Elon Musk - Artificial Intelligence, Neuralink & New Forms of Government on Mars

talk about your time allocation I think one of the things you spend an awful lot of time thinking about I know is artificial intelligence it's something that you and I have a shared interest that it's something that our audience is interested in as well the question here is a lot of experts in AI don't share the same level of concern did you do about the day traffic rules what what's his last words what's what specifically do you believe that they don't well the biggest issue I see with so-called AI experts is that they they think they know more than they do and they think they're smarter than they actually are in general we are all much smarter than we think we are much less smart dumber than we think you are by a lot so this is this tends to play plague smart people that you just can't that they define themselves by their intelligence and they they don't like the idea that a machine can be way smarter than them so they discount the idea which is fundamentally flawed that's the wishful thinking situation I'm really quite close to I'm very close to the cutting edge in AI and it scares the hell out of me it's capable of vastly more than almost anyone knows and the rate of improvement is exponential you can see that some things like alphago which went from in the span of maybe six to nine months it went from being unable to beat even a reasonably good go player so then beating the European world champion who was ranked 600 then beating Lisa doll for five what would be in a world champion for many years then beating the current world champion then beating everyone while playing simultaneously then then there was alpha zero which crushed alphago 100 to 0 and alpha zero just learnt by playing itself and it can play basically any game that you put the rules in for if you whatever rules you give it it literally read the rules play the game every superhuman for any game nobody expected that great of improvement to guess those so those same experts who think AI is not progressing at the rate that I'm saying I think you'll find that their predictions for things like go and and other and other AI advancements have therefore their batting average is quite weak it's not good that we'll see this also with was self-driving i I think probably by anammox year the self-driving will be well encompass essentially all modes driving and be at least a hundred to two hundred percent safer than a person by the end of next year we're talking maybe eighteen months from now and it's a third study on on Tesla's autopilot version one which is relatively primitive and found that it was a forty five percent reduction in highway accidents and that's despite what a pilot one being just version one version two I think will be at least two or three times better that's the current version that's running right now so the rate of improvement is really dramatic we have to figure out some way to ensure that the advent of digital super intelligence is one which is semiotics with humanity I think that's the single biggest existential crisis that we face and the most pressing one and how do we do that I mean if we take it that it's inevitable at this point that some version of AI is coming down the line how do we how do we steer through them well I'm not normally an advocate of regulation and oversight I mean I think it's once you generally go inside minimizing those things but this is a case where you have a very serious danger to the public and it's therefore there needs to be a public body that has insight and then oversight on to confirm that everyone is developing AI safely this is extremely important I think a danger of AI is much greater than the danger of nuclear warheads landlocked and nobody would suggest that we allow anyone to just blow up nuclear warheads if they want that would be insane and mark my words AI is far more dangerous than nukes far so why do we have no regulatory oversight this is insane what question you've been asking for a long time I think it's a question that's come to the forefront over the last year where you begin to realize that it doesn't necessarily I think if we've all been focused in on the idea of artificial superintelligence right which is clearly a danger but maybe you know a little further out what's happened over the last years you've seen artificial what I've been calling artificial stupidity talking about you know algorithmic manipulation of social media like we're in it now it's starting it's starting to happen how do we how do we is it what's the intervention at this point so I'm not really all that worried about the short-term stuff things that are other like narrow AI is not a species level risk it will it will result in dislocation in lost jobs and you know that sort of better weaponry and that kind of thing but it is not a fundamental species level risk whereas digital super intelligence is so it's really all about laying the groundwork to make sure that if if humanity collectively your science that creating digital super intelligence is the right move then we should do so very very carefully very very carefully this is the most important thing that we could possibly do building on that other other than AI and the the other issues that you're you're tackling transportation energy production aerospace what issues should our next generation of leaders be focused on solving what else is coming down the line well I mean there there are other things that are on a longer time scale the obviously the things that I believe in like extending life beyond Earth making life multiplanetary but I'm a big believer in a sort of Asimov's Foundation series or the principle that you you really want to you know I recommend reading the foundation series but it's like if you if you know that there's a there's likely to be you know you don't know but there's likely to be another Dark Ages which it seems my guess is the father will be at some point I'm not on a predicting that we're about to enter a Dark Ages but that there's some probability that we will particularly if there's a third world war then we wanna make sure that there's enough of us of a seed of human civilization somewhere else to bring civilization back and perhaps shorten the length of the