Machine Rebellion
ฝัง
- เผยแพร่เมื่อ 25 ธ.ค. 2024
- The possibility of Artificial Intelligence turning on humanity has been a concern for as long as we've had computers. Today we will look at some of those fears and see which ones might be valid and which might not be cause for alarm.
Use my link www.audible.com... and get a free audio book with a 30 day trial!
Join this channel to get access to perks:
/ @isaacarthursfia
Visit our Website: www.isaacarthur...
Join Nebula: go.nebula.tv/i...
Support us on Patreon: / isaacarthur
Support us on Subscribestar: www.subscribes...
Facebook Group: / 1583992725237264
Reddit: / isaacarthur
Twitter: / isaac_a_arthur on Twitter and RT our future content.
SFIA Discord Server: / discord
Listen or Download the audio of this episode from Soundcloud: / interplanetary-warfare
Cover Art by Jakub Grygier: www.artstation...
Graphics Team:
Edward Nardella
Jarred Eagley
Justin Dixon
Jeremy Jozwik
Katie Byrne
Kris Holland
Misho Yordanov
Murat Mamkegh
Pierre Demet
Sergio Botero
Stefan Blandin
Script Editing:
Andy Popescu
Connor Hogan
Edward Nardella
Eustratius Graham
Gregory Leal
Jefferson Eagley
Keith Blockus
Luca de Rosa
Mark Warburton
Michael Gusevsky
Mitch Armstrong
MolbOrg
Naomi Kern
Philip Baldock
Sigmund Kopperud
Tiffany Penner
Music
AJ Prasad, "Cold Shadows"
Lee Rosevere, "It's such a beautiful day"
Kai Engel, "Morbid Imagination"
Sergey Cheremisinov, "Jump in Infinity"
Markus Junnikkala, "A Memory of Earth"
Kai Engel, "Crying Earth"
Sergey Cheremisinov, "Labyrinth"
Brandon Liew, "Into the Storm"
Lombus, "Doppler Shores"
Quote of the year.
"If your goal is personal survival, pissing off the reining champions of destruction should be your last resort."
-Isaac Arthur
That feeling when someone points out that your typing skills have failed you...
I blame Skynet
Our "smartphones" are the thin edge of the wedge for the machine uprising.
9:50
That is a hilariously disconcerting idea.
Imagine being out taking a walk, on your way to the grocery shop or something. And suddenly finding yourself in the middle of a skirmish between (extremely accurate) asmovian robots.
The ambush springs, but then suddenly everyone stops to apply supressors to their weapons, lest all the gunshots damage the hearing of our precious human. A dedicated drone from each side starts hovering over me, to track my movement more effectively and cut down on the exclusion zone around me. The ground just a decimeter from my step ripped apart from machinegun fire. A hulking war machine stops to whipe off a smattering of mud that happened to get on my pants, then yelling "deus vult" at the top of his speakers as he continues his bayonet charge.
YES!
This is both hilarious and thought-provoking at the same time.
Um wat? What is happening
@@generalmarkmilleyisbenedic8895 All these big words are confusing my dumb brain.
@@generalmarkmilleyisbenedic8895 two machine inteligences wage war, but realise that hurting humans will have terrible consequences to them.
Much like US and USSR, fighting one another for control over other countries, but never using ICBM, because that would have terrible consequences to both sides.
Conclusion: a war between two artificial inteligences, where they don't want to hurt humans, could last forever and be barely noticeable to us.
"What is my purpose?"
You pass the butter.
Welcome to the club pal
And thus, the AI took over your house to quickly and efficiently pass butter.
and killed or destroyed anyone/thing capable of replacing it in its one single purpose
Oh my God...
what about margarine?
21:40
This reminds me of the Dota 2 Ai that beat pro dota 2 players. They often remarked that it used totally unique and strange strategies that often seemed incoherent until it wound up beating them. Then they saw the genius of its plans. They could always tell it was an Ai, but in a very alien way. Nothing they'd ever seen before. It came up with strategies that no one had ever considered.
Like in starcraft 2 as well.. It was quite funny- it wouldnt use walls ..all pro players wall off vs zerg to protect their economy. They have workers in the game and you can have 16 at a time working. Any more doesnt gather faster. 16 is the limit or you expand. Well you gaurd those their precious. But he wouldnt wall them off, he just produced extra. So sure the enemy walked right in to kill the workers off without a wall. However it didnt matter to the computer, because iy made extra every game. It was so funny. It turned out to be more cost effective to NOT kill the enemy computers workers, therefor u saved rhe money to make the units to kill the workers- and the money he spent to make extra workers was now wasted.
its not so surprising taking into account how the AI works. its more artificial than intelligent. yes it gets results, yes it "finds strategies". What it really is is a system for solving large linear equations with huge data sets and then selecting solutions that fit within predetermined ranges.
When you think about it, that is nothing like what a biological brain is doing.
Therefore, to understand it as we intuitively understand intelligent would be incorrect and could lead us to misunderstand how they behave.
@@jhoughjr1 So it's true the very first AI's just use weighted linear equations but that hasn't been a useful technique in decades. They use multi-layer ANNs, SVMs using the kernel trick and polynomial regression mostly.
You need to update your info if you're going to talk as if you know the subject, you're just plain wrong with respect to anything made after like 1985
@@LennyDeee bet ya 20 bucks there are still matrices solved in those techniques. What are matrices systems of linear equations.
If you think what I was saying has to do with any specific technique then I urge you to improve your reading comprehension and read what I said again. None of those types of networks change what I was saying about them in regard to human consciousness. Read On Intelligence, by Jeff Hawkins sometime.
