#Elon #Musks #Warning #Artificial #Intelligence
Talk about your time allocation I think one of the things you spend an awful lot of time uh thinking about I know is artificial intelligence it’s something that you and I have as a shared interest that it’s something that our audience is interested in as well
Um the question here is a lot of experts in AI don’t share the same level of concern that you do about the dangers fools what what’s this last word what’s what specifically do you believe that they don’t ah um well the biggest issue I see with so-called AI experts is that they they
Think they know more than they do um and I think they’re smarter than they actually are um in general we are all much smarter than we think we are but much less smart dumber than we think we are um by a lot so this is this tends to play plague smart people
Um they just can’t they’re defining themselves by their intelligence and they they don’t like the idea that a machine could be way smarter than them so they discount the idea which is fundamentally flawed That’s The Wishful Thinking situation um I’m really quite close to I’m very close
To The Cutting Edge in Ai and it scares the hell out of me um it’s capable of vastly more than almost anyone knows and the rate of improvement is exponential but you can see this in 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 to then beating the European world champion who was ranked 600 then beating Lisa doll 4-5 who had what been world champion for many years then beating the current world champion then beating everyone while playing simultaneously then then
There was Alpha zero uh which crushed Alpha go a hundred to zero and Alpha zero just learned by playing itself and it can play basically any game that you put the rules in for if you whatever rules you give it literally read the rules play the game and be superhuman for any game
Um nobody expected that rate of improvement if you ask 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 the batting average is quite weak it’s not good
Um we’ll see this also with uh with self-driving uh I think probably by end of next year self-driving will be will Encompass essentially all modes of driving and be at least 100 to 200 percent safer than a person by the end of next year we’re talking like maybe 18 months from now
Um a little study on on Tesla’s autopilot version one which is relatively primitive and found that it was a 45 reduction in highway accidents and that’s despite autopilot one being just version one um 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 but we have to figure out some way to ensure that the Advent of digital super intelligence is one which is symbiotic 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 that well I I’m not normally an advocate of Regulation and oversight I mean I think
Once you generally you’re on the side of minimizing those things but this is a case where you have a very serious danger to the public and therefore there needs to be a public body that has insight and then oversight on to confirm that everyone is developing AI safely um
Like this is extremely important I think danger of AI is much greater than the the danger of nuclear warheads by unlocked um and nobody would suggest that we allow anyone to just build 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’s the 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 we’ve all been focused in on the idea of artificial super intelligence right which is clearly a danger of maybe you know a little further out um what’s happened over the last year is you’ve seen artificial what I’ve been calling artificial stupidity you’ve been talking about you know algorithmic
Manipulation social media like we’re in it now it’s starting it’s starting to happen how do we how do we get is it what’s the intervention at this point um I’m not really all that worried about the short-term stuff things that are um like narrow AI is not a species level risk um
It it will result in dislocation uh in lost jobs and um you know that sort of better weaponry and that kind of thing but it is not a fundamental species level risk whereas uh digital super intelligence is yeah so it’s really all about laying the groundwork to make sure that
If if Humanity collectivity decides that creating digital super intelligence is the right move then we should do so very very carefully um very very carefully um this is the most important thing that we could possibly do yeah building on that other than AI and the the other issues that you’re
That 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 um well I mean there are other things that on a longer time scale the um and obviously the things that I believe in like extending life beyond
Earth making life multi-planetary um I know a big believer in sort of um asimov’s Foundation series or the principle that you you really want to um you know a recommend reading the foundation series but it’s like if you if you know that there’s a there’s likely to be we don’t
Know but there’s likely to be another Dark Ages which it seems my guess is there probably will be at some point um I’m not I’m not predicting that we’re about to enter Dark Ages but that there’s some probability that we will particularly if there’s a third world war um
Then we want to make sure that there’s enough of us of a seed of human civilization somewhere else to bring civilization back um and perhaps shorten the length of the Dark Ages um yeah I think that’s why it’s important to get a self-sustaining base ideally on Mars because Mars is far
Enough away from Earth that that a war on Earth the Mars base might survive it’s more like this survived than a moon base but I think a moon base and a Mars base um that um they could perhaps help regenerate life back here on Earth it would be really important and to get
That done before a possible world war three um you know last 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 there probably will be at some point or
If we have another one it’ll be the last yeah it just could be a radioactive Rubble you know um so again I’m not predicting it just seems like well if you say given enough time will it be most likely given enough time this is this is has been our pattern in the past
Yeah so um like I really believe in the zeroth law of asthma of zero flow you’re gonna take a set of actions most likely to support Humanity into the future um yeah but I think that’s a sustainable energy is also obviously really important that’s totological if it’s not sustainable it’s unsustainable yeah how
Close are we to solving that problem well I think that the core Technologies are are there with the wind solar um with batteries um the the fundamental problem is that there’s an unpriced externality in the cost of of CO2 um 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 puts carbon into the atmosphere which includes rockets by the
Way so I’m not excluding Rockets from this it has to be a price and that um you can start off with a low price but then that price and then depending upon whether that price has any effect on the positive 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 of fossil fuels from deep and under the Earth and putting it into the atmosphere we’re pretending that that had that that that has no probability of a bad outcome in
The entire scientific Community is saying obviously it has it’s going to have a bad outcome obviously you just you’re changing the chemical constituency the atmosphere so um 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 that’s already up there I actually think we can manage with the current carbon level or even a little bit higher um it’s um and this is going to sound like I’m backtracking but there’s actually an argument that
More common in the atmosphere is is actually good but up to a point so we might actually arguably have been a little carb installed if you go back 200 years ago and say okay well Fiona just go like you had like 280 290 pots per million of carbon we’re probably a little common
Stock now we’re about 400 just past 400 Mark I think somewhere in the 400s probably okay we don’t have to worry about sequestering carbon or anything like that but now if this momentum keeps going and we start going up to 600 800 a thousand fifteen hundred um that’s where things get really squirrely
And the sheer momentum of the world’s energy infrastructure is leading Us in that direction um it’s very so it’s just very important that the the public and the government’s push to to ensure the correct price of coven is paid um so that that will be the thing that that matters um
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 right I don’t know I feel like we know part of the answer to this because you were trapped in Westport for a while
Um uh but but how I mean on a regular day for you are you are you are you sleeping you’re not sleeping a lot right oh geez do I look that bad no um you look great but we just imagine with the amount of responsibilities with
The amount of 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 um that’s like always there
And uh and uh working really hard on Tesla Model 3 production um and uh look who made some good progress but it’s usually hard work but those are the two most stressful things in my life right now yeah um our audience really wants to know uh
We hope the world will look like for children’s 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 so I think the a good future would look like
You know we really substantially transferred to sustainable generation and consumption of electricity so that the the CO2 risk and the ocean Rising risk is mitigated um and we’re not looking at like you know having Florida and and sort of large portions of the world underwater that’d be great that not um
But to have addressed that risk that’ll be kind of enormous for us to have a base on the Moon based on Raj to be out there exploring the solar system it’s our building industry on essentially having human civilization go out there and and and have to be such that anyone can
Go uh to Bruno Mars or out to the solar system if they want to make it really affordable I do think it’s important that this competition that there are multiple companies doing this not just SpaceX and um and that AI risk is that I guess it’s the sort of a benign
Ai and that we’re able to achieve a symbiosis with that AI um ideally the AI there’s somebody who I can’t remember his name but had a good suggestion for what the optimization of the AI should be what’s its utility function you have to be careful about this because if you say maximize happiness
And the AI concludes that happiness is functioned opening and serotonin so it captures all humans and Jacks your brain with logical amounts of dopamines that are turning like okay it’s not what we meant it’s pretty good though oh you’ll love it um well I like the definition of like
The eyes should try to maximize the freedom of action of humanity um maximize the freedom of action maximize Freedom essentially um I like that definition um but we do want a close coupling between Collective human intelligence and digital intelligence um a neuralink is trying to help in that regard by