dark ages I think that's why it's um it's important to get a self-sustaining base ideally on Mars because Mars is far enough away from Earth that occur that a warrant earth the Mars base might survive is more likely to survive than a moon base but I think a moon base and a Mars base that that could perhaps help regenerate life back here on earth it would be really important and to get that done before a possible World War 3 you know last century we had two massive world wars three if you count the Cold War I think it's unlikely that we'll never have another world war again the parlor will be at some point or if we have another one it'll be the last you Madhvi radioactive rubble yeah so again I'm not predicting this seems like well if you say given enough time will it be most likely given enough time this is this is this is has been our pattern in the past yeah so like we really believe in the zeroth law of Asimov zeroth law you're gonna take the set of actions most likely to support the humanity of the future but I think that sustainability is also obviously really important that's tautological if it's not sustainable its unsustainable yeah how close are we to solving that problem well I think that the core technologies are are there with the wind solar with with batteries the the fundamental problem is that there's an unprecedented the cost of of co2 the market economics works very well if things are priced correctly but when there's when things are not priced correctly and something that has has a real cost has zero cost then that's where you get distortions in the market that inhibit the progress of of other technologies so essentially anything that that produces carbon it will push push carbon into the atmosphere which includes rockets by the way so I'm not excluding rockets from this there has to be a price and that you start off with a low price but then that price and and then depending upon whether that price has any effect on the past a million possibility of co2 in the atmosphere you can adjust that price up or down but in the absence of a price we sort of pretend that digging trillions of tons of fossil fuels from deep under the earth and putting it into the atmosphere pretending that that had that that that has no probability of a bad outcome and the entire scientific community is saying obviously it has it's gonna have a bad obviously you just you're changing the chemical constituency the atmosphere so so it's really up to people and governments to put to put a price on on carbon and and then automatically the right thing happens it's really straightforward what do we do with the carbon surrender I actually think we can manage with the current carbon level or even a little bit higher it's and this is gonna sound sound like I'm backtracking but there's actually an argument that more carbon in the atmosphere is is actually good but up to a point so we might actually arguably have been a little carbon starved if you go back 200 years ago and say okay well Furies go like you had like two hundred eight hundred ninety parts per million of carbon we're probably a little cough and stuff now we're about four hundred just passed four hundred mark I think somewhere in the 400s probably okay we'd have to worry about sequestering carbon or anything like that but not if those momentum keeps going and we stopped going up to six hundred eight hundred a thousand fifteen hundred that's where things get really squirrelly and the sheer momentum of the world's energy infrastructure is leading us in that direction so it's just very important that the the public and the governor's push to ensure the correct price of carbon is paid so that that will be the thing that matters yeah our audience is very interested in knowing how many hours of sleep you got last night I don't know about six five or six I think all right I feel like we know part of the answer this cuz you were trapped in west wall for a while [Music] but but how do I mean on a regular day for you or you are you are you sleepy you're not sleeping a lot right well geez do I look that bad huh you look great okay just imagine with the amount of responsibilities we think about you know with what you've got going on do these problems still keep you up at night or do you think we're on our way to solving well right now the only things that are really stressing me out in a big way are AI obviously that's like boys there and and I'm working really hard on Tesla Model 3 production and we're making good progress but it's usually hard work but those are the two most restful things my life right now yeah our audience really wants to know we hope the world will look like four children born today when they're your age wins right what do you hope for the world to look like what's the best-case scenario say we solve these problems what's that world look like let's see I think the a good picture would look like you know we're really substantially transferred to sustainable generation and consumption of electricity so that the the co2 risk and the ocean rising risk is mitigated and we're not looking at like you know having Florida and it's sort of large portions the world underwater that'd be great that not but to have addressed that risk that'll be there you know was for us to have a base on the moon face a mirage big out there exploring the solar system so welding industry on it essentially having you know civilization to go out there and and have you such that anyone can go the Moon or Mars or this whole system if they want making it really affordable I do think it's important that this competition of their multiple companies doing this not just a sex and and a I risk is that I guess it's the sort of a benign AI and they were able to achieve a symbiosis with that AI ideally the AI there's somebody who came over his name but had a good a suggestion for what the optimization of the AI should be what's its utility function you have to be careful about this because you say maximize happiness and the AG concludes that happiness is function of dopamine and serotonin so she captures all humans and jacks your brain with large amounts of dopamine serotonin like ok so he meant it sounds pretty good though I love it well I like the definition of like the I should try to maximize the freedom of