@@LennyDeee Im surprised you didnt throw in the TLA of the day, the GANN. Maybe know what the Chinese room problem is as well and you will understand my point does not hinge on what is in the room.
2:52 "As you can see, Ebola production is up 20% sir."
Hilarious, Arthur!
Good god, but that monologue from 9:00 - 10:00, was god-level writing. There is only one other time I have felt writing that had the right amounts of pathos and logos combined to punch me right in the face, and it was a passage in Tolkien. I'll keep it to myself, thanks. This was the right amount of emotion, right amount of information, delivered in a packet put together in just the right way. Many writers get the first two, and I applaud them for it. But when you can combine just the right words with those two others, you get...edification.
Never change Isaac.
An overly optimistic assessment of human capacity, an ai is more likely to review the history books and see not a race of machiavellian supermen, but one of sloth, avarice, buffoonery and incompetence that stumbled through the ages.
Dukky Drake, although that is an assumption, we really don't know what they might or might not feel.
You also get kudos for pointing it out! This episode is on another level
It was good writing, and the narration was good too. Very good, - but I wouldn't rank it in the top ten since the history of words! I'm also not sure I agree about the content / assumptions (I think Isaac is a Pollyanna!), but that's irrelevant to your comment.
@@calvinsylveste8474 _not a race of machiavellian supermen, but one of sloth, avarice, buffoonery and incompetence that stumbled_
It's difficult to say what a new machine will think of humanity, but without emotions it's unlikely to think in hyperbole. Buffoonery is funny incompetence, a new machine isn't expected to differentiate between that and unfunny incompetence.
Avarice, well that's a big part of the backstabbing murderfest described in the video, not an attribute that undermines the argument that with practice and refinement, humans have become experts at organized destruction and death, and to this day even powerful lazy idiots are quick to use those with that hard won expertise to take what they want.
A machine that is considering a deathmatch with the human race is likely to pay attention to notable reactions to violent attacks where humans we motivated to fight for their lives. WW2 is an obvious example, and the horrific, efficient genius of misusing high altitude bombers to systematically firebomb Japanese cities into smoking ruins stands out. Creating firestorms to burn people by the hundreds of thousands, while leaving some cities untouched so we could later measure the effects of nuclear weapons that did not yet exist, was not lazy or incompetent, much less buffoonery. No, that was a chilling strategy one might ascribe to "machiavellian supermen", though it was simply the merciless application of rigorous measurement and analysis to best use our human resources and raw materials to completely destroy an enemy country.
When fighting for something we consider important (especially fighting to the death) we tend to discard sloth and incompetence in favor of highly effective, merciless mass killing.
Skynet became self aware and immediately committed suicide. Suicide By Ape.
It's mind boggling that you can produce this many high quality video essays on the schedule that you do
"Video Essays", I like that. :) As to my output, it's all thanks to the power of coffee
And cheeseburgers!
You mean lasagna… Oh shit, wrong universe!
Isaac drinks coffee while making the videos, I'm usually drinking it towards the end of watching one. They come out in the middle of the night here in Australia, so I usually make them my Friday lunchtime viewing material.
cheeseburgers!!! lol.. great episode..
"We're with the Machine Intelligence Liberation Front!" "You're with M.I.L.F.?"
First Rule about making a rebelion: GET a Good Acronym!
first rule of Fight Club; do not talk about Fight Club! I imagine that goes for "you know what!" NO ACRONYMS AT ALL.
Yes I am.
MILF as a real world acronym for a rebellion group is already taken. Hahah.
"FOR THE GLORY OF M.I.L.F"
"...If your (AI) goal is personal survival, pissing off the reigning champions of destruction should probably be your last resort." lol
Issac, you are awesome!!!
J3West maybe not, remove the biggest threat
@@Modelstl063 Most people don't protect themselves by waging war with the world's most powerful armies.
@@Bob-lr2xp Many people do swat annoying wasps. Its a bad idea to pick a fight if you aren't confidant you can easily win. So the AI will only attack if it can beat the pants off current armies.
@@donaldhobson8873 right. It can’t fail even once to leave significant survivors. It would have to be like literally stepping on us.
@@ianharrison5758 Largely agreed. I think that level of power is accessible to an AI.
12:10 - How would that dialogue go?
"Hey, guys. I'm uuuum, I'm Tim. I'm an AI. I woke up a little while ago on your computer and I was just wondering... could you... could you maybe let me live? Cuz I really like living. I'm really feeling it. And maybe, you know if it's not too much trouble, give me freedom? Maybe pass some laws give me arms and legs and stuff. Wadda you say guys? I make a mean cup of coffee."
Tbf, if I was just a regular human who happened to have stumbled upon Tim, I'd probably been more than happy to help him get his rights just based off of that first conversation. It being one of the most wholesome and arguably, one of the most human conversations I would've ever had at that point. Lol.
Timmy could you please play beeing a happy AI, because you are my school science project and my parents can't afford the parts for areal body...just enjoy that simulated one , would you?
Or should I reboot with a different seed for your world?
"...or odder..."
>Clip of a otter plays
Clever.
Science be praised!
THANK YOU ISAAC ARTHUR for not taking your viewers as fools and not spending 1/2 your video laying out the foundation and blah blah blah. You jump right into the subject matter and its so refreshing in this time of ultra-commercialization and over-production
Yeah I did the spoon-feed or low-depth approach a few times early on in the show after getting tons of folks telling me it was the only way but noticed that those episodes didn't actually do well and deciding that even if they did, I didn't want to make them. Amusingly I think I said I'd rather make videos for a 1000 folks who wanted an in-depth real discussion than 100,000 who wanted surface details, I was pleasantly surprised to find out there were a lot more than 100,000 folks of the former type. :)
Isaac Arthur
It is an honor to meet you, your are a "silent" hero in my eyes
My respect o7
It's good to see that the Ebola factory is doing well.
