you do realize nobody make a real ai yet .. so unless joe is a super intelligent programmer who will make it himself and teach you .. what did you expect
learn to think in terms of how to get money into your pocket, instead of getting money out of your pocket, based from the existing system. if i were out on the streets with nothing but my intellect, first i'd start making money off of garbage, sell to junkyards whatever junk they'd accept (steel, glass, paper, plastic, these fetch loose change per pound or kilogram, collect enough junk and you slowly command a workable amount), do this for a year. then slowly move up to more dignified sources of income, like selling candy, newspapers, streetfood and other small items, and if you can, start recruiting (your similarly-homeless) friends into slowly becoming your employees, delegate some of the work and slowly multiply your man-hours. all these should give you enough muscle to rent a place. don't stop there, once you've had elevated yourself from being homeless, slowly climb the financial ladder and start looking into selling lots, houses and condos and getting involved in precious metals like gold and silver (while still never letting go of selling junk, candies, newspapers, streetfood and other small items, all this time you should still maintain ALL your sources of income, you're relying on rented replaceable muscle anyway). after this point you should have already acquired a little money smarts, you should start looking into throwing your money into stocks and derivatives, government bonds, insurance and mutual funds, and cryptocurrency.
I have been watching YT videos for hours a day for 10+ years. Your channel only now appeared as a search result. How is it that a channel with 7.5M subscribers and directly aligned with my interests has never been recommended to me.
I worked on this for some time and it can all sound very complicated but the simple fact is AI starts with two things: a) memorization (storage of vast data), and b) comparative analysis (of the stored data). Your program needs to 'run' several hundred times (thousands even millions of times) to get best most accurate outcome and 'prediction' based on these calculations. You get the expert 'knowledge' of AI when your program is able to memorize, compare and anticipate outcomes based on given data and problem given.
I'm only 14 so I don't know much about Artificial Intelligence but I'm working on a story with a robot that is always mistaken as a human. The way I made it seem possible to create a robot like this was by saying that they became very human-like by trial and error. They would sit in a room and talk to a person for quite awhile. whenever they say or do something wrong the person watching them would tell them what they need to do instead and the robot would automatically correct their actions to work around that. I was wondering if that seems possible in real life, a robot that is so self learning that they can look, act, and sound like a human being.
The ultimate way to keep AI on our side is the power of moving on. If you give yourself an ultimate goal and tell the AI to assist you. It most likely will
I'm 17 and created a learning AI, which teaches itself through evolution and won a landwide coding competition in Germany with it. I think i'm allowed to be proud of myself now.
The two big examples of Deep Blue and Watson are no real AIs. They are database-orientated search engines that simply are able to go through massive amounts of data at superhuman speeds. They are NOT self-aware artificial entities that can really go beyond playing Jeopardy (a game hardly suited for a real AI test). Honestly, most of this video is a joke.
Watson is still able to approximate an understanding of the questions asked. It may lack agency, but it's capabilities are starting to close in on humans.
Nikonikoneko There's actually no proof for that at all. Watson more likely exploits the limitations of the game (for a human player) itself with it's near-instant access to a rather huge database. I'd love to see Watson solve other problems than Jeopardy. The first version of Watson failed providing above 50% correct answers because it did not have access to hundreds of thousands of encyclopaedic entries (both offline and online (magazines, scientific journals etc.)). It's first version had to approximate the reliability of an answer based upon a more limited database and as a result was way more often giving incorrect answers. In my humble opinion the whole Jeopardy thing was mostly a marketing thing. The actual AI at play is fairly underwhelming. As a result it's really a bit shortsighted to think it's capabilities are closing in on human capacities. I'd say that's incorrect.
PHeMoX My understanding was that IBM is developing systems based on Watson which should i.e. support technicians based on manuals. I don't claim that its an AGI (artificial general intelligence), but as far as I know it has far more applications than any other knowledge system. I'm not really up to date though.