action of you of humanity maximize the freedom of action maximize freedom essentially I like that definition but we do want to close coupling between collective human intelligence and digital intelligence and a neural link is trying to help in that regard by creating a an interface between a high bandwidth interface between AI and your and human brain there were already we're already a sidewalk in the sense that your phone and your computer a kind of extension of you just low bandwidth input-output exactly it's just low bandwidth particularly output I mean two thumbs basically so how do we solve that problem giver the bandwidth it's a bandwidth issue but we've all we've also come to it now we're all we're all cyborgs we're just low efficiency cyborgs so how do we how do we make it better I think we've gotta both a we've got apples in her face like we didn't evolve to have a communications Jack you know some so there's gotta be essentially vast numbers of of tiny electrodes that are able to read right from your brain of course you know security is pretty important in the situation say the least obviously saying I'm not coming with I'm keeping my brain air-gapped yeah well I think a lot of people will choose to do that but it's a bit like Ian banks is in your lace but not made it but in the case of your lace it sort of felt that's there from when you're born or it's sort of it's not a it's working so tobacco yeah kind of a backup this would be this there's a digital extension of you that is an AI the AI extension of you a tertiary layer of intelligence so you've got your limbic system your cortex and and the tertiary which is the digital AI extension of you and that high bandwidth connection is what achieves the tights and biases I think that's the best outcome I I hope so if you know he's got better ideas and pull up to here and you talking about another project that you're working on that her audience wants to know a little bit more about Starling oh can you tell us anything doing Skynet hopefully not Skynet its Internet in the sky well we we don't talk that much about StarLink but essentially it's intended to provide low latency high bandwidth internet connectivity throughout the world that actually will not be enough cognitive processing car onboard the satellite system to to in any way be a Skynet thing like it's the digital AI requires a lot of super intelligence requires lot of big servers on the ground to power intensity but this is attempt to be to provide people with who don't have any internet connectivity with Internet connectivity and it's very good for sparse and responsive operated in moderate moderate least boss equated populated areas and forgiving people in cities a choice of a in your low-cost choice of internet access but I do think it's gonna be important the stalling system will be important in providing the funding necessary for SpaceX to develop interplanetary spacecraft and at the same time yeah helping people who have even though or super expensive connectivity and giving people in urban areas more of a competitive choice very cool I have to ask you because it's the number one question just going back to Mars what kind of government do you envision for the first Martian colony and what's your and what's your title yeah yeah exactly Emperor oh God Emperor I don't know it might be too much if you were to watch my jokes yeah not everyone gets irony you know I remember so I think the I think most likely the the form of government on Mars would be somewhat of a direct democracy where you vote on issues where people vote directly on issues instead of going through representative government in in when the United States was formed representative government was the only thing that was logistically feasible because people there's no way it was all your people to communicate instantly a lot of people's didn't even have really access to mail boxes well there wasn't even the post office was very primitive a lot of people couldn't write so you had to have some form of representative democracy or things just wouldn't work at all but I think I'm most likely it's gonna be people at everyone votes on every issue and that's how it goes a few things I'd recommend which is keep Lois shirt long Lois it's like that's that's something suspicious is going on if there's long Bowl you know if you if the size of law exceeds the word count of Lord of the Rings which it does amazingly then it's like something's wrong so there should be a limit to the size of the law that I should be able to digest it like how come you can read the Constitution and all of the amendments like you can read those and maybe an hour and and we haven't so much of a civilization by that and yet modern law is this obtuse super boring tome that's indecipherable to almost anyone so I think direct democracy lowers lows that are comprehensible I think having some kind of hysteresis on like it should be easier to remove a law than create one because things just get to inertia you have to have something that's gonna combination so probably I don't want the right number of you maybe it's like 60/40 maybe you don't quite a 60% to get a law in place but any number above 40% can remove a little otherwise things that you just get lowers just accumulate over time they cure their time and it's sort of like Gulliver where you just get trapped by all these tiny strings and can't move you get hardening of the arteries of civilization with law with rules and rules rules rules so it should be just easier to get rid of rule then let's put one in maybe it should even have like a some kind of sunset clause so that they just automatically expire unless there's enough of an impetus to keep them around I know I know there's an affirmative which is I'm interested in hearing a little bit more about the very early days with tassel on how it came together brother Kimble is here so we bring him out you guys could talk a little bit about it

Channel: The Artificial Intelligence Channel

Views: 583,174

Keywords: singularity, ai, artificial intelligence, deep learning, machine learning, deepmind, robots, robotics, self-driving cars, driverless cars, Elon Musk, Neuralink, SpaceX, Falcon Heavy, Mars, Tesla

Id: 6tBZA2rygcM

Length: 23min 36sec (1416 seconds)

Published: Mon Mar 12 2018