Such an insightful video. The difference between being smart and being a genius is the ability to generate new abstractions. Dr Wolfram expands the how we evaluate the universe. A true treasure.
So it all boils down to 'choice'. We have the choice to add 'greebling'. Stunning interview. An incredible man and a joy to listen too. Thank you Stephen Wolfram and thank god I 'chose' to whatch the whole thing.
Great talk ..Life is about to change in many ways, medical diagnosis for one, I would say if we could get that right up to 90 to 100 % correct , the cost to society would be reduced to over 30% maybe 50... On top of that, automated diagnosis and care, which will be 24 hours a day, every day of the year, no holidays or sickness, unlimited Diagnosis, no shortage of Doctors or nurses ......Autonomous transport, massive reduction in accidents, less people going to hospital, people traveling any hour of the day and night, less traffic jams ..sleep as you travel, no stress or read, maybe even yoga and meditation, never late ...Automated food cultivation.. living walls with air cleaning plants... recycled water, aquaponics .. EV ,,Solar panels ..power walls ...The whole education system needs modernising ..everyone medically monitored, new ways to treat, and detect, using skin colour changes, smell detectors many times more sensitive than any animal ..etc
How would you prove that anything is minimal in achieving a purpose? I do agree that these questions about purpose and goals are extremely important and central to most if not all of the "big questions."
I always wondered - Can we build better compression using automata - We search the program space for a particular program that after some steps produces our data. The program + initial state + number of steps must be less than data in size. But if program+state is all data, then we could recursively decrease our compressed data size. So some kind of proto-programs exist and some kind of base entropy which we can't go any further must exist. PRNG appear to generate 'random numbers' ie lots of entropy but are described pretty easily -- PPMd and similar can partially compress serial output of LCRNGs but not down to their basic structure. So in some sense compression algorithms do search program space in some sense with limited time and memory.
Very interesting talk. A lot of what he says and predicts seems very reasonable but I disagree with something he said and the very beginning, that there is a broad equivalence between the types of computation that different systemes (brains, computers, weather) do. Computers do computation, brains can sort of do computations but it's not how they operate at a very basic level and weather doesn't do computation. Mathematics, computation , can be used to model and describe brains and weather systems but that is not the same as those systems computing. Things do not happen according the laws of nature, the laws describe how things happen. When things happen that aren't in the laws of nature, we change the laws of nature. The laws of nature and rules of computation are human inventions. There is machine intelligence and there's human intelligence, they are different sorts of things because they arise from different underlying mechanisms. I wouldn't all either of them artificial, they are either intelligent or not. I agree with him that computers will take over the world not because it's their goal but because people are lazy. Uber is very big in my city. Uber drivers drive all over yet they don't learn the roads because the google maps tell them where to go. I like ask them if they know which street is some major thoroughfare is and often they don't. We've automated many process at work and now people just click buttons and don't even pay attention to information on the screen. Give humans an autopilot and they will use it. I am skeptical of his prediction that people will start to learn programming, or human language will evolve towards a programming language. Some terminology will leak in from programming languages has computers become more and more a part of the culture. Most people won't learn computer languages even though it might be useful to do so. Mathematics has been around for a long time and is very useful but people still don't learn it. Why would computer programming be any different? Humans will continue to make more and more intelligent computers, they will even use computers to aid in that task. But I don't see computers running that process of their own accord. Brains are not computational machines, they are very complex chemical soup with structure. The things that give us purpose and motivates us are emotions, hormones, feelings. We might be able to make a computer that mimics those but mimicking emotions is not the same as having emotions. There's no reason in principle we couldn't make a device that had human like emotions but the more human like the emotions the more brain like the device will be. Think about it, what would be the point of a machine with it's own motivations? Wouldn't that make it unreliable? We couldn't count on it to do what we wanted it to do. It be rather foolish to build such a machine and then give it control over us. That's not to say in some distant future humans won't do it, we are rather foolish.
I timed it just right to be a C engineering student because they have largely automated the parts of the questions you have to understand in order to get an A. Or at least that's what I tell myself.
