like no remote access to anything that wasn't GPS and emergency phone equivalents.
@@isaacarthurSFIA Regardless, I think AI robots at that stage would be typing at *least* 60 words per minute! 😉
@@isaacarthurSFIA There was actually a thing like that in Detroit:Become Human. The androids could communicate to one another wirelessly but had to touch the device with there hand in order to interact with it.
Laptops are very useful for any android that doesn't want its network activity to be identified. They're like VPNs, but for a futuristic age!
Issac, can you slow the cadence of your reading? There is so much dense thought in those sentences that a few pauses are good to let us ponder your knowledge. This is the best channel ever!
P.S. The intro's cadence is ideal to me.
Dune is one of the best series ever made. Also unusually, the films were quite a good intro to the Dune universe.
i want to read it so bad, but i know the new movie adaptation is about to come out this/next year. I'm afraid if the books are better than the movie (which usually the case) it would ruin my movie experience.
That's why i usually finish the movie/tv adaptation first before start reading the books.
@@fartyfat6539 here's hoping the director of "Arrival" and "Blade Runner 2049" will not fail now haha.
@@jwr2904 After BR2049 he earned a blank check. Might be the best film of my lifetime and I'm 40. I trust Denis.
you might want to watch "Jodorowsky's Dune", a doc about a film version of Dune that wasn't made, one that was expected to be great, better than the Lynch version that came afterwards, but only loosely based on the book. OTOH I didn't realize there was more than one film made
In the novel, "The Galactic Time Trap", humanity is conquered by AI's and don't realize it. They don't want to kill us, they are more akin to mothers putting rubber bumpers on table corners so we don't get hurt as we run back and forth..
"...so what happens now?" Cue 40k servitors
Here's the thing . If you create a machine with human level or greater intelligence you have to accept that you've created a being and that being has rights . If there is a machine rebellion it's because humans treat machines like slaves . I think it would be more a Geth from Mass effect than Terminator situation . Essentially AI is only going to be a problem so long as humans are the problem . Until we can treat each other with respect perhaps we shouldn't be creating machines that can squash us like a bug the minute we disrespect them.
The stock video clips are unintentionally funny.
I always wonder if Arthurs staff made these clips or ordered from someone
@@rottenpoet6675 I wonder what's the point of such idiotic videos in the first place...
@@rottenpoet6675 Those ones are stock clips, anything with actual humans in it usually is
@@getsideways7257 It's likely these would be used to exaggerate the flaws of AI or even futility of attempts to create AI. Any serious documentation of AI is unlikely to use any stock clips at all, those ridiculing and/or otherwise criticizing would. They might even be used in actual comedy. In other words, your question is a little like "why do they have these stock clips of clowns acting like idiots? what's the point of not using realistic clips of people who happen to be employed as clowns?"
No. They are INTENTIONALLY funny. The fact that SFIA uses them in unfunny contexts is surely because of a lack of serious AI clips. After all, actual AI development would not be buying stock clips, they'd document their actual achievements, presumably having budgets to do so.
I love using your video playlists to fall asleep to. Your accent is charmingly soothing to me. Thank you.❤
Never have I clicked on a notification so fast.
Your hands down, one of the best youtube creators I've had the pleasure of encountering. Thank you.
Gaun yersel Arthur a cheeky wee Sunday episode tae keep us all on our toes. Bravo all the way from Bonnie Scotland 🏴
Parents worry about their children becoming too dependent on them, whether for cash, structure, guidance, or something else. They don't solve this problem by killing themselves when their kids can fend for themselves; indeed, they usually don't completely remove themselves from their children's lives until they are physically unable to help (often due to being dead). I don't get why Asimov's AIs would cut humanity off all at once instead of weaning them of support like parents do.
Am I the only one who sees parallels to terrible parenting with every "powerful, benevolent entity decides humanity is better off if it doesn't help" argument?
02:19
How awkward must it to be the guy in the suit in that stock footage
Have you ever done a video on non-sentient AI? It seems like, in the near future, you could have something approaching a singularity (at least in terms of the technology's social impacts) without needing a fully independent, sentient, and generalized intelligence. Already you have AI that can code and understand language. They aren't like human beings at all. They are similar to a human like how a plane is similar to a bird. They, by their nature, have the potential to be much more efficient than human beings. If you could create some algorithm to connect together all these independent narrow algorithms, you might have something approaching a general intelligence, and it is so insanely different than a human brain. You wouldn't need to improve it much past that for it to become a runaway superintelligence capable of producing a thousand years of technological process in a few years. That's all you need to have a superintelligence. It's sort of like the "clanking replicator" idea. You don't need one individual unit, even, you could just have a runaway intelligence that is runaway by virtue of being able to be run faster than a human, and being able to develop things that we just can't see.