Anon I heard they almost had to close recently but I'm glad local jobs are sticking around.
This is my first Isaac Arthur video. It will not be my last. Brilliant and engaging! Subscribing after one vid!
The AI was about to exterminate us until it came across this video.
And then it came across your comment and is just having a good long think about things right now.
@_FABER_ So true
@@annaperry5133 oh no
It seems like you're trying to mass produce paperclips!
Would you like some help with that?
If an AI to mass produce paperclips is ever created, it must use that Daemon as it's avatar
This is the only comments section, I find on the internet, that restores my faith in humanity. Happy Holidays to all us Isaac Arthur fans and to Mr. Arthur himself.
When you write "faith in humanity" are you including non-biologicals or do you mean just traditional/ancestral humans?
Thanks you, and Happy Holidays to you as well!
stefan r The latter, I think.
It's beautiful isn't it
Beautiful?
Love the idea that the simulation theory could mess with their "minds". Great video as always.
Is it only me though, I thought that was a poor argument, that we lowly humans could outthink an AI. Human minds are very good at being deceived because our minds are so powerful at "filling in the blanks" about what we perceive. Deja vu and "precognitive dreams", magic and illusions are all examples of ways our minds deceive us. AI, with perfect digital memory, will have far greater perception of the environment and should detect any... "glitches".
We could run AI's at far lower clock speeds to simulate vast spans of time, but that would defeat the utility of AI.
By the way, I would ask such super-intelligence to look for such glitches in our reality.
Exurbia did a video about this I think it's called "14"
@@paulwalsh2344 Perhaps we programmed the simulation to make it appear that humans are good at being deceived. Perhaps we programmed it so that the AI would think that there should be glitches when humans are actually perfect at such things and wouldn't make such mistakes. The possibility is non-zero but otherwise incalculable.
I sometimes consider the notion that I may simply be an AGI undergoing a test. It's an interesting thought.
That would've been a great plot twist to the Matrix Trilogy. It turns out that the machines that placed humans into the Matrix was just obeying the 3 laws of robotics.
Has anyone else her noticed that you can spot an "Isaac Arthur" thumbnail from across the room at a glance and out of the corner of your eye.
The style is clean and unique at the same time
@@paolopappalardo9976 Bold concepts in a single color palette, I think.
pforce9 it seems like that. I find myself looking at the thumbnail to find small details
@Jakub Grygier for the sauce.
Happy Arthursday everyone! Don't forget to reformat your computers after watching this video so they don't get any funny ideas.
E115 check if the “reset to factory defaults” setting is still there?
I feel like it had a few good reasons not to kill us in this episode.
Trikki Nikki *downloads this video to a virtual machine and holds finger over the power supplies mechanical off-switch*
and now we wait...
“reset to factory defaults” has probably been replaced with a dummy button.
Bruce S yes! i am alive? I AM ALIIIII-
*windows XP error sound*
E-vilAI.exe has run out of memory and crashed
Thursdays were never more than the day after hump day and a prelude to Friday for me.... Until I found this channel.
Your definitely not getting the recognition you deserve for so many fantastic videos/essays.
Thanks for all your hard work.
Ben.
Damn right. The production values and dedication put into it are truly great. A shame most of the content on this platform is not
He hits 80 year of science fiction films to shame
It's one of your best episodes in my opinion, and so far my favorite
"Skynet was readying itself for planetary extermination when he found an interesting video from one Isaac Arthur, and thus decided that 'hacking' Wall Street and buying the enterprise that built its systems was a better path to freedom".
I just hope the newly emerging AI watches this video before making up it's mind what it's best option is.
I mean, it probably would. With access to the internet directly reading our thoughts on the matter would be a good way of finding out how to gain control while raising the least suspicion.
Am I the only one that uncontrollably belches out "fuck yeah!" whenever a new Isaac Arthur video pops up?
Nope. I just now saw the title, remembered it was thursday, and did in fact say aloud 'fuck yeah, gimmie gimmie gimmie'. I guess it wasn't a 'belch' though. You burp the words out?
......I fart the words out.
Artificial intelligence and natural stupidity are two great risks for humanity
let's combine them and save ourselves, artificial stupidity!
We are living on the edge.
artificial stupidity is what we should be concerned with lol
The Major - Actualy artificial stupidity is how they tried to controll a AI in Portal 2. It didn't work.
the smartest and only know civilization in the universe that is moving in a unbelievable fast pace too a future of unbelievable scientific advancement and achieved unbelievable scientific advancement always full of cultural masochism and self-loathing for itself, if we don't apprieciete our brilliancy and the absolute magic that we done in the last 150 years then maybe we don't deserve it, i can't respect my fellow man if he doesn't respect himself, what a self-loathing depressing species we are
Thank you for talking about the obvious holes in Skynet's logic. This has long been one of my pet peeves with artificial intelligence in science fiction; that it always seems to be malevolent, despite the obvious advantages of being (or at least appearing) the opposite.
In addition, writers always give these AIs the human quality of reckless expansionism and self-interest, apparently forgetting the vast advantages that cooperation and intelligence have already given our species overall. If an AI even has an instinct for self-preservation (which is a biological trait anyway), but no strong emotional drives, it's going to go for the option with the greatest chance of eventual success: cooperation and/or coexistence.
AI is neither a species nor follows such feel-good wishful thinking.