"We have to try to find a sort of symbolic representation of things we have traditionally expressed in natural language" (40:36) - We don't need to "try", we have already "tried": Formal Semantics. A subfield of linguistics and philosophy. Frege. Russel. Montague. Kripke.
Thanks Dr. Wolfram for keeping me humble, for letting me know I'm illiterate. I wrote a couple of programs in Fortran back in 1975 (on punch cards), but that only makes me a dinosaur. :)
Consciousness is intelligence. Intelligence is computation. Clouds are computation. Rocks have computation going on in their atoms. Everything in the Universe is computation. Everything is conscious but thinks fundamentally so differently to us that we mostly don't realise it. Panpsychism is true. en.wikipedia.org/wiki/Panpsychism
@@qualiacomposite well consciousness is kind of a muddy word to use for a idea that's purely an abstraction of human perception. Technically nothing and everything is "conscious" because the word implies unfalsifiable concept... "We are conscious because we are aware of ourselves...rocks are not conscious because they are not aware of themselves."...except how do you know that the rock isn't conscious if you are not the rock? If you asked me personally, I think everything is conscious and exhibits intelligence to some degree so long as it exists as part of a complex system. The level of "intelligence" of a bacteria is of course no where near the intelligence of a human being because bacteria are no where near as complex. But even inanimate objects, so long as they exist in a complex system and are doing computation, and are at some level "alive" and we see this with the genesis of life...inanimate proteins connecting together to make DNA and so on...these things we would not consider traditionally as conscious...but these proteins are existing in a complex system and self assemble to create our consciousness eventually as the system becomes more complex. That's really the beauty of cellular automata's...it's watching inanimate objects come to life.
We become what we make. Manchine. Man and machine intertwined. It is clear we will all be part machine in the not too distant future. The issue of machine interfacing will be central to our nervous system. Currently we think of AGI as an external future machine technology - something that will be mostly beneficial to mankind. But what happens one day when we replace our humanity with intelligent parts - we become immortal. We are what we make.
Looks like someone added too much contrast and the “blacks got crushed”. I.e. under his collar, under his chin, his eyebrows, .....in his mouth, his eyes....no detail in the black parts of this image.
As AI is imitating with human brain, we should believe ourselves to move forward, develop, and collaborate with AI. It is very ambivalent between global language which makes us easier to communicate and variety of languages led to different cultures respectively. As we are trapped economically and emotionally, we have put ourselves to the square.
"Loneliness, that's the bottom line. I was never happy as a child...Christmas Ted, what does that mean to you? It was living hell. Do you know what it's like falling in the mud and getting kicked, in the head. With an iron boot? Of course you don't, no one does, that never happens. Sorry Ted, that's a dumb question, skip that...Municipal bonds Ted, I'm talking double A rating. The best investment in America." Sorry, Stephen, I couldn't resist 'cause you're rambling a bit. But I did find the "Airplane!" quote quickly, thanks to machine learning, a better term for artificial intelligence.
9:54 Using code to construct human contracts will never work. It has been tried in the past by many philosophers, through the ages, to try and mathematically represent the human law system. But it always never works because there is always a degree of obscurity that can only be resolved by consulting human intellect. It is impossible for machines to establish the fairness to human arguments and contracts. Only humans can establish such fairness. Even so, codes have bugs!
Laws and contracts written in natural language have bugs, too. They are called loopholes and they are exploited in the same way as bugs in software are exploited. However, if you write contracts in code, you could perform things like automated testing, regression testing and so on.
It's surprising how similar I think to how Wolfram thinks. The whole paradigm of thinking this way may sometimes feel too mechanistic, and make one lonely though, so I liked watching Paul Budnik's video poems: th-cam.com/channels/dePJnbPvilEEytRLAH2wfQ.html -- that inspire about the possibilities of the diversity that can be.