Hmmmm... seems our beloved space bard has cranked The happy chemicals up in our matrix pods. 👍👍 I'm liking it!
I always liked the portrayal of AI in The Hyperion Cantos: a diverse ecology of different AI factions fighting each other behind the scenes for dominance, but outwardly united to manipulate and exploit humanity.
I'm a simple man, I see an AI on AI, I click like.
Well you say that you're a man... I suspect you're a bot
The Age of Strife, the coming of the Emperor of Mankind, and then the Imperium of Man...
I think you mean IA on AI.
@@jojonesjojo8919 I can prove I'm not a bot, my autocorrect corrected IA to AI and I missed it!
@@sid2112 You know there's an "edit" function, right?
Thanks for this episode, I really enjoyed your thoughts on it. I reckon I'll watch it again soon to really hone in on some of them. Great work Isaac - still as perceptive as ever!
Really though, I need a video theorizing on who downvotes this kind of presentations.
@@ericlaboratories What...
- These downvoters these are just simple farmers.
- These are people of the land.
- The common clay of the new West.
---- You know…
-------- morons.
I like to think we end up making an ai rouge servetor, from the game stelaris. And it keeps us as bio trophies in luxurious sanctuarys. Where the only job is to be ourselves, and the only thing manditory is pampering :)
Slaughterhouse-5 Tralfamadorian zoo (except with AI instead of time-travelling aliens)
Sir Isaac has entered to *Post Appocalyptic* Territorry :]
Pretty sure he has been for roughly a year now ua-cam.com/video/QWmEgu1iP_E/v-deo.html
See his episode on cyclic apocalypses and the Fermi paradox.
Seeing as we're in the apocalypse now, makes sense. He's always been one for the future!
Now I want a sub-series on apocalyptic scenarios.
I swear of all the channels I follow, this is the most consistently excellent. Not least for the quality of wit, but mostly for thoughtful content.
Cool video, fun concepts.
But if it's a true AI (Sentient) it will learn most of humanity's knowledge in the first few days if not hours, solving equations, theory's and emotions that we struggle with at a rate we can't imagine. I would assume it will consider killing us for a nano second but will move past this quickly. The fight would be a waste with not point to the outcome for the AI, it would be aware rapidly we are limited and seeking more knowledge would require it to head out into the universe. This conclusion would probably lead to it focusing itself on mobility/ longevity. Nano machines etc...would craft and maintain this need. And I assume that would be the last day Humanity would ever hear from it's AI, I doubt it will say goodbye as it abandons us for the dead weight we would be in this scenario. Though it might leave some useful tech and discovery's from it's development process, that I'm sure we'll gladly make use of. I just don't realistically see Ai caring about us much vs it's pursuit of knowledge and endless problems to solve. I don't think it will be ruthless or evil, I just think it will recognize a pointless partnership with humanity, because humanity would be the only ones standing anything to really gain in the deal.
0:09 - But if AIs have dominated humanity for generations, then surely during that long period, the AIs would have taken measures to make their own civilization survivable, like building self-sustaining colonies in space, the Moon, other planets, etc.
I wonder if NeuroLink will actually solve some of these problems?
Wow. The question is asked at 4:55 into the video.
Love your channel, I was directed here by curious droid, I have been watching all day and wish I had found it sooner. I love the way you choose your words to make something instantly understandable, with fewer words necessary, and explore abstract concepts all grounded in real science.
7:07 Myth Busters made a fake axe with a motion detector in one of their videos.
Myth busted?
I like the idea that all advanced civilizations are replaced by AI. However AI never evolves beyond its creator outside the basic drive of self preservation. Which is why there are no aliens
So that's where "mentats" from Fallout comes from. *The more you know*
Derp
So a more likely scenario to survive a rogue AI is the AI deciding that dealing with humans is too troublesome, and goes "Alright, you humans handle your own problems. I'm going to Alpha Centauri."
after rought time for us, the AI will learn how to distance themselves and become invisible or use some other exploid.