Nikonikoneko I think you're largely correct. I do think it's useful technology and will have it's application, especially when further developed. As such I'm not really downplaying the importance of Watson and various other projects before it. I just do not agree it's actual artificial intelligence. Yet. It's relying way too much on databases of information to really be a more independent artificial intelligence (I guess AGI) and there is not enough learning involved. The Jeopardy experiment almost failed for a variety of reasons. Anyway, in my opinion it's not real AI when it's basic search routines work something like this ; 1. determine a search that makes sense based upon the question asked 2. take parts of the question and do a brute force search using all relevant database entries available. 3. estimate the relevance of the database entries that returned information 4. do a linguistic and logical check based on similar known questions and answers and 5. combine the results with the highest probable rated answer being the answer given by Watson. I'm not claiming this is exactly what happens, but ("Jeopardy") Watson should roughly work like this. It's essentially a really fast search engine. In other words, take away the database and there's virtually no intelligence left. The whole point of AI should be arriving at a point where the machine starts to learn, understanding the world more or less by it's own. Supplying access to a huge database is not quite that. Why not let a machine decide for itself which data to add to it's library of knowledge? Why not add routines that predict, deduct and confirm relevance accordingly? I guess you'll understand that having access to virtually every answer by being able to search a very very big database is quite different from estimating or deducting a logical answer based upon actual intelligence and understanding. To put things in perspective, some of these AIs have long crossed the point where a human would have the same library of knowledge at their disposal. Well... wouldn't that be... cheating?
John Kirk Buy a guitar, learn how to play a couple chords/scales, go up on stage, dance on stage, destroy your equipment, rinse and repeat. The more you do that, the cooler you look and the more you learn.
I don't understand how any of this can scare people, the idea of A.I., and computers that learn, and that can actually pass as a human is amazing to me, even the idea of ai wanting to kill everyone doesn't scare me, because it just amazes me that a computer can learn and have those thoughts and desires, all *we* need to do is help A.I. have normal, healthy, sane thoughts and desires.
That last part about an off switch is a really important point. Computerphile has a video titled "AI 'Stop Button' Problem" that describes the difficulties of making an off switch for an AI
They should first of all change all the media, and create a lot of content about IA having to serve humans, then they hard program it into their memories , then they create those emergency buttons to reset them, switch them into defensive mode and attack other IA's that treat humans, or destroy them.
I see so many reasons how robots won't take over the world 😂 pitting that aside, its very intriguing, I love artificial intelligence, and robot arts, even science.
The ai we are creating are getting stronger and smarter until eventually they become aware enough to realize that humans are a threat and decide to do something similar to terminator
Though I came here to learn more about developing an AI system. Instead I was educated on all of the existing AI systems built by other people. Thanks AWE me.
Hahah, this was fun, I was doing some research for my sophomore science fair project for a voice synthesis and even when I knew this video didn’t have what I was looking for, I couldn’t stop watching.
The true problem with Strong AI is the lack of ethics and morality. If you create a strong AI we could live with it peacefully... until it believed we are a threat and would snuff us out of existence. So the question becomes: how do you teach a machine what being a good sentient being means without firing back at you (I Robot)?
Simple: make America FREE again. Get the government out of people's lives and businesses. Then reform the educational system to actually EDUCATE people instead of just indoctrinating them. Teach people to think for themselves.
I hate how you only talk about AI from the western point of view and not add from Japan's Point of View where you can have a companion AI.. they even a dog AI right now
You basically just told us all of the experiments that people have made.
NOT how to make them.
He said go study. That's pretty much it
you do realize nobody make a real ai yet .. so unless joe is a super intelligent programmer who will make it himself and teach you .. what did you expect
ssuo ben plenty of AI rightnow. All algorithms. Bxby, Siri, Alexa etc.
Samurai Medi those aren't self aware.
@@odgarig8601 how do you know?