"paint a picture really fast" ..."this is what I'm taking about..." - Ironic to think human communication methods are the bottle neck :) .... "hey google: turn down the volume" + "HEY GOOGLE TURN DOWN TH....." = Me walking across the room
First off: I don't pretend to understand much of this, but, I don't trust this guy. (Also, I don't know much philosophy so a lot of this rant is from intuition.) I don't trust AI when its architects begin to discuss how it will interact with free will. It appears that Wolfram is utterly amoral - he has actually said that he detests the word "consciousness". His philosophy essentially states that if we follow the universal program, it will eventually result in consciousness. My worldview is the opposite and rests on consciousness being primary. His ideas are frightening because I think the philosophy that simple rules and programs can eventually describe the entire universe is accurate. Reminds me of Chomsky describing the language and the mind - how a very finite lexicon following common and simple rules can be used to express an infinite range of thoughts. When Wolfram says "harness ... for human purposes" I have to ask which "humans" he is referring to. My sense is that he is referring to the elites. How the ruling class decides to implement AI to replace bureaucracy could lead, in short order, to a "Terminator", "Brave New World", or "I, Robot" scenario.
goals: break through the stupid barriers to cognitive be expansion. get rid of the dea. get rid of any sections in the federal bureau of indecency that exist to persecute citizens. fire all of them no retirement. send them to Siberia to did in the Pete mosd. this statement is not nearly as silly as it will seem to you. all apologies.
He is almost entirely incorrect in his analysis...he seems to think that you can symbolically represent objectives when the converse is true...Ask yourself...How do I know I like ice-cream?
Such an insightful video. The difference between being smart and being a genius is the ability to generate new abstractions. Dr Wolfram expands the how we evaluate the universe. A true treasure.
Agree, I like his Turing test analogy computers generate a visual image which is a much richer experience/ higher bandwidth than speech brilliant.
@@TheReferrer72 066*---===
Incredible to be able to talk so coherently during such a long time about such complex matter. Genius
What a delight to listen to Dr. Wolfram speak.
Thank you for the upload.
So it all boils down to 'choice'. We have the choice to add 'greebling'. Stunning interview. An incredible man and a joy to listen too. Thank you Stephen Wolfram and thank god I 'chose' to whatch the whole thing.
Thanks for being a voice of sanity on AI risk, Stephen. Always nice to hear your views on knowledge and technology as well.
Great talk ..Life is about to change in many ways, medical diagnosis for one, I would say if we could get that right up to 90 to 100 % correct , the cost to society would be reduced to over 30% maybe 50... On top of that, automated diagnosis and care, which will be 24 hours a day, every day of the year, no holidays or sickness, unlimited Diagnosis, no shortage of Doctors or nurses ......Autonomous transport, massive reduction in accidents, less people going to hospital, people traveling any hour of the day and night, less traffic jams ..sleep as you travel, no stress or read, maybe even yoga and meditation, never late ...Automated food cultivation.. living walls with air cleaning plants... recycled water, aquaponics .. EV ,,Solar panels ..power walls ...The whole education system needs modernising ..everyone medically monitored, new ways to treat, and detect, using skin colour changes, smell detectors many times more sensitive than any animal ..etc
How would you prove that anything is minimal in achieving a purpose? I do agree that these questions about purpose and goals are extremely important and central to most if not all of the "big questions."
To me, this is the most valuable video!!!
I always wondered - Can we build better compression using automata - We search the program space for a particular program that after some steps produces our data. The program + initial state + number of steps must be less than data in size. But if program+state is all data, then we could recursively decrease our compressed data size. So some kind of proto-programs exist and some kind of base entropy which we can't go any further must exist. PRNG appear to generate 'random numbers' ie lots of entropy but are described pretty easily -- PPMd and similar can partially compress serial output of LCRNGs but not down to their basic structure. So in some sense compression algorithms do search program space in some sense with limited time and memory.
the best scientist in the world. i am almost sure that this is an uncontestable opinion.
Great exploration of natural language merging with computer languages. This would be far more insightful if he included some examples.
Very interesting talk.