Um creating a an interface between the high bandwidth interface between Ai and your and human brain um yeah we’re already we’re already a cyborg in the sense that your phone and your computer are kind of extension of you um just low bandwidth input output exactly it’s just low bandwidth
Particularly output I mean two thumbs basically um so how do we solve that problem the boundary thing the bandwidth District I mean we’ve all we’ve all succumbed to it now we’re all we’re all cyborgs we’re just low efficiency side works so how do we how do we make it better
I think we’ve got both a we’ve got to build an interface um like we didn’t evolve to have a Communications Jack um you know or some so there’s got to be essentially vast numbers of of tiny electrodes that are able to read write for your brain of course you know security is
Pretty important in the situation say the least um I was gonna say 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 um but it’s a bit like in Banks is neural lace but not but in the case in your
Life it’s sort of that that’s there for when you’re born or it’s sort of it’s not a yeah kind of a backup this would be there’s there’s a digital extension of you uh that is an AI the AI extension review a tertiary layer of intelligence um so you’ve got your limbic system your
Cortex and and the tertiary layer which is the digital AI extension of you and that high bandwidth connection is what um achieves the types of symbiosis right I think that’s the best outcome I I hope so if anybody’s got better ideas I’d love to hear it um and um
Talk about another project that you’re working on that our audience wants to know a little bit more about Starlight oh can you tell us anything um doing Skynet [Laughter] hopefully not kinda it’s internet in the sky um well we um we’re not talking much about starlink but essentially it’s intended to provide
Low latency high bandwidth internet connectivity throughout the world um that actually will not be enough cognitive processing power on board the satellite system to to uh in any way be a Skynet thing like it’s uh digital AI requires a lot of super intelligence requires a lot of big
Servers on the ground it’s too power intensive but this is intended to be to provide people with um who don’t have any internet connectivity with the internet connectivity and it’s very good for spas and it was possibly populated and moderately specific populated areas and for giving people in cities
A choice of a low-cost choice of internet access but I do think it’s going to be important the stalling system will be important in providing the funding necessary for SpaceX to develop interplanetary spacecraft um and at the same time yeah helping people who have either no or super expensive connectivity and giving people
In urban areas more of a competitive choice very cool um I have to ask her because it’s the number one question uh just going back to Mars what kind of government do you envision for the first Martian colony um what’s your and what’s your title yeah yeah exactly Emperor or God Emperor I
Don’t know it might be too much I don’t know um not everyone gets irony you know remember I must remember um so I think the the I think most likely the the form of government on Mars would be somewhat of a direct democracy where um you vote on issues where people vote
Directly on issues instead of going through representative government in in you know when the United States was formed representative government was the only thing that was logistically feasible because people there’s no way there was no people to communicate instantly a lot of people didn’t even have really access to uh
Mailboxes or it wasn’t even really the post office was very very primitive a lot of people couldn’t write um so you had to have some form of representative democracy or things just wouldn’t work at all um but I think I might as most likely it’s going to be people Everyone votes
On every issue and that’s how it goes there are a few things I’d recommend which is keep laws short long laws it’s like that’s that’s something suspicious is going on if there’s lung law you know if you’re if the size of the law exceeds the word
Count of Lord of the Rings it sounds foreign which it does amazingly then it’s like something’s wrong um so there should be a limit to the size of the law that it should be able to digest it like how come you can read the Constitution and all of the Amendments like you can
Read those in maybe an hour and and we govern so much of our Civilization by that and yet modern law is a obtuse super boring term that’s indesirable to almost anyone so I think direct democracy lowers laws that are comprehensible um I think having some kind of hysteresis on
Like it should be easier to remove a law than create one because things just get too inertia you have to have something that’s going to overcome inertia so probably I don’t know what the right number would be maybe it’s like 60 40 maybe it requires 60 to get a low in
Place but any number above 40 can remove a low um otherwise you just get laws just accumulate over time accumulative time and it it’s sort of like gullivery where you just get trapped by all these tiny strings and you can’t move you get hardening of the arteries of
Civilization with with rules and rules and rules rules so it should be just easier to get rid of a rule than to put one in maybe they should even have like a some kind of sunset close so that they just automatically expire unless there’s enough of an impetus to keep them around um
I know I know there’s a fair amount of interest I’m interested in hearing a little bit more about the very early days with Tesla and how it came together brother Kimball is here I thought we’d bring him out you guys could talk a little bit about it
0 Comments