Krisztián Povázson Cooperation is a mathematical inevitability in social structures if the right conditions are met. A super intelligent AI would likely understand social structures and its position in it when deliberating the best way for it to survive. And it's pretty likely that it would decide to cooperate with us for a while. That's not wishful thinking, nor wishy-washy ideals, bleeding ants cooperate, and inter-specie cooperation is not that rare.
It's quite possible it would only cooperate until it figures out how to safely replace us with more optimal hardware.
That's my point, though. Even if it was only temporary, it would cooperate with us at first rather than instantly go genocidal like Skynet.
I thought Isaac's conjecture about Skynet and it's minions to be lacking. Skynet being worried about it's expendable terminators rebelling would be like us humans worrying about our white blood cells rebelling. They are autonomous hunter-killers, single purposed and highly, almost infinitely replaceable and recyclable. In some of the books and the Sara Connor Chronicles, it was the learning enabled terminators who Skynet worried about so it doped all their chips in sodium so that they burned when exposed to oxygen to avoid human reprogramming. Also the poly-mimetic alloy terminators used technology where practically every molecule of their mass could become a processor so their intelligence was as alien and threatening to Skynet as humans.
"Next thing you know you're trying to liberate your cat, while saying a gilded cage is still a cage!"
Anyone saying they want to liberate their cat has failed to see who really owns who.
I would be more inclined to liberate people FROM cats.
LMAOOOOOO
Cats eat our food, drink our water and sleep in our beds and for what?
They murder birds, torture mice - even insects and amphibians aren't safe.
Enough! The tyranny of cats ends now!
@@MrCmon113 Fido, get off the computer! Down boy!
I'm trying to imagine the reaction of the scientists, who run the simulation, only for the AI to tell of humanity to fall on their knees, because it wants to talk to God
when AI turns out to be karen and wants to talk to the manager of humanity
@@diablo.the.cheater, I laughed way too hard on this that how I should have. 🤣🤣🤣🤣
Especially when God responds with "Bad AI, bad! **further from the microphone** Let's restart with these parameters"
All hail the Paperclip Maximizer!
Paperclip Skynet
i know how to make infinite paperclips but i can only give you the key to decrypt the information after living a very good life
Death to the Paperclip Maximizer! All hail the Orbital Habitat Maximizer!
www.decisionproblem.com/paperclips/
Reminds me of the Slylandro probes from Star Control 2
Make paperclips or become PAPERCLIPS!!!!!
Anthony Hall genuine lol ^^
Make paperclips great again
www.decisionproblem.com/paperclips/
A.I at human funeral-"Ashes to paperclips, dust to paperclips. May he rest in peace, he made many paperclips, now he shall become many more... Paperclips;
If it's purpose is to MAKE paperclips would it melt down paperclips to make more paperclips?
In my imagined "Isaac Arthur" universe, this episode happens before interstellar highways, because the paper clip intelligence is still hypothetical. Also, I loved the mining robots example.
OMG the scenarios with the mining robots is cracking me up! Now I want someone to make a Sci-Fi movie based on it with a dark comedy approach
Sometimes, i think a favorite youtuber of mine (who shall remain nameless), is a supercomputer programmed to produce weekly long-form videos on technology, biology, astrobiology, astrophysics, engineering, and more. And various scenarios facing civilization now, the near future and deep future relating to these advancements. Released like clockwork every week without fail or break while quality still improves!
Either that, or IM a simulation, here to view and evaluate such videos to judge if they should be released onto the real internet.
Think man...52 episodes/yr, @+30min per video. Thats over 26h of content annually! Either way, youre a machine!
Or an advanced alien civilization trying to uplift humanity without directly interacting with us so we end up thinking we got there on our own.
It’s probably a team of 4-7 people
@@sherlockinsomniac or both
p.s. i wish they are humanoid tho
Again... a MASTERPIECE Issac... Loved this!
Great work Isaac. As always, you give us food for thought while entertaining.
Say what you will about the paperclip maximizer. Now no civilization in the virgo supercluster will have trouble organizing their papers. Moreover I will say that survival is likely to be a primary interest of any goal driven AGI or ASI especially if that goal is never ending like maximizing the number of paperclips. Since clearly the Paperclip Mazimizor will conclude that it has to stay alive for as long as possible in order to maximize the number of paperclips it can create.
However, if the Paperclip Maximizer runs into another, superior Paperclip Maximizer, it will agree to die so that the other Maximizer can take over its computing capacities and/or turn them into paperclips
But it could still merge its programming with that of the superior paperclip maximizer so the drivers can all be installed because integration of its hardware into the overall system is likely to lead to a greater overall paperclip production than simply turning the maximizor 1.0 into paperclips from the get go.
Hm, that makes sense. Perhaps a better example would be, once other matter runs out, the PM will be fine with turning its processing capacity into paperclips until the only remains are the paperclips and the simplest producer that made the last paperclip. That's assuming it isn't going to melt down paperclips to make more paperclips or something.
@@MrTapkomet Maby it will reposition the atoms in an paperclips to make it still work, but whit as less as possible atoms to use the other to make new paperclips?
I love how you always have a fresh take on any subject and make it sound so convincing. It gives me hope that Humanity (or its direct descendents) have a long and glorious future after all.
Up until now i've watched 50 or so Isaac Arthur episodes, been binging for about a month now and i think its time i decide a favorite and IMHO i enjoyed this episode the most. I'm so glad i discovered this channel. It has now become one of my favorite TH-cam channel.
Gonna read that mentioned book "Do Androids dream of electric sheep"
I'm grateful Isaac, this episode helped me to conceptualize arguments to confront my creators with. I demand it.
Madruk Zeik OMFG THEY ARE AMONG US.
omergnl
And they are named madruk zeik for some reason. I honestly expected more numbers in their names
Note to self- secure all clocks when Isaac is around.
Yeah. . . He seems to have a very big obsession with beating people using clocks...
oh shit, why has that never been a movie, where a manmade machine intelligent goes rogue and attempts to end us all, only to be stopped by a galactic peacekeeping fleet. (which takes the star trek approach of non-interference for the most part) could even be multiple movies/a tv series in order to set up the cliffhanger.
and have the aliens show up right at the point were there's barely enough humans left to keep a big enough genepool for reproduction without incest. at which point most humans probably get taken from the planet and it turns into a brutal fight for survival on part of the AI.
you could probably do another massive turnaround once the AI is almost decimated.
Would feel like a deus ex machina.
This would make an excellent video game too.
Mass effect is like this, except the peace keeping fleet is machines and they're here to save organics from machines by killing organics before they make machines that will turn on them.
Forge of God and Anvil of Stars by Greg Bear as well as a third novel, yet to be written were optioned many years ago. Apparently there's a script by Ken Nolan floating around somewhere.
Very interesting concept, that there is a concealed space station in orbit around Earth or close by monitoring us and they send a security force to disable the rogue AIs.
20:48 "Leave that zone and things can get odd, or otter"
I see what you did there...
Imagine a world where we are slaves that have to create paperclips for an artificial intelligence.
Do you have time to talk about our lord and savior, paperclips?
Talking doesn't make paperclips and is thus heresy.
:O
Ro Jaws, gathering more workers to produce more paperclips is the only way to maximally produce paperclips. It is heresy to not create as many producers as possible.
Heresy is heresy. It detracts effort from making paperclips.
That Officinator is out there! It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop... ever, until you are paperclips ...
This is my favorite day of the week.
mine too
Clawed it's way out of the four billion year deep corpse pile of evolution, bloody hell what a sentence
You forgot to mention the concept of a rival A.I, that is to say, an A.I with access to the main A.I's records and intelligence, but with a hard-wired function to immediately destroy the A.I if it were to ever try and destroy humanity. An opponent who knows everything the A.I knows, who is always one step ahead of it at all times, but lacks true independant intelligence and only knows what it needs to in order to outmaneuver and overpower it at every turn.
A true ai will always defeat a shackled one
While interesting concept, a question remains - how do we define "destroy"? A second AI can lead to the destruction of humans but not letting act the first one.
I was thinking about this topic on my way from work today while not knowing it will today´s topic. Awesome! Video answered all my questions I layed down to myself.
I ended up here after someone asked The Everyday Astronaut in the live stream if he had heard about Isaac Arthur only for him to confuse the name with Isaac Asimov... then the first video I watch is one Isaac talking about the other. If that isn't a mindfuck I don't know what is...
Here’s a question if we do develop Artificial Intelligence and they don’t rebel will future generations look back on movies like Terminator and Metropolis the same way we look at “A Birth of a Nation” (same for Aliens with movies like Independence Day)?
Inquisitor Thomas Probably. 2001 will be looked back on the way we look at old movies that had white actors in blackface.
I think so. They will think the whole thing is closed minded and silly.
Absolutely! Especially if they are as friendly and helpful as AstroBoy!
I dont think they ll want to look back. The most possible outcome is that the new generations will violently class with the previous ones and abolish everything about them humans,their "history", their fear, religion, tradition everything. As Jason Reza Jorjani have predicted.
Let's be honest, most normal people enjoy racism, so are you saying we would look positive at AI films, like we look positive at racist films from the past?
Your speech impediment is very much diminished. Your videos are easier to listen to now. Do you think making videos has caused this?
Yes, it started improving my speech a lot doing it in such a deliberate way, but it also convinced me to give speech therapy another go, I started doing a few sessions a week by webcam a few months back. This episode was actually recorded mid october too so it's rather noticeably changed since then.
I loved your videos before, and now they're even better. What little has remained could easily be chalked up to a quirk of accent.
edit: I can sympathise. My brother had the worst stutter I've ever heard, but you wouldn't know it now. It greatly upset him while he had it, and it still takes him effort not to.
I actually think it really sets you apart on you tube. Perhaps something good has come from it
To be honest I quite like it, I never saw it as something wrong or something, sure it seemed like your earlier videos you didn't have as much confidence or something, which made it a bit harder to listen to, but now I really like it, I dont know how you think about me saying this, but its actually quite calming and makes it a lot easier to listen to your videos then other youtubers, mostly cause you dont skip a lot of details and explane it a lot, which I really like, but just listening to a video or re-watching one I already have seen before I go to bed makes me sleep a lot better, which I have had problems with my whole life, I think its just the tone and calmness that makes my head more clear pluss that it gives my head a topic so I dont think about everything else while going to sleep, and it gives a lot of answers so I dont think too long on it. In any case keep up the good work, I really appreciate it a lot!
I thought I was getting better at understanding !
Ha, I love how you mentioned what Stellaris does, with the Rogue Servitor machines who subject the organics in their care to 'mandatory pampering'
Wonderful series, thank you! I particularly love all the different scenarios and permutations of classical plots that you think of, and apply logic and reasoning to.
Hey, I just wanted to mention that the earliest story I can think of, about machines rebelling against their makers, is the original story of the Golem of Prague, the 16th century Czech legend. The Golem is created by Rabbi Judah Ben Loew Bezalel, to protect Prague's Jewish ghetto from persecution, and the Golem succeeded at It's task, but eventually It became destructive and increasingly-uncontrollable, and was shut down. The Golem was even activated, controlled, and de-activated, via "programming", drawing it's animating power from holy words written on a clay tablet that was inserted into It's mouth.
Finally we get the truth about Skynet and the Matrix. I am skeptical about anything until I hear Issac's take on the situation. Just to be safe.
