This was an insightful discussion on various aspects of Tesla Autopilot that I hope catalyzes further nuanced conversation on the future of AI-assisted driving. Starts at 2:35. The full outline of the video is as follows: 0:00 Introduction 2:35 Start of conversation: Autopilot motivation 4:01 Display the vehicle's perception of the driving scene 7:11 Algorithms, data, and hardware development 10:23 Edge cases and common cases in driving 12:18 Navigate on Autopilot 13:57 Hardware and software path toward fully autonomy 17:08 Driver supervision of Autopilot 20:13 Human side of Tesla Autopilot (driver functional vigilance) 23:13 Driver monitoring 24:30 Operational design domain 26:57 Securing Autopilot against adversarial machine learning 28:29 Narrow AI and artificial general intelligence 30:10 Physics view of love 31:53 First question for an artificial general intelligence system
It was a clever reply in a pseudo clever kind of way, and Elon was obviously enjoying the humor of it, but I don't think it was a particularly serious answer. I'm not sure it was a very serious question either. The implication in the question, and perhaps a better question might be "Do you think an AGI would have any better insights into any questions than people do now?"
I love how Elon always completely listens to the other person whom he is talking to and then thinks before answering.Nowadays very less people do it and I have started to practice that.
@@gsh64 more than politeness, though, because a lot of people might listen to most of what people say, but they will still try to finish people's sentences, and then respond without taking time to consider what they heard. Im as guilty as anyone else, but it's something im working on
Lex kept asking the same questions with different sentenses. Not a very good interview imo. Also Musk seemed a bit irritated for having to repeat himself.
Elon has talked abou base reality and simulation theory numerous times. I found the interview was simply recycling old topics in a somewhat fresh way...
@@tamie341 I think there's something appealing about this interview. It's kind of an unfiltered conversation between analytical minds and I like it. We've seen Elon smoking weed and laugh, this is good too
@@Tyraell01 it's good to see him converse with someone who can engage with him on his kinda level. Probably my favourite Elon interview I've ever seen was with a guy who runs a TH-cam channel (everyday astronaut or something like that I think?), at the unveiling of the starship prototype a while back. He asked Elon about the choice of rocket engine design, the pros and cons and why it was chosen over a more conventional type. It's definitely not my area of expertise so it went way over my head, but Elon's (almost childlike) enthusiasm was awesome to see, and some of the thoughts he shared really inspired me and stuck in my brain.
I was little shocked to see that a child of Alex Fridman (kind of obvious Jewish name) could have a child named Alexey Fedotov then intealised during Soviet times it was pretty common to name your child this way to protect them from antisemitism. He doesn't look Jewish so it might have worked. Here's his own tweet - A healthy mix of Russian soul, Jewish wit, and American guts makes for a good soup recipe, or so I heard.
I really enjoyed this interview, finally somebody who know enough about AI to talk with Elon about autopilot. And not some random journalist who ask always the same bonehead questions.
I've heard that Zuckerberg said that Musk doesn't understand AI. I'm not saying either person is right or wrong but does anyone know why he said that or what their difference of opinion is?
His answer about adversarial examples shows he does not have much of a clue himself what neural networks are. His solution to adversarial examples certainly is nothing of the sort. The problem with adversarial input isn't a lack of negative examples during training. That said the problem of adversarial input is overstated; finding adversarial examples requires exact knowledge of the parameters of the network in question, and the ability to feed it exactly controlled fake input. If you have that level of access to a self-driving car you might as well adversarially turn the wheel around and be done with it; why bother going to the trouble of misleading the cameras?
Probably the best ending to any TH-cam video i've ever seen. Well done Lex, I couldn't have asked for a more insightful, and surprisingly motivational dialogue.
Wow, just WOW! Can you imagine the CEO of ANY other company having this sort of discussion? Meanwhile, the rest of the industry fires off all these platitudes about how they're going to take down Tesla and seem to expect Tesla to quit innovating and just wait for them to catch up. The fact that Tesla would have to go outside the established computer industry and develop their own computer to accomplish the requirements of Full Self Driving speaks volumes as to where the rest of the industry is OR should I say, is not.
The problem is that the majority of other companies are money/greed driven. Tesla is driven by the passion of someone with a childhood in scifi and the ruthless efficiency of a prodigy with an immeasurable IQ. You can't judge Tesla as as a regular company, it's not.
When Musk got involved with Tesla he realized the two key technologies for the company were batteries and autonomy. Main stream automakers couldn't see that. They are miles behind on both technologies. They are in a seriously no-win situation. ICE cars will go extinct. But who would want to buy their electric cars?