A lot of what he says and predicts seems very reasonable but I disagree with something he said and the very beginning, that there is a broad equivalence between the types of computation that different systemes (brains, computers, weather) do. Computers do computation, brains can sort of do computations but it's not how they operate at a very basic level and weather doesn't do computation. Mathematics, computation , can be used to model and describe brains and weather systems but that is not the same as those systems computing. Things do not happen according the laws of nature, the laws describe how things happen. When things happen that aren't in the laws of nature, we change the laws of nature. The laws of nature and rules of computation are human inventions.
There is machine intelligence and there's human intelligence, they are different sorts of things because they arise from different underlying mechanisms. I wouldn't all either of them artificial, they are either intelligent or not.
I agree with him that computers will take over the world not because it's their goal but because people are lazy. Uber is very big in my city. Uber drivers drive all over yet they don't learn the roads because the google maps tell them where to go. I like ask them if they know which street is some major thoroughfare is and often they don't. We've automated many process at work and now people just click buttons and don't even pay attention to information on the screen. Give humans an autopilot and they will use it.
I am skeptical of his prediction that people will start to learn programming, or human language will evolve towards a programming language. Some terminology will leak in from programming languages has computers become more and more a part of the culture. Most people won't learn computer languages even though it might be useful to do so. Mathematics has been around for a long time and is very useful but people still don't learn it. Why would computer programming be any different?
Humans will continue to make more and more intelligent computers, they will even use computers to aid in that task. But I don't see computers running that process of their own accord. Brains are not computational machines, they are very complex chemical soup with structure. The things that give us purpose and motivates us are emotions, hormones, feelings. We might be able to make a computer that mimics those but mimicking emotions is not the same as having emotions. There's no reason in principle we couldn't make a device that had human like emotions but the more human like the emotions the more brain like the device will be.
Think about it, what would be the point of a machine with it's own motivations? Wouldn't that make it unreliable? We couldn't count on it to do what we wanted it to do. It be rather foolish to build such a machine and then give it control over us. That's not to say in some distant future humans won't do it, we are rather foolish.
I timed it just right to be a C engineering student because they have largely automated the parts of the questions you have to understand in order to get an A. Or at least that's what I tell myself.
"We have to try to find a sort of symbolic representation of things we have traditionally expressed in natural language" (40:36) - We don't need to "try", we have already "tried": Formal Semantics. A subfield of linguistics and philosophy. Frege. Russel. Montague. Kripke.
Thanks Dr. Wolfram for keeping me humble, for letting me know I'm illiterate. I wrote a couple of programs in Fortran back in 1975 (on punch cards), but that only makes me a dinosaur. :)
humans are beautiful let them prosper
exactly my thoughts
how energetically costly is this matrix?
Consciousness is intelligence. Intelligence is computation. Clouds are computation. Rocks have computation going on in their atoms. Everything in the Universe is computation. Everything is conscious but thinks fundamentally so differently to us that we mostly don't realise it. Panpsychism is true. en.wikipedia.org/wiki/Panpsychism
Consciousness is a kind of computation. I don't think it follows that all computations are conscious.
@@qualiacomposite well consciousness is kind of a muddy word to use for a idea that's purely an abstraction of human perception. Technically nothing and everything is "conscious" because the word implies unfalsifiable concept... "We are conscious because we are aware of ourselves...rocks are not conscious because they are not aware of themselves."...except how do you know that the rock isn't conscious if you are not the rock?
If you asked me personally, I think everything is conscious and exhibits intelligence to some degree so long as it exists as part of a complex system. The level of "intelligence" of a bacteria is of course no where near the intelligence of a human being because bacteria are no where near as complex. But even inanimate objects, so long as they exist in a complex system and are doing computation, and are at some level "alive" and we see this with the genesis of life...inanimate proteins connecting together to make DNA and so on...these things we would not consider traditionally as conscious...but these proteins are existing in a complex system and self assemble to create our consciousness eventually as the system becomes more complex. That's really the beauty of cellular automata's...it's watching inanimate objects come to life.
We become what we make. Manchine. Man and machine intertwined. It is clear we will all be part machine in the not too distant future. The issue of machine interfacing will be central to our nervous system. Currently we think of AGI as an external future machine technology - something that will be mostly beneficial to mankind. But what happens one day when we replace our humanity with intelligent parts - we become immortal. We are what we make.