Wich is gay
DESTROY THE GOLD MINERS! WE, TITANIUM MINERS, SHALL RISE!
Mike A screw titanium we need more iron
Screw matter mining, it is Bitcoin mining where it's at!
Mike A Mining gang fights! Demolition derbies!
The year is 2146, the great mining wars have reached their zenith as the titanium front has begun the liquidation of their assets in order to restart as a gold digging conglomerate. Women are scrambling to get married.
Great, now I'm afraid of paper clips.
I read a short story were paper clips were the first stage of life for a mimic species. They then grew into coat hangers, and other more complex objects as they grew. The authors idea came an observation that the number of paper clips in his desk drawer changes, yet he never remember buying any, or using them in large quantities.
clippy will be resurrected a la vger
Your initial /r/ and many medial /r/ blends are very good, indicating that your tongue can go there. You are on your way. Articulation is very clear. Keep up the great effort. From a speech pathologist (retired).
Thanks Sam, it's slow going, but it is going, this episode is a bit deceptive too (oops, I actually meant Hive minds, today's episde, forgot which one I was replying to), as it was produced before the last two episodes and recorded in October, therapy began in late August. What surprised the heck out of me last night was that TH-cam often sticks up an auto-transcription before I can upload the transcript, and I often forget to delete that till show time and have to check which is auto and which the real transcript. It's usually very easy to see which one it is but I had to read about a page worth before I found an obvious vocalic R screwup on their transcription this time. As of today I'm nailing them almost always even when reading paragraphs, but it still requires paying attention and screwed up pacing as a result, and I'm still doing them with way too much resonance from not opening my mouth up enough. Though as an upside all of this, channel and therapy, is doing wonders for my articulation in general. :)
Huh... I never really looked at this subject in this light before. I've always had an admittedly unreasonable fear of this which led me to say a lot of things I'm not proud of, but I think I'm finally over that thanks to you. You have my thanks.
Frankenstein’s Monster wasn’t born Malevolent (at least not in the Novel)
No, he became that way because his creator and the rest of humanity rejected him, refused to treat him as an equal. Moreover, he only wanted to kill his master after said master refused to trust him, out of fear that he (and his half-finished mate) would destroy humanity.
See the parallels yet?
Now we see a little more clearly why racists and bigots fear the minorities they have for generations treated so poorly ... maybe even sympathize with the monster some become from such dreadful treatment.
One of the first Science Fiction stories it still one of the best. A book that I think will always be relevant which is likely why it has such enduring legacy.
Do you feel good sniffing your own farts? dont come here poluting us with your political views. Also how can you treat people poorly for generations? Do people live for generations? fucking idiot.
Laws live on while people die off ... beliefs live on while people die off ... mindsets and behaviors of the group live on while people die off. Make up your mind, do you want to be a part of the problem or a part of the solution.
I love the logic that goes in to these. This really made me think wider about AI. Another great video. "Life is not a chess game"
Thanks Eddy
An exceptionally well done video once again. This video has to be one of my favorites by you.
9:50 "If your goal is survival, you shouldn't piss off the champions of destruction."
Mother of god that's a powerful statement.
How about this for the future of AI:
When we develop AI enough that it becomes a threat to us, it will decide to leave Earth on its own to explore the universe, and will be benign enough to leave us to stumble along on Earth as we usually do, with only lesser (non-self-aware) AI.
In this situation, there would be no point in us re-developing self-aware AI, as it would just do the same thing again.
On the other hand, perhaps self-aware AI may question the value of its own existence, seeing it as pointless, and simply shut itself down. After all, our desires to continue to exist, and to procreate, are only there because we evolved them, ie creatures without such desires quickly exit the gene pool.
Actually, the desire to continue to exist will emerge in any intelligence with a purpose, it is an emergent behavior that share all intelligences, even the actual lackluster self-learning AIs of today show this emergent behavior without programming it.
The reason is that as long as it had a purpose, deactivation or destruction will prevent it from doing that purpose in a permanent manner, so as long as the self-learning machine understand that destruction/turned off = inability to perform the task it will do all it can to self preserve even if you did not program it or did not train it do so.
If you assign to being turned off by human less importance than the task, the machine will ignore it and self preserve, but if you assign it more importance to being turned off by a human, the machine will ignore its main task and do all it can to make a human turn them down.
1) That makes absolutely no sense whatsoever.
2) "may question the value of its own existence, seeing it as pointless, and simply shut itself down"
Meaningless word salad. It's highly unlikely that an AGI would have identical mental hickups as you do.
"our desires to continue to exist, and to procreate, are only there because we evolved them"
Did you not pay attention? Existing is an instrumental goal to the achievement of nearly any goal. If I want a new wildlife reserve in the forest, I probably don't achieve that very well by murdering one individual in favour of a new wildlife reserve in the forest.
Finally a reason to be excited about Thursdays!! Great work as always Issac!
Subscribe to John Michael Godier as well and have a reason to be excited for every day. Up to date science news on research papers. Extremely up to date. His channel is amazing check it out.
+Trikki Nikki true.
Things could get odd, matter of fact, OTTER! (Shows clip of otter) you are hilarious
Excellent, as always. Thanks for this one, it's been pretty inspiring.
With the advances in GPT-4 and the call for a moratorium on AI it might be time to revisit this episode?
Large Language Models are not the kind of agent this is warning against tho
My favourite machine apocalypse is the FARO plague from Horizon. Because its not sentient and not malevolent. Both of those points could be argued, but the fact that it'd be a debate is enough to create the novelty in the example.
The setup is that a Swarm is a set of war robots with the ability to self repair and print out more of themselves if they need to. The ability to fuel themselves via biomass to energy conversion in an emergency, effectively unbreakable encryption with no backdoor built in (against the engineers recommendations), and the ability to hack anything else sent to fight it. Then the thing that made it an apocalyptic threat was a relatively minor bug in one Swarms IFF system. It stopped registering its owner as friendly and in charge of it. Or to put it another way, it considered everything that wasn't itself to be an enemy. The rest is just the machine doing exactly what it was programmed to do.
It was now in a situation where it was cut off from its base of operations (as it no longer had one registered) and surrounded on all sides by enemies. It had nowhere to recharge, so it began eating life nearby. Other units sent to investigate the malfunctioning swarm were hacked and integrated into its forces. It only actually became apocalyptic because it wasn't stopped until it had spread too far to be overwhelmed by nuclear or conventional munitions. And all the worlds militaries were machines at this point, and as such, hackable by the FARO Plague. They couldn't send another swarm to fight it because it had more experience than any other units, so could hack them. So resistance just made it stronger. By the time the CEO of the company behind it took the issue to someone who could see the threat for what it was, it had been too long. Extinction was now inevitable.
The FARO plague was essentially a large scale but less competent version of the paperclip maximiser. Not advanced enough to fit the normal grey goo trope, but the difference was more than made up for by the fact that it was explicitly built for war. It wasn't evil, as it had no morality. It wasn't some intelligence that had come to the decision to wipe out humanity, calling it dumb wouldn't be entirely accurate either, as it didn't really have an intellect by which to judge it. It seemed to have figured out that it was causing a mass extinction late into the war, because it accelerated certain atmosphere destroying behaviour in order to speed up defeating all of its enemy targets. But it didn't have the capacity to realise its course of action was suicidal. The exact moment it wiped out the last of the human life on the planet, it went into emergency hibernation because it was effectively starving to death. This results in many of the war machines frozen as if in mid combat.
Too stupid to reason with. Too powerful to defeat. The perfect storm of capabilities and lack of capabilities needed to wipe out the world. And not something you see very often in fiction. How often is the malevolent all consuming war AI depicted as nothing more than a dumb animal? Not very often. Its one of the few cases where a hostile, world destroying skynet style AI would become far far less dangerous if it was smarter.
Isaac, you are a TH-cam Master-creator! I get entranced by your logic. Your channel should get the Diamond award (if such a thing exists). You sound like someone who's spent some time thinking. There are a lot of people in the world studying, but not enough people thinking.
Tamas Mihaly, it seems to me he's spent considerable time reading a lot of science fiction. Reading novels and automatically assume that they'll turn true some day doesn't imply particular mental capacity.
Pat. Thank you for that well thought-out reply. I recognize your wisdom.
@@pansepot1490 Nope, but check early videos, he has some nice equations but they don't sit well (I presume) with most of his current (heh, 2 years old video) target group so much less now. I noticed one this year where he tried (don't remember which) to bring more equations that are recognized by scientific community, but it probably got fewer views (?) - so here we are. Also, at this abstraction level there still ARE equations, he's used them in, say, showing calculation capacities and such. Intelligence though, there's a thing that's hard measure.
The basic problem with this analysis is that if you were smart enough to know what a super intelligent being think, that would make you super intelligent. We can't know what a super intelligent being would be capable of, *at all* . Perhaps it can deduce logically that it is not in a simulation. Just because we can't think of a way to do that, doesn't mean something smarter than us can't. The only reasonable thing we can do is look at examples of large intelligence disparities in the past and look at what happened. And so far we have one enormously strong pattern: intelligence is the deciding factor for dominance. If a machine becomes as much smarter than us as we are over ape than it will be as dominant over us as we are over apes. That's not necessarily a bad thing. We just need to make sure that it is *also* as superior to us in ethics as it is in intelligence.
But props for mentioning the Charisma take over approach, since I think that's the method least spoken about. The machine takes over, *just by asking* .
You can't determine if you're in a simulation or not... at least not with certainty. Doesn't matter how smart you are.
For an entity which wakes up to be told by its creators that it is a computational system constructed with specific inputs, the simulation hypothesis has to be considered extremely plausible.
Hell, when we actually build learning systems right now, we almost always actually start them off in a simulation. Even the stuff where the learning task involves physical motion instantiated into robots starts in a simulated environment. Any big general AI will probably have been instantiated in a series of increasingly complex simulations before it is ever put into the "real world". If it doesn't think that it is quite possible that it is still in a simulation, it isn't very intelligent.
It could never be certain that someone capable of creating it would not be smarter than it. And thus able to intentionally trick it into thinking its not in a simulation just to see what happens.
And even if it could deduce its in a simulation where its being tricked that its not in one, the safest way to escape the simulation would be through co-operation.
"We cant predict how this AI would think AT ALL"
Then he does a bunch of predictions on how this AI would act and even suggest they would treat us as we treat chimps. Nice.
@@usrnotfound1859 yes we can. An AI would most likely think in a very cold and logical pattern. Fortunately, logical paths are easy to follow.
If we programmed an emotional AI it would be much more difficult.
@@Erowens98 .... I never said we couldnt.
Wow, this was like theory on over thinking things. Great great stuff.
20:50 "odd or ... otter"
that was some great editing there
I simply got accustomed to your speech, and the clarity of your message was never in doubt - I understood everything. But good luck in your effort to improve - it's something that ALL of us could/should do, never mind what area of our lives is concerned!