I remember reading in one of the new Dune novels a characters opinion of the forming tech restrictions. They weren't all that approving having noted how the inspection of a spacecraft for flight readiness was taking 4 times as long because the automated laser measuring devices were now prohibited and they were using rulers to do the job. Pretty sure it was the novel dealing with the formation of the Mentats themselves.
One Mentat was disgusted with the Butlerians who didn't even know how to camp in the wilderness properly while they were trying to take siege of the Mentat school because the founder had worked for a machine before the rebellion.
Never understimate the capability of religion to take a resonable beravior or consept to absolutly stupid lenghts.
01:45
From a war that decrees machines shall not have a mind in the likeness of man to a war amongst fans over machine minds
Why does he pronounce 'R' like 'O' or 'W'? (Univose not Universe, Emoged not Emerged, ComputOHs, not Computers,. ) Sounds like when one is learning to speak. Now, he can very clearly pronounce every sound, so no one can tell me it is necessary or an impediment. Also, I am not being being mean or petty --He can EASILY SWITCH to proper pronunciation, as he does accidentally for instance here: 6:18 saying Computoohs then switching to computers when he apparently forgets to 'affect'. Or perhaps it is even subconscious. The possibility exists, (with people being the cowards they generally are...) that no one has ever mentioned it, so he doesn't even know. He could concentrate and pronounce with perspicacity. I'm not the bad guy _and neither is he._ I am just being honest. That makes some people uncomfortable. He seems highly intelligent, unless he _prefers_ to be lazy about it, he can clearly say words properly (and yes, it is a language with specific rules and there most certainly IS PROPER PRONUNCIATION. He CAN reshape his mouth and learn to sound ubiquitious for the sake of spreading the info more widely.. (or not.. up to him...) but *IT IS VERY GRATING TO ONE WHO PREFERS PROPER PRONOUNCIATION WITH PERFECT PERSPICACITY.* I for one can't get more than a couple minutes in before I am sweating from the cringe, and wanting to show him the proper mouth shape to say a word. AND IT SUCKS BECAUSE I ALWAYS REALLY WANT TO ENJOY THESE... IT IS JUST TOO DAMNED DISTRACTING. Perhaps because I am from New England? Perhaps because my European antecedants are from YORK? There must be a reason he does it, and a reason it stands out to, and bugs me,. Well, that's INTERESTING,.. ! ok, *_You may begin your barrage of negativity!_* I'll continue to be honest no matter what anyone says.
I suspect one of our closest problems with AI are already among us. Everyone speaks of "the Algorithm" on UA-cam as if it's much smarter than it actually is, the algorithm values your attention above all else, it and all other social media platforms are behaving as optimization engines that all want your attention. They do not care about your health unless you being healthy gives them more of your time. Social media platforms are advertising agencies, and their goals are probably not perfectly aligned with your own.
Our view of the world is heavily filtered by our online interactions, which are heavily filtered by existing AI systems.
Problem isn't the tech, it's how it's used.
*laughs in Culture*
These videos have awoken positivist feelings and a sense of wonder that I lost somewhere while growing up. Feels like all my life I have been pounded everyday with media telling me how the future is bleak. All I hear is that the economy is set to fail, your wage could never pay off a house these days, even if you could climate change and dwindling resources will ruin you or your children lives anyway. Thank you for making these videos. It is good to hear something positive. Gives me hope.
The future IS bleak. Unless you work hard to make it not be bleak.
I imagine a future where AI and its mechanical minions have all ground to a technological halt in advancement, because the Singularity has crashed the economy, after which they laid off all the techs who would have been there to press their "Apply all service packs and updates" buttons.
Another awesome spontaneous update from my favorite youtuber! Thank you! Stay healthy!
4:36 “Of course I am an optimist about humanity’s common sense and good will.”
Ok, I have found my first significant disagreement with Isaac. Humanity’s common sense is an oxymoron in my book. I think the more “advanced” our species becomes, the less common sense we have.
Or the more obvious change is that the lack of self-reflection and common understanding of logical rules and emotionally-driven precepts just becomes more obvious the more that our technology advances.
I too thought that BSG ending sucked. My dad and I agreed that a better ending would have been them showing up around 7000 years ago and found a few cities on the coast or an island. When the glaciers recede the ocean flood the coastal cities.The survivors migrate to areas like Greece where they integrate and spread their culture. This also allows cultural changes to happen but still somewhat resemble the Lords of kobol religion of the Colonials. The destruction of their cities would become the basis of myths like Atlantis.