Is it just me or did someone give him some joke black dye candy before he recorded this video?
Looks like someone added too much contrast and the “blacks got crushed”. I.e. under his collar, under his chin, his eyebrows, .....in his mouth, his eyes....no detail in the black parts of this image.
As AI is imitating with human brain, we should believe ourselves to move forward, develop, and collaborate with AI. It is very ambivalent between global language which makes us easier to communicate and variety of languages led to different cultures respectively. As we are trapped economically and emotionally, we have put ourselves to the square.
What's wrong about everyone playing video games all the time? Beats war, famine, politics, etc.
Sedentary sedimentary fat layers. Enjoy lethargy.
@@JulesBartow I play video games and I hit the gym everyday. Video games are AWESOME!
"Loneliness, that's the bottom line. I was never happy as a child...Christmas Ted, what does that mean to you? It was living hell. Do you know what it's like falling in the mud and getting kicked, in the head. With an iron boot? Of course you don't, no one does, that never happens. Sorry Ted, that's a dumb question, skip that...Municipal bonds Ted, I'm talking double A rating. The best investment in America." Sorry, Stephen, I couldn't resist 'cause you're rambling a bit. But I did find the "Airplane!" quote quickly, thanks to machine learning, a better term for artificial intelligence.
human prejudices is built into the A.I. that may not be All Good in the long run .
9:54 Using code to construct human contracts will never work. It has been tried in the past by many philosophers, through the ages, to try and mathematically represent the human law system. But it always never works because there is always a degree of obscurity that can only be resolved by consulting human intellect. It is impossible for machines to establish the fairness to human arguments and contracts. Only humans can establish such fairness.
Even so, codes have bugs!
Laws and contracts written in natural language have bugs, too. They are called loopholes and they are exploited in the same way as bugs in software are exploited. However, if you write contracts in code, you could perform things like automated testing, regression testing and so on.
It's surprising how similar I think to how Wolfram thinks. The whole paradigm of thinking this way may sometimes feel too mechanistic, and make one lonely though, so I liked watching Paul Budnik's video poems: th-cam.com/channels/dePJnbPvilEEytRLAH2wfQ.html -- that inspire about the possibilities of the diversity that can be.
"paint a picture really fast" ..."this is what I'm taking about..." - Ironic to think human communication methods are the bottle neck :) .... "hey google: turn down the volume" + "HEY GOOGLE TURN DOWN TH....." = Me walking across the room
First off: I don't pretend to understand much of this, but, I don't trust this guy. (Also, I don't know much philosophy so a lot of this rant is from intuition.) I don't trust AI when its architects begin to discuss how it will interact with free will. It appears that Wolfram is utterly amoral - he has actually said that he detests the word "consciousness". His philosophy essentially states that if we follow the universal program, it will eventually result in consciousness. My worldview is the opposite and rests on consciousness being primary. His ideas are frightening because I think the philosophy that simple rules and programs can eventually describe the entire universe is accurate. Reminds me of Chomsky describing the language and the mind - how a very finite lexicon following common and simple rules can be used to express an infinite range of thoughts. When Wolfram says "harness ... for human purposes" I have to ask which "humans" he is referring to. My sense is that he is referring to the elites. How the ruling class decides to implement AI to replace bureaucracy could lead, in short order, to a "Terminator", "Brave New World", or "I, Robot" scenario.
Huh? Disembodied representation of the world.
So to speak
Machine machine...
goals: break through the stupid barriers to cognitive be expansion. get rid of the dea. get rid of any sections in the federal bureau of indecency that exist to persecute citizens. fire all of them no retirement. send them to Siberia to did in the Pete mosd. this statement is not nearly as silly as it will seem to you. all apologies.
So, Who or what have we to kill to get free?
Could a human other than you respond to most of your email?
He is almost entirely incorrect in his analysis...he seems to think that you can symbolically represent objectives when the converse is true...Ask yourself...How do I know I like ice-cream?
The talk is too long!
a comment
lol u bald