@@alilwittiness You mean like Hyundai cutting production of the Kona, or the Audi E-tron's "Real World Range" coming in at a paltry 204 miles? Or the mythical solid state battery that is always just months around the corner? Or Nissan's supreme battery management technology after about a decade in the EV field? Or Toyota's release of outdated EV patents whilst they concentrate on Hydrogen Fuel Cells? Oh, I forgot the holy grail of Graphene Technology. Sorry! Perhaps I am not "Fully Versed" in what is going on in the rest of the industry but I do believe Sandy Munroe has a pretty good idea of what is going on as it relates to Tesla, and especially in China. When FCA has to beg Tesla to take their millions of dollars to bail them out of their own shortsightedness, it just further re-iterates the dilemma Legacy Manufacturers are in. I will grant you that VW group spits out lots of interesting concepts but they need to actually deliver something. Whatever the future holds, there is LOTS of room for multiple players in the EV space. At least ELON and team have kick started the momentum.
Wow Sir your comment is extracting the exact reason why Tesla is going to rule ! The reason is their CEO, completely different and not profit ruled but innovation driven.
I've never seen Elon so relaxed and authentic during an interview I have a new respect for him and Lex as well, what an ideal interviewer this was an excellent experience on so many levels
It was probably that Elon knows that Lex is at a certain level of understanding and intelligence, so he doesn't need to try to explain the things that he would normally be required to. Plus Lex asks great questions.
@C0mmand 0ption his method works to get some of the most interesting guests on. He gets on my nerves often but it’s often my best (and only) option to hear hours of real intellectuals. Maybe he’s just plain enough
My favorite part in this interview...When Elon Musk was asked by Lex, “When you or someone else creates an AGI (artificial general intelligence) system, and you get to ask her one question, what would that question be?” After a 1 minute pause, where he’s deeply thinking, he finally responds, “I’d ask, ‘what’s outside the simulation’”?
A system can never know what’s outside the simulation because it’s only working with data it’s been fed. “In order to find that which is immeasurable, measurement must first come to an end”
Very very very informative. Nothing compared to trashed pumped out by CNBC, FOX, Etc. for economic benefits. Thank you both for such powerful interview which helps general public have a glance at the future of AI and full self-driving.
Thomas Edison approach. Create the desire for something that doesn't exist, promise it does exist, then hire people to create it in your name, sell it. Repeat step 1. He's said before Edison is one of his heroes and models.
So thats why the Techpriests of Mars worship Toasters! It all comes down to Elon making a divine toaster and bringing it to Mars. This kickstarted the Stellar Exodus and very much later the Cult Mechanicus...
I often think when people interview Elon : "Stop wasting his time, he's got work to do!" But this is one of the rare cases where his time was well used. Well done Lex
But how's this for a mind bender... what if we were to assume that everyone had the potential to achieve great things like Elon, but if you want to get there then there's "work to do"? What if the person whose time is being wasted the most... ...is _you?_
@@joshnic6639 I thought the "other guy" did a really great job. He clearly knew what he was talking about, asked great questions, and really brought the best out of Elon.
@@joshnic6639 you should check who is Lex. You will then appreciate better that his questions were spot on. Also, Elon Musk will never take 30 minutes of his precious time to talk to some random guys. Elon Musk likes to be challenged by smart people. Lex is one of them.
Say what you may, Elon Musk is a genius! And Lex asked very probing and very informed questions. This interview makes me want to buy Tesla stock right now!!
Feralz right but are those two things that related? The energy crisis from dwindling fossil fuels is millions of times more a threat than traffic fatalities.
@@robertblake1228 i used to think like that but elon is 100% right in the fact that in the future people will say we were crazy for driving cars , i actually think about that sometimes when im driving fast on the highway or anywhere else, just a tiny shift of my hands of 10 degree can make me crash and possibly kill me or others around me.
Fascinating how Elon's brilliance is so focused while concurrently immensely flexible. When he's thinking, you can practically see his mind and imagination working together to weight out and explore different outcomes. Inspiring watching how passionate and immersed he is in the journey itself, not just arriving the destinations. Thank you, Lex!
Its like elon said before in the interview: the better the AI is, the more you can engage the autopilot. with the automatic lane change there will be a huge increase in time spent with autopilot engaged. therefore you get a lot of more data from these critical moments, thereby vastly improving the learning curve of the AI. this in turn alows you to engage autopilot even in more tricky situations and creating even more useful data. the better the AI, the more you can use it and the more you use it, the better the data -> better AI.