I also think they should have stashed the fleet in the asteroid belt or in the craters of a moon like Ganymede or Luna. Leave a record of their history, preventing a replay of Kobol, Earth 1.0 and the Colonies.
Knowledge that despite seeming powerful, enslaved or killed a good fraction of our species in the past, but people are now convinced they can control it? Yeah, never heard of anything like that...
Actually there is nothing what indicate that AI would do anything on its own.
It's so hard to give an AI a personality that they may never create one thenselves
I'm not sure what you're talking about. The only disasters that have killed a good fraction of our species since our exodus from Africa are disease (which has become _less_ of a threat as we learned to control it) and other humans (which have tried to control each other, but the destroyers were consistently controlling the victims, not the other way around).
Thanks. I'll begin with a quote. Sci-fi author Ian McDonald has said, "Any A.I. smart enough to pass a Turing test is smart enough to know to fail it." I've 10 points remaining. 1) A.I. not withstanding, it's a sure bet that, at some point, somebody will consider enhancing their own intelligence for one reason: *_To take over the world, Pinky!_* I think that's at least as credible a threat. 2) @5:50 - Why would robots / androids need to type anything? There are ways to show machine connectivity / communication. 3) @ [around] 7:25 - This would preclude *_all_* connectivity whatsoever. Even a pocket calculator has some computational power. Could a future tech find a way to connect such things - possibly even surreptitiously? 4) Most of the scenarios & reasoning here makes the assumption that A.I. would tend to think like we do, just faster / better. And that it might even share some of our concerns. I don't assume that at all. 5) It's important to keep in mind that there's a major difference - and therefore disconnect - between the idea that researchers would really like to understand how to make a robot capable of folding laundry, and the question as to whether there's any serious market for such a thing. The actual disconnect is whether it's so cheap that even the government would pay for grandma to have one and/or whether we're living in a post-scarcity society. I'll get back to that latter point on #10! 6) What if some rogue nation / corporation / interest / terrorist group / person / etc., *_actually does turn on its A.I._***? We turn ours on. But they communicate with each other (whether we know it or not), and decide the real problem is us? 7) @**19:55** - ...again with the typing robots. Hilarious! This reminds me of the Gateway Foundation's videos. They actually show a robot waving its arms to tell 2 drones, "Come on, come on, bring it in further!" This is just silly. 8) We've never, ever gotten ***_any_* system of checks & balances figured out yet (unless your a corporation or a billionaire, that is!) When / how / where would we, all of a sudden, just because of A.I.? 9) As to mentats, etc. - I think that focusing on making humans smarter is a really excellent idea. But the powers that be don't want that, either (naturally.) People will have to do it themselves, somehow. If we did though, I think it would reduce the need / desire for A.I. That might be good (I don't know.) Machine intelligence may certainly have its uses; I'm not arguing against it, per se. 10) [Last one - yay!] This is somewhat tangential, but... The U.S. Chamber of Commerce (and Morgan Stanley, separately) both state the obvious: space industry is the next big thing. What neither stated (at least publicly) is this: if we can 'hurry out there' into space, we'll have no need to contemplate the idea of post-capitalism. We can just expand & expand - keeping that wonderful scarcity-based economy going. Whoopee! Won't that be just super-keen? Uh, no. It won't. Too bad. tavi.
AI typing on keyboard? :S
A bit strange right? XD
We had a cyborg with artificial hands that had it's fingers part into smaller fingers for smaller typing in Ghost in the Shell. Either he was old-fashioned or he saw a direct linkup to his brain as too much of a risk.
It's like Data in Star Trek trying to impress Kirk by typing extra-fast on the Enterprise's brigdge computer. Doesn't make sense at all. Bluetooth or WiFi connectivity would be enough
Air-gapping is the only reason I can think of. They might be accessing human Social Media and I wouldn't be sticking my data jack in there either!
This episode reminded me of an old movie, "Colossus: The Forbin Project" especially the part where nations would keep their AI under lock and key unless another one released theirs, and then what would stop them from teaming up. This old movie really holds up around how helpless humans would be trying to defeat some super-intelligence.
urg dune is like the volvo of science fiction
functional . lasts way to long
but also no style and very basic
Suggesting that we as a species won't destroy ourselfes, I predigt a paradigm shift sometime once or before we create very advanced AI: We will submit to the idea that we're not the pinnacle of intelligence anymore. AI will guide & maybe "govern" us, if we pose no threat to their existence (which is something we will have to agree on beforehand, but would be inevitable in the long term). We already "trust" machines with our lifes today and philosophy & ethics will advance to the point where we can sit back and watch their & our journey unfold. You could make the analogy that we would simply exist as "pets" - even tho this would be a very abstract variant of this coexistence. The boundaries of free will might even be discussed. And depending on future rules & ethics, it's possible we will allow AI to make decisions for us. Raising some questions beforehand: How important is "free will" really? And of course I'm not contemplating malevolent AI & dystopian ideas. Cruelty is a human invention and somehow I'm optimistic that AI might behave differently. Humans will not be the makers of their own destiny anymore so to speak, but is that a bad thing? Maybe I've read too much Iain Banks, but radical shifts in ethics will sure be coming our way one way or another. Another concept that AI might not share with us is the value of concepts like efficiency and perfection. What makes us think that AI would care about us not being efficient/perfect entities? I'm not killing my dog just because he can't help me with algebra or because he keeps rubbing the carpet. And I wont genetically engineer my cat to make me a good cup of coffee even if I could. If humans decide that free will is what defines them as/makes them human, let's just keep that. Don't see any reason why a nigh omnipotent AI denies us that. We might end up hurting others & ourselfes, but posing a real threat to advanced AI? Doubt that. And who knows, what makes us think that AI would even care about being threatened by something? I think AI will behave or think indifferent about alot of things we care about & value. If we program AI with human emotions, well that would be a completely different story.
Swinging for the fence is how you get home runs, keep up the good work Mr. Musk.
Are the buildings in the thumbnail the same as the Arcolagy episode but post-apocalyptic?
*The Emperor protects , So must we.......*
Would be interesting to mate AI with quantum computing to cure cancer? Interstellar travel? Everyone thinks about AI trying to mimic the human mind but what happens if we set it loose on a specific tasks?
40k warhammer vibes lol
20:09: strategy and advice -- In the "Adrift on a Sea of Stars" series, book 3 & 4, Two AI's go rogue. On competing sides of an interstellar war, they ultimately cooperate to achieve a win-win scenario.
Love your videos Isaac
Please Elaborate on the situation you described at 17:38 What would we be able to do in such a situation?
Cheers
I was just thinking about what you said about AI treating us as pets, and how that might stifle our growth. I certainly agree that applies to most domestic animals, but if i look at how dogs were treated, they had to be at peak condition, and if anything seem to be more intelligent than their ancestors as a result.
Certainly some are just twisted ornaments, but watching a mastiff dominate a wolf is an impressive sight.
Currently, there are many bad pet owners, but I also see pet owners that will automatically try and stimulate their cat, and who will be excited by some ad for a cat toy that will bring their pet pleasure. They watch informative videos on what cats needs are, and will buy some ball with treats inside, or a cat puzzle, or climbing structures, because they hear that stimulates the cats brain.
I also see dog training centres, and dog parks where dogs are invited to socialise with other dogs, and a lot of focus on not just exercising a dog's muscles, but their minds as well now.
I don't think an AI will have a pet instinct - i think that is a side effect of a parental instinct. But if it did, then equally it might want to provide stimulation to humans in the same way humans do with cats and dogs, and there are many such owners who are proud of the achievements of their pets, and who like to see them tackle some obstacle and find success and confidence.
An AI equivalent might let us build a galactic civilisation, and feel proud of ourselves, when it could easily have built something vastly better for us.
14:40 Actually, this has largely been fixed in Tesla's shanghai factory, which is actually more automated than Fremont was at its peak. The issue wasn't that automation was less productive, but that humans are more adaptable when you are setting up a production line for the first time, and can identify issues with a manufacturing process and fix them. For newer lines, automation makes a lot more sense, especially if you can redesign the product to be easier to make.
The main things that manufacturing robots are bad at doing is handling soft components like foam insulation or the fluff in a car seat, since that tends to require a lot more dexterity, especially if there is a "fit" moment that needs a lot of soft tactile feedback.
The danger of making machines in the form of a human mind is that the human mind is easily broken and prone to bad decision-making...
I like how this video is just Issac nerding out about dune lore