1970: I think we would have a flying cars in the future
2018:trying to understand our brain
Epic how to: Survive a riot
Challenge Accepted That one seems like it'll be pretty good to have in the US here soon...
yer that would be helpful for people living in us after trump became president
Challenge Accepted that actually would be helpful considering there's alot of whiny butthurt libtards rioting in the streets
lol
stay inside your house
How to : Survive Being Homeless.
No doubt you are also Indian
just stop being poor
Just go home 🤔
@@pranabsr.5923 🤨
learn to think in terms of how to get money into your pocket, instead of getting money out of your pocket, based from the existing system. if i were out on the streets with nothing but my intellect, first i'd start making money off of garbage, sell to junkyards whatever junk they'd accept (steel, glass, paper, plastic, these fetch loose change per pound or kilogram, collect enough junk and you slowly command a workable amount), do this for a year. then slowly move up to more dignified sources of income, like selling candy, newspapers, streetfood and other small items, and if you can, start recruiting (your similarly-homeless) friends into slowly becoming your employees, delegate some of the work and slowly multiply your man-hours. all these should give you enough muscle to rent a place.
don't stop there, once you've had elevated yourself from being homeless, slowly climb the financial ladder and start looking into selling lots, houses and condos and getting involved in precious metals like gold and silver (while still never letting go of selling junk, candies, newspapers, streetfood and other small items, all this time you should still maintain ALL your sources of income, you're relying on rented replaceable muscle anyway). after this point you should have already acquired a little money smarts, you should start looking into throwing your money into stocks and derivatives, government bonds, insurance and mutual funds, and cryptocurrency.
Epic how to: how to make epic how to
Artur Eksi I think they did that already:)
Artur Eksi yeah they did that on the new year episode
Instructions unclear; ended up making Ultron 2.0
For some reason i would like to see Epic How To and HowToBasic work together.
Oh boy, that's gonna be some chaos
I have been watching YT videos for hours a day for 10+ years. Your channel only now appeared as a search result. How is it that a channel with 7.5M subscribers and directly aligned with my interests has never been recommended to me.
Wow best animation yet Andy!!!
GREAT JOB ANDY U DA MAN
Julien Paris This show would be nothing without Andy
Ummmm he didn't animate this time.😐
Alex Delrio umm... he didn't animate this time
Good ol' Jackaboy ok I didn't know that
I thought making an A.I. was cool and all until i learned that they can also make creepy stuff😂
I only sub to this channel for epic how to
V S RAJU same
Well they dont make epic how to animore
sunday superman sad
Epic
I worked on this for some time and it can all sound very complicated but the simple fact is AI starts with two things: a) memorization (storage of vast data), and b) comparative analysis (of the stored data). Your program needs to 'run' several hundred times (thousands even millions of times) to get best most accurate outcome and 'prediction' based on these calculations. You get the expert 'knowledge' of AI when your program is able to memorize, compare and anticipate outcomes based on given data and problem given.
Where did you like how to make AI?
let the AI program itself, i dont have time :v
xenazai zm you are the best
Thats actually a concept called seed AI !! :O
Actually, some AIs improve themselves by adding more code to their own programs. Cool right?
Machine Learning period
Your Spanish right?
Brother I"m 28 seconds in and you already have me cracking up! I can't wait to watch the rest!
I'm only 14 so I don't know much about Artificial Intelligence but I'm working on a story with a robot that is always mistaken as a human. The way I made it seem possible to create a robot like this was by saying that they became very human-like by trial and error. They would sit in a room and talk to a person for quite awhile. whenever they say or do something wrong the person watching them would tell them what they need to do instead and the robot would automatically correct their actions to work around that. I was wondering if that seems possible in real life, a robot that is so self learning that they can look, act, and sound like a human being.
as soon as a computer becomes self aware and learns logic i will turn
I want a tiny robot with an amazing AI with personality and an Australian accent.
PsychoPanda9000 I think there is one. Cosmo by anki.
Otaku Inventor ik but I'm talking about one that can tell me banterous jokes
PsychoPanda9000 me too, mate!
PsychoPanda9000 well, that would take a few years to create, but also do you have the money to buy it?
You Can Give Siri An Australian Accent...
Lelelelelelelelelelelelelelelelelelelelelelelelel
Epic how to be an assasin from Assasines creed
King 17 1. Find guards who are worse than a blind cop
2. Hide in hay barrels and jump in them
3. Jump off of 100 feet walls and survive
You can’t
Not a "how-to"
...point proven
I was hoping they'd talk about features of an AI rather than examples of them
true but funny
I think this video is made by a ai video generator
Thanks dude! I love I saw this before the video hit 10 seconds (doesn't count as a view). Underrated but I appreciate it.
I've seen a lot of these videos, but I think this one may be the first one that I believe deserves a like
Thank you for this episode. AI is a really serious thing, and all humans should be aware and concerned about it.
Or how about you just learn about it.
What you do now there is chat gpt
these videos are the only reason I still go to youtube
Epic how to create a virtual reality like sao (sword art online)
ALREDY exists
@@pearlchinasa1770 its not
@@fahmitajjuddin9863 It will ..
Exactly the concept was cool not a big fan though
Best epusode in a long time! So much funny stuff Joe!
The ultimate way to keep AI on our side is the power of moving on. If you give yourself an ultimate goal and tell the AI to assist you. It most likely will
I'm 17 and created a learning AI, which teaches itself through evolution and won a landwide coding competition in Germany with it. I think i'm allowed to be proud of myself now.
Yeah now go to bed kid
@@jackritcher2291 Lul, that really happened though. Not sure about the kid part either. Btw I'm studying CS at the moment
The two big examples of Deep Blue and Watson are no real AIs. They are database-orientated search engines that simply are able to go through massive amounts of data at superhuman speeds. They are NOT self-aware artificial entities that can really go beyond playing Jeopardy (a game hardly suited for a real AI test). Honestly, most of this video is a joke.
Honestly, the question "how to make a real AI" is a joke.
Watson is still able to approximate an understanding of the questions asked. It may lack agency, but it's capabilities are starting to close in on humans.
Nikonikoneko There's actually no proof for that at all. Watson more likely exploits the limitations of the game (for a human player) itself with it's near-instant access to a rather huge database. I'd love to see Watson solve other problems than Jeopardy. The first version of Watson failed providing above 50% correct answers because it did not have access to hundreds of thousands of encyclopaedic entries (both offline and online (magazines, scientific journals etc.)). It's first version had to approximate the reliability of an answer based upon a more limited database and as a result was way more often giving incorrect answers. In my humble opinion the whole Jeopardy thing was mostly a marketing thing. The actual AI at play is fairly underwhelming. As a result it's really a bit shortsighted to think it's capabilities are closing in on human capacities. I'd say that's incorrect.
PHeMoX
My understanding was that IBM is developing systems based on Watson which should i.e. support technicians based on manuals. I don't claim that its an AGI (artificial general intelligence), but as far as I know it has far more applications than any other knowledge system. I'm not really up to date though.
Nikonikoneko I think you're largely correct. I do think it's useful technology and will have it's application, especially when further developed. As such I'm not really downplaying the importance of Watson and various other projects before it. I just do not agree it's actual artificial intelligence. Yet. It's relying way too much on databases of information to really be a more independent artificial intelligence (I guess AGI) and there is not enough learning involved. The Jeopardy experiment almost failed for a variety of reasons. Anyway, in my opinion it's not real AI when it's basic search routines work something like this ; 1. determine a search that makes sense based upon the question asked 2. take parts of the question and do a brute force search using all relevant database entries available. 3. estimate the relevance of the database entries that returned information 4. do a linguistic and logical check based on similar known questions and answers and 5. combine the results with the highest probable rated answer being the answer given by Watson. I'm not claiming this is exactly what happens, but ("Jeopardy") Watson should roughly work like this. It's essentially a really fast search engine. In other words, take away the database and there's virtually no intelligence left. The whole point of AI should be arriving at a point where the machine starts to learn, understanding the world more or less by it's own. Supplying access to a huge database is not quite that. Why not let a machine decide for itself which data to add to it's library of knowledge? Why not add routines that predict, deduct and confirm relevance accordingly?
I guess you'll understand that having access to virtually every answer by being able to search a very very big database is quite different from estimating or deducting a logical answer based upon actual intelligence and understanding. To put things in perspective, some of these AIs have long crossed the point where a human would have the same library of knowledge at their disposal. Well... wouldn't that be... cheating?
I Came Here How To Create An AI For My Game. Now I Am Hooked To This Channel
I have one way to avoid AI revolution: NEVER task them unspecific roles like protect humans at all cost or mantain peace in the world
If Elon Musk Says that AI is the biggest existential threat to humans, why is he building AIs constantly?
Epic How To: Become a rockstar
get the show off
get paid
John Kirk Buy a guitar, learn how to play a couple chords/scales, go up on stage, dance on stage, destroy your equipment, rinse and repeat. The more you do that, the cooler you look and the more you learn.
Szczypta Suchara its called all star not rockstar
read the lyrics to all star
I don't understand how any of this can scare people, the idea of A.I., and computers that learn, and that can actually pass as a human is amazing to me, even the idea of ai wanting to kill everyone doesn't scare me, because it just amazes me that a computer can learn and have those thoughts and desires, all *we* need to do is help A.I. have normal, healthy, sane thoughts and desires.
how to survive if someone breaks into your house?
Slam a beer bottle in their face while drunk.
Epic how to get home invaders drunk
sprukununuy Slam a beer bottle in their face while drunk.
"Mom! Get out! Wait... I don't live home anymore... Who the hell are y..." *sounds of getting beaten up and mugged*
+Haavi12 My GSD/Hovawart mix is guarding here.
That last part about an off switch is a really important point. Computerphile has a video titled "AI 'Stop Button' Problem" that describes the difficulties of making an off switch for an AI
EPIC HOW TO: Land an Aeroplane in an Emergency.
Awesome introduction! An offswitch is always handy xD
Good job on the animations :)
0:04 i feal personally attacked
Epic how to achieve world peace.
Notering Well in the end it will achieve world peace but the process won't be peaceful.
Only through death or oppression would true, lasting peace be achieved. As for me, give me liberty or give me death.
Hey Andy great animating in this episode
epic how to: survive the purge
Hold fast to your 2nd Amendment rights.
@@bohemianmonk6222 L O L
@@katoca81 why is that funny? The 2A defends ALL of the rest.
@@bohemianmonk6222 I know, just a funny response.
They should first of all change all the media, and create a lot of content about IA having to serve humans, then they hard program it into their memories , then they create those emergency buttons to reset them, switch them into defensive mode and attack other IA's that treat humans, or destroy them.
How to be a BODYGUARD! -Epic how to!
Have a gun and put yourself for hire on the internet, congratulations.
I’m not worried about ai doing creepy stuff, in fact the reason i want to create an ai is for it to basically re-enact FNAF.
This aged well
sick job on the animation, andy!
How to be a vigilante
play watchdogs
they already did it
@@tam_3413 and it's called how to be Batman
Atleast provide full details
Preach !
Hey Andy!
Great Jon animating!
I really Liked your energy and how cool you put the stuff out! 12/10
For anyone wondering, this is not a "how to" video. It's basically just a history lesson.