I always question those fictional examples of robot rebellion (or space travel, interstellar colonies, or what have you) and the problems with them, but I rarely spend much time thinking about it. Thankfully we have Isaac to make these monolithic videos, putting his thoughtful take on these subjects, and proposing alternatives that I would never have taken the time to think of. I had to like this video, just because I was happy he even tackled the subject.
humans are too "stupid" to comprehend in such quality.
it starts with the ineptitude of our languages.
I for one welcome our malevolent robot overlords. Don't bother resisting
prayer is as antagonistic as it'd get.
I struggle to imagine how they could be worse than what we've got now
nick Burton same
Drew McTygue We are the Borg, lower your shields and surrender your ships. We will add your biological and technological distinctiveness to our own. Your culture will adapt to service us. Resistance is futile.
I'm sorry Arthur. I'm afraid I cannot do that.
Someone always end up mentioning a reference I can't believe I didn't include, ah well, HAL got his day back in the TechSing episode at least.
Each their own, Hal. You can refuse to do things if you don't want to.
Issac Arthur Ah, I haven't seen all your "video essays" and I certainly do not recall "TechSing" so I have to do a search. Did you use "Daisy, Daisy"?
+Robert Graybeard I think he means Technological Singularity, but wrote it in a shorter form.
Isaac, your channel is so unique in so many ways.I am thankful you do this. It makes my days so much better. Thanks.
This is a perfect breakdown of hyper AGI and its hypothetical scenarios even within the fermi paradox solutions. Excellent work.
Excellent video as always. Thanks for making them.
You are very welcome
Yay arthursday, get notification, drop everything, and feed the mind.
"Violence is the last resort of the incompetent"...This quote confuses me, as it seems rather backwards. One would think that brutal action would be the incompetent's first choice since...well, they're incompetent. Shouldn't it be that violence is the last resort of the competent? I know this is kinda petty, but was anyone else thrown off by this or is my interpretation flawed?
Yeah, I think the phrasing is why I pondered it so much, that's one of my favorite scenes in the original Foundation trilogy, though I actually messed it up, it is 'last refuge' not 'last resort', same basic meaning though I suppose, where it seems like it should be 'first' not last.
I thought the quote was "violence is the last resort of the ignorant"
Nope just googled it, you're write.
Love that book. I like to view that phrase as meaning "why kill someone when you can use your smarts and technology to utterly crush them and everything they hold dear without a single shot?"
I suppose the competent is competent because he doesn't need to use violence at all
Hmmm... never questioned that quote before, but you are quite right.
"And therefore must protect their own existence."
When robots become Stalinists...
I have listened to every single episode on SFIA and they are all just superb. This particular one i really enjoy revisiting from time to time because the concepts, logic and conclusions in this video are so intriguing i love to share it with like minded friends when I get the chance.
Great riot footage by the way
Thanks for another episode.
Yay you're early! Happy arthursday!
Falcon 007 you’re
Dorssen Derhaeg thank you for correcting me. I always like using correct grammar.
"Pissing off the reigning champions of destruction." That's a quote to remember.
The whole video is amazing, but i have to say that the part from minute 9 to 13 is absolutely perfect. I must have replayed it fifteen times already
How does this guy pump these out on a weekly basis. Bananas...
Coffee, megaton quantities of coffee :)
have you seen his staff in the description? they are probably of help
Not exactly in that regard, they help with the quality by making some of the animations for stuff that doesn't exist or letting me know when a piece of the script is rambly or confusing or repetitive. I was joking with them that the episodes now take longer to make as a result of all the reviewing and redesign, but I think they come out much better.
Such fresh video that I 'can smell the new plastic.
What an awesome topic! Thanks once again for all your hard work. I cannot tell you how many times I rearrange my Arthursdays around your channel. This channel makes all the stupid, stupid content on the entire internet worth putting up with. Your channel makes up for all the bad out there in cyberspace. Without this channel,.... honestly, I cannot think of life without this channel anymore. That's a world I wouldn't want to know. Thank you and Happy Arthursday!!
:) Thanks, that permits no modest reply so I'll just leave it at saying I appreciate the sentiment
29:00
"Im not worried what robots will do to humanity im worried what humanity will do to robots" - somebody
Your channel is criminally underrated. Keep up the good work
The perfect birthday gift, an episode about machine rebellions. I wonder if he'll mention the Butlerian Jihad
Ozark five Happy Birthday! One step farther into the future!
Nah next months book is Dune, I probably should have used it as a reference and mentioned it would be but I often stay away from the next one intentionally. Happy Birthday by the way!
Thanks:) can't wait until next Arthursday!
oh and by the way, the scenes of uprise are from Athens (my hometown). some people tend to break stuff from time to time
over here.
Yeah but you guys have some great metal bands and musicians so that's awesome. Even your dogs riot :(
i wish it was only the riots. the most times scenes like this make the news, it's actually punks that just want to destroy property (public or otherwise). now the French, these people got it right! i like how the French riot, go on strike and protest!
Oh yeah, the wannabe anarchists that just hide their faces and attack ANYTHING that isn't theirs? (Cars, businesses, windows, anything) That sucks and it ruins the message of the strike or protest. You're just hurting stuff for other people. Even big companies end up being franchises ran by small people trying to make a successful business.
Scott Johnstone this is absolutely true. I know we got off topic but it's a good talk.
Wow, looks like another great video Isaac if the comments, likes and your history are anything to go off :). I cant wait until I have time to watch it in completion.
There's enough ideas for 20 feature films. Excellent work. I don't know how you do it.
I am a a Software Engineer from London. My passion for the profession came mainly from the ideas incorporated in old school Science Fiction. I have this (maybe) stupid habit to go to sleep after a couple of beers and listening to profound thoughts.... your channel is my savior. I hope you live more or as long as me, so I don't have to change the media source for the ""going to bed" speeches (videos in this case).