@@Stan_144 I think he was often like that during his interview with Joe Rogan as well; he seems quite reserved about what he'll say. Not sure about the "disclaimer" at the beginning as well. And, at the time I made that statement, I really was roughly at the 2 min mark, so yea, my expectations were higher.
@@brando3342 I think it would be very difficult to come up with an honest true answer to this question in the few seconds provided. Of course Elon would have to use some humor and take reference from the previous topic to get through that question so quickly. However, you are undoubtedly spewing the "good news" and want us to all accept the undeniable fact that God walks among us, testing our ever-lasting souls for the attributes he deems acceptable in his white billowy country club he runs in the after-life. No Jews, Muslims, Hindu, or atheist need apply, God tosses their applications straight away.
@@brando3342 I'm not trying to make fun of anyone. To answer directly to "Not their fault and you don't know who's going to heaven, neither do I." I would say, you don't even know if there is a heaven to go to.. You were born into Christianity. I was also born into Christianity, but have discovered reasonable doubt. "It isn't a "simulation". This is Satan's world, the delusion God sent. Ignore it, go find your family and live happily." is a quote from you on a seperate thread in this video's comments. You seem pretty confident in your knowledge of god, and his intentions. I admit I have jumped to conclusions based on your other thread, and it was insensitive and lazy to assume to know the meaning of "Elon is not being honest." What did you mean when you said Elon was not being honest in his answer?
Nice to see Elon get interviewed by someone with the ability to challenge him with real probing questions. It's interesting to Elon struggle with his desire to speak frankly to a fellow engineer against his obligation to speak as the CEO of Tesla.
I feel like After listening to many other Elons interview I have noticed that he will have the answer prepared but always respond like he is struggling.
I love the way Elon thinks. Things are just so obvious and automatic to him but they're revolutionary to the rest of us. His ability to see the future crystal clearly shows he's a true visionary.
31:30 "there may be ways to test if it's a simulation, but you can imagine that once an entity in the simulation found a way to detect the simulation, it can do something that corrects for that error" 31:52 "so when someone creates an AGI system and you get to ask her one question, what would that question be?" 32:16 "what's outside the simulation?" 32:20 Simulation: "time to end this podcast"
Lex is so pure. Until now he's still himself. How time flies. This is the first Lex interview I've ever seen. Keep up the good work Elon & Lex! I love you guys! 😘
@@joseph1541 Assuming for a moment that hardware processing power wouldn't be a bottleneck, a true AGI would be able to self-improve exponentially and therefore its intelligence would be effectively infinite. Now, whether its knowledge is effectively infinite is a whole other question that's dependent on the data it has access to. But if it's supplied sufficient relevant data, it should be able to answer any question. Of course, this is all hypothetical science fiction currently, as we don't really know yet if true AGI is actually possible.
A simulation is an attempt to replicate the universe. The universe can't be a simulation you fools. By definition. Can it be more than you ever imagine? Well, you bet your ass it can. But that doesn't mean any of it is fake, when, obviously, it's the realest thing you can perceive at this time of your existence
@@chibula I think this question is maybe more important than Isaac one, I mean If there is no entropy outside the simulation, clearly is not the last question. Elon question is not the last one neither, because, maybe there is another universe in which our simulation is running, but there is more simulations? which one was the first one? is a cycle? it has a end or a beginning?. Still, Elon's question approaches this concept more
The difference of comfort from 1st appearance to round 4 is nearly opposite ends of the spectrum. It's neat to see these two have conversations and how they evolve over time each appearance.
A friendly suggestion if I may, Pull the camera back just a little, bring it up slightly to drop your eye level, and leave some head room (space between the top of your head and the frame of the video) May be useful for reducing the sense of being imposing and general claustrophobic vibe. Just a suggestion based on general observation. You might have technical constraints I'm not aware of or what not, and I'm not trying to tell you how to run your show. Just minor tweaks that may yield improvements. Cheers.
serious bro whats up with this weak intro.. lets get into it already he still talkin about stuff we already been knowing since the JRE podcast. alrighty then its about to start... oh wait its not hes still talking to himself where is ellon ???
@@MealTeam6 If by "dumb" you mean people who don't take themselves too seriously and know when to let go, then I agree. Pretentious cunts need not apply.
My instant respect for this person - Lex Friedman - have never seen him before but it’s instantly obvious that the guy is brilliant. Very sharp and insightful discussion, really enjoyed it, thank you.
@@tamