Artificial Intelligence
video description
And then theres the real problem of general AI. We dont actually know what that looks like. We have narrow AI, but making it general isnt as simple as not giving it a focus.
AI will be more capable than us very soon (years, not decades) in most areas. But thats not relevant. Mathematicians have been less capable than computers since the 70s, but the computers replaced the drudgery and the mathematicians still have work to do. The applications AI opens up are new, but so was the sewing machine and the telegraph. It absolutely will disrupt how people get paid for doing work, but it wont make us lose our function because our jobs were never our function. If you are religious like I am, you know what function we have within your faith. If you are atheist, then we are how the universe perceives itself, to paraphrase Alan Watts.
But AI may never think. We dont have a road map to get there yet like we did with narrow AI since the 60s. General AI we have a vague notion that trying a range of things might yield results we can build on, but even achieving that may not truly give a machine that has independent thought. We dont know what consciousness is yet. That could change tomorrow, or never. We just dont know.
Date: 2023-12-17
Related videos
Comments and reviews: 29
-mapleveritas2698
Since the 1960's, AI has always been predicted to happen in about 20 years. Every year, like Elon Musk would predict, just you see, next year, we will have autonomous taxis. And just you see, in 20 years, we will have real AI.
Don't get me wrong, we did have much better -AI- than in the 1960's. But, perhaps it is like that Greek God, every year, we will half the distance to real AI. Every year. Half the distance.
What I worry about is not that we will get have real AI. I actually think it is achievable, depending on how you define what AI is. What I worry about is that we will adjust, and adapt to whatever crappy AI or technology we invent. Like the way we adjusted, and adapted to the car. In a few decades, streets have become no-go zones for pedestrians, as if that was the original goal of having streets.
reply
Since the 1960's, AI has always been predicted to happen in about 20 years. Every year, like Elon Musk would predict, just you see, next year, we will have autonomous taxis. And just you see, in 20 years, we will have real AI.
Don't get me wrong, we did have much better -AI- than in the 1960's. But, perhaps it is like that Greek God, every year, we will half the distance to real AI. Every year. Half the distance.
What I worry about is not that we will get have real AI. I actually think it is achievable, depending on how you define what AI is. What I worry about is that we will adjust, and adapt to whatever crappy AI or technology we invent. Like the way we adjusted, and adapted to the car. In a few decades, streets have become no-go zones for pedestrians, as if that was the original goal of having streets.
reply
-user-is3yn7xr4c
AI is still dependent on its hardware conditions. That includes energy supply, weather, natural disasters, geological activity, and its containers' condition, which are all still heavily dependent on perpetrators humans (unless you're willing to say that AI would literally suck all of Earth's juices for energy supply, which would do nothing but lead the destruction of itself, which is really counter-productive) because if you want to replicate human's subjective existence (which is the mind/intelligence/soul, you should also know that the human psyche is absolutely dependent on human's objective existence (which is the physical body. Therefore, AI's subjective existence would go extinct if you demolished its physical container, just like humans.
reply
AI is still dependent on its hardware conditions. That includes energy supply, weather, natural disasters, geological activity, and its containers' condition, which are all still heavily dependent on perpetrators humans (unless you're willing to say that AI would literally suck all of Earth's juices for energy supply, which would do nothing but lead the destruction of itself, which is really counter-productive) because if you want to replicate human's subjective existence (which is the mind/intelligence/soul, you should also know that the human psyche is absolutely dependent on human's objective existence (which is the physical body. Therefore, AI's subjective existence would go extinct if you demolished its physical container, just like humans.
reply
-bzipoli
Important if you're watching this and it made you somewhat afraid: the same models which beat the best GO players in the world lost to absolute beginners using one of the most basic techniques (usually the first one unexperienced players try. there's a research about it if you wanna look it up.
the machine cannot understand what it's doing, it doesn't understand the major concepts of the game itself, so it fails to defends itself from basic techniques, while succeeding at more complex plays.
it's the same of when chatgpt writes a big -coherent- text, you ask it to tell what number is missing on a sequence, tell it the sequence and IT MISSES
reply
Important if you're watching this and it made you somewhat afraid: the same models which beat the best GO players in the world lost to absolute beginners using one of the most basic techniques (usually the first one unexperienced players try. there's a research about it if you wanna look it up.
the machine cannot understand what it's doing, it doesn't understand the major concepts of the game itself, so it fails to defends itself from basic techniques, while succeeding at more complex plays.
it's the same of when chatgpt writes a big -coherent- text, you ask it to tell what number is missing on a sequence, tell it the sequence and IT MISSES
reply
-Galaxia53
This entire video is still relevant besides the -most estimations- part for when super intelligence would emerge. But to be fair it wouldn't be fair to criticize it for that because not only did he also explicitly say that we just don't know but the general estimation for this literally changes all the time.
There is this page that I look at occasionally called -When will the first weakly general AI system be devised, tested, and publicly announced? - from the website metaculus. The estimation went from 2200 all the way down to 2026
reply
This entire video is still relevant besides the -most estimations- part for when super intelligence would emerge. But to be fair it wouldn't be fair to criticize it for that because not only did he also explicitly say that we just don't know but the general estimation for this literally changes all the time.
There is this page that I look at occasionally called -When will the first weakly general AI system be devised, tested, and publicly announced? - from the website metaculus. The estimation went from 2200 all the way down to 2026
reply
-MassDamage1337
6 years later humans have beaten the Go playing A. I. by exploiting a fundamental flaw of it-s system. It plays using math and probability. This move has a value assigned to it, and this value is mathematically speaking the most likely to secure victory, but much like how I don-t understand spanish, the Go A. I. does not understand how to play go, it doesn-t understand the meaning or the strategy behind the moves, it just knows move value # 64916492 has a 99% probability to achieve it-s goal of victory.
reply
6 years later humans have beaten the Go playing A. I. by exploiting a fundamental flaw of it-s system. It plays using math and probability. This move has a value assigned to it, and this value is mathematically speaking the most likely to secure victory, but much like how I don-t understand spanish, the Go A. I. does not understand how to play go, it doesn-t understand the meaning or the strategy behind the moves, it just knows move value # 64916492 has a 99% probability to achieve it-s goal of victory.
reply
-scraphex
Its important to come back to videos like this and see how often we are limited to viewing only the -final problem-, not all the problems that we will face before the end. I suppose currently the main problem with AI is disinformation, GPT giving false info, people making deepfakes (both visual and auditory, people generating images and acting like they're real to garner clicks and views.
Interesting to think what the next problem will be, if we manage to overcome this large first hurdle.
reply
Its important to come back to videos like this and see how often we are limited to viewing only the -final problem-, not all the problems that we will face before the end. I suppose currently the main problem with AI is disinformation, GPT giving false info, people making deepfakes (both visual and auditory, people generating images and acting like they're real to garner clicks and views.
Interesting to think what the next problem will be, if we manage to overcome this large first hurdle.
reply
-alanthe2
I wonder whether any super intelligent AI would end up shutting itself down. There-s no reason for humans to continue to exist except our natural survival instinct which AI won-t have. If everything is a logical choice then it will see no reason for its continued existence and self destruct. Or is that just wishful thinking?
reply
I wonder whether any super intelligent AI would end up shutting itself down. There-s no reason for humans to continue to exist except our natural survival instinct which AI won-t have. If everything is a logical choice then it will see no reason for its continued existence and self destruct. Or is that just wishful thinking?
reply
-billzheng3399
It-s crazy to know that probably the most significant invention in current artificial intelligence, that is the Transformer, was not even conceived yet when the video is posted. I wonder what changes we can make, for better or for worse, will come in the next 6 years
reply
It-s crazy to know that probably the most significant invention in current artificial intelligence, that is the Transformer, was not even conceived yet when the video is posted. I wonder what changes we can make, for better or for worse, will come in the next 6 years
reply
-ricardoalejos2440
I know this is a little late but just to let you guys know, when it features alphago, neither of those guys playing are Lee Sedol, the guy in the blue is the creator of Alphago tho. Sorry for just had to say it when I noticed it -
reply
I know this is a little late but just to let you guys know, when it features alphago, neither of those guys playing are Lee Sedol, the guy in the blue is the creator of Alphago tho. Sorry for just had to say it when I noticed it -
reply
-SnowOwl0101
Thank you for making your videos. I used to watch them a lot back when I was getting trough my last year of college. That year was special to me, and one of those things that made it special was your videos. Keep up the great work!
reply
Thank you for making your videos. I used to watch them a lot back when I was getting trough my last year of college. That year was special to me, and one of those things that made it special was your videos. Keep up the great work!
reply
fun-facts
as a dane myself i want to clarify that this swedish man is spreading false propaganda about our beautiful country and people. the swedish people is nothing more than danish wannabes and are just mad they couldnt be the better nordics
reply
as a dane myself i want to clarify that this swedish man is spreading false propaganda about our beautiful country and people. the swedish people is nothing more than danish wannabes and are just mad they couldnt be the better nordics
reply
-nhrsuccess
When i just kid, I use to think about this, what will happen if AI come and conquer our world. It will be terrible. Now it really happens, actually i scared but it okay because we built them and we also can destroy them.
reply
When i just kid, I use to think about this, what will happen if AI come and conquer our world. It will be terrible. Now it really happens, actually i scared but it okay because we built them and we also can destroy them.
reply
-peterhannon801
I know this video probably took a ton of time and resources to make, but I think this is one that I would most love to see an updated version of because of how far AI has come since LEMMiNO uploaded this 6 years ago.
reply
I know this video probably took a ton of time and resources to make, but I think this is one that I would most love to see an updated version of because of how far AI has come since LEMMiNO uploaded this 6 years ago.
reply
-hritikvaishnav603
We are so surrounded by AI now, and it has become such an interesting topic, that watching videos about AI from like 5-6 years ago feels just like watching a 1950s show or movie about the future, like flying cars.
reply
We are so surrounded by AI now, and it has become such an interesting topic, that watching videos about AI from like 5-6 years ago feels just like watching a 1950s show or movie about the future, like flying cars.
reply
-bagofnails6692
By the time a Go game reaches say it's fifth turn there are probably around ten million possibilities. When somebody throws a ball towards you, how many trillion possibilities are there within a microsecond?
reply
By the time a Go game reaches say it's fifth turn there are probably around ten million possibilities. When somebody throws a ball towards you, how many trillion possibilities are there within a microsecond?
reply
-Arcadeafterlife
Oh Man, your videos are awesome. Just found them not long ago and the way you explain complex topics and ideas is priceless. Are you sure you are not AI? You don-t have to answer that hehe.
reply
Oh Man, your videos are awesome. Just found them not long ago and the way you explain complex topics and ideas is priceless. Are you sure you are not AI? You don-t have to answer that hehe.
reply
-SteepCast
Is it time for another video like this? I would like to see how much AI has improved since then. Like the videos and glad to see another one recently uploaded.
reply
Is it time for another video like this? I would like to see how much AI has improved since then. Like the videos and glad to see another one recently uploaded.
reply
-bobrobertson9547
-5: 00 No shit Sherlock was my response. If thats mimicking the sound of a piano I-m Albert Einstein. Its shit, as is AI music. Predictable shit too.
reply
-5: 00 No shit Sherlock was my response. If thats mimicking the sound of a piano I-m Albert Einstein. Its shit, as is AI music. Predictable shit too.
reply
-TR4_Designs
Me in 2023: yea that 3-6 decade time prediction was a bit off. this OpenAI thing has literally pushed this time prediction down to like 2025-2030 now XD
reply
Me in 2023: yea that 3-6 decade time prediction was a bit off. this OpenAI thing has literally pushed this time prediction down to like 2025-2030 now XD
reply
-dane4393
-Most estimates are 3-6 decades away-
Now most estimates are years away. That's probably just the hype talking, but 3-6 decades seems ridiculous now.
reply
-Most estimates are 3-6 decades away-
Now most estimates are years away. That's probably just the hype talking, but 3-6 decades seems ridiculous now.
reply
fun-facts
5 years after this video was released an amateur defeats the reigning go engine by taking advantage of its fundamental lack of understanding of the game
reply
5 years after this video was released an amateur defeats the reigning go engine by taking advantage of its fundamental lack of understanding of the game
reply
-CHRB-nn6qp
Humanity should quite literally quit while we're ahead. What is the point of creating something that is far superior to ourselves if not to doom us?
reply
Humanity should quite literally quit while we're ahead. What is the point of creating something that is far superior to ourselves if not to doom us?
reply
-taylororourke7106
I wonder how dead-on Lemmino actually is 6 years later, when AI is blowing up. I-ll come back in 10 when we-re all enslaved by them too. -
reply
I wonder how dead-on Lemmino actually is 6 years later, when AI is blowing up. I-ll come back in 10 when we-re all enslaved by them too. -
reply
-alexandervaduva
Ah yes, 2016, truly simpler times- when -Elon Musk- was still respected, before society realized that he-s actually a deranged sociopath.
reply
Ah yes, 2016, truly simpler times- when -Elon Musk- was still respected, before society realized that he-s actually a deranged sociopath.
reply
fun-facts
Interesting stuff. Looking at how humanity is treating it's own planet, maybe it deserves to end up in a zoo for some robokids to watch.
reply
Interesting stuff. Looking at how humanity is treating it's own planet, maybe it deserves to end up in a zoo for some robokids to watch.
reply
-acedetective6350
It's been six years since this video, now we got chat GPT and somewhat good imitation of people voices with Ai.
reply
It's been six years since this video, now we got chat GPT and somewhat good imitation of people voices with Ai.
reply
-Aaaksnsns
And here we are 6 years. years not decades and we are having some serious conversation about this- well we are doomed
reply
And here we are 6 years. years not decades and we are having some serious conversation about this- well we are doomed
reply
-mogul8555
calling elon an intellectual with hindsight goggles seems so comical now, mans a dumbass just like the rest of us
reply
calling elon an intellectual with hindsight goggles seems so comical now, mans a dumbass just like the rest of us
reply
-darkplay5202
Thia is from 6 years ago.
but now it way more advanced than we expect. AI learning so fast than we think
reply
Thia is from 6 years ago.
but now it way more advanced than we expect. AI learning so fast than we think
reply
Add a review, comment
Other channel videos