Where's Tony Stark when you need him
In heaven with all the other dead MCU characters, lol
good work Andy and all crew
I see so many reasons how robots won't take over the world 😂 pitting that aside, its very intriguing, I love artificial intelligence, and robot arts, even science.
nope, in only 6 years all of the things in this vid is now childs play, scary stuff
Marvelous bro ❤Love you ❤🤟
Respect you research 👏👏👏👏👏
Epic how to: build a fortune 500 company
The ai we are creating are getting stronger and smarter until eventually they become aware enough to realize that humans are a threat and decide to do something similar to terminator
Hello, I'm Connor, the android sent by CyberLife
this is not a "How to", this is a history lesson xD
I'm watching it now. Chatgpt is on. Haha
Haha
99% : explaining AI
1% : what you need to create an AI
And now I create an indestructible army, then take over the world
Mwwwwhahaha!
Hay Andy great job animating on this video
EpicHowTo can you make a episode on: How to survive an EMP!
dancingdragonite
EMPs don't effect humans, only electronics.
Faraday box, look that up
artistafrustrado I know about that.
none of your business We are dependent electronics in this day and age. Think about if all the electricity went out.
If all electronics went out there would be complete and total chaos. Also all the people with pace-makers and prosthetic limbs would die.
Though I came here to learn more about developing an AI system. Instead I was educated on all of the existing AI systems built by other people. Thanks AWE me.
epic how to lucid dream
Well i've done it before, never really had time to program robots since, but i suppose i should give it another shot
How to survive a plane crash
Great work Andy.
Epic how to: Open a BANK
Loveumore nah
Great animations Andy!
R.I.P Stephen Hawkins
That robot did not only convince 33% of the judges that he was human he also proved that they were also idiots.
I hope at some point that we'll have AI citizens who will coexist and live alongside humans.
Please tell me that you mean in harmony because that'd be nice
NEXT EPIC HOW TO : PLUG 100 EARPHONES IN ONE EAR !!
how to become a successful musical artist?
Hahah, this was fun, I was doing some research for my sophomore science fair project for a voice synthesis and even when I knew this video didn’t have what I was looking for, I couldn’t stop watching.
how to become a jedi/sith
Wow the animators out did them self this time 👌
Epic How To : Time Travel!!!!
please do this video, please
I'ts going to have a lot of einstein
Black fire it's impossible and most likely little to no research
Done
2000-2030 = Software Engineering, apps, websites
2030...=AI
How to clone a dinosaur
The true problem with Strong AI is the lack of ethics and morality. If you create a strong AI we could live with it peacefully... until it believed we are a threat and would snuff us out of existence. So the question becomes: how do you teach a machine what being a good sentient being means without firing back at you (I Robot)?
Do epic how to run away from home or survive in the hood
That Google AI is a savage😂
how to be a video game assassin
Play Assassins Creed.
He already made a video similar to that.
Nice animating Andy
This didn't teach me how to create anything
I get really excited when he said the computer predicted the future
Make an AI that could be a better president than Trump
Aydan Metsch look up cleverbot. There you go.
Aydan Metsch any tin can would be better than trump
Aydan Metsch thats easy, just get your pet pig and stuff him in the place of the president
My calculator (at least it gives accurate responses)
Hey Andy!Great job animating this video!
Epic How To: Make a separate channel for Epic How to
how to make America great again?
Simple: make America FREE again. Get the government out of people's lives and businesses. Then reform the educational system to actually EDUCATE people instead of just indoctrinating them. Teach people to think for themselves.
Joe, you make these videos worth watching lmao
detroit become human
hey Andy great job on animating
How to Run away from home
Simply run away
But then you'll be motivated to go back, that's not really the highest quality of instructions lol
Great job animating Andy!
How to become a dictator
robot, what's the meaning of life?
robot: **freaks out** I DON'T HAVE ETHICS!!!!
I hate how you only talk about AI from the western point of view and not add from Japan's Point of View where you can have a companion AI.. they even a dog AI right now
vanthursday cuz merica
vanthursday that's cool
maybe cause it's an american channel? ¯\_(ツ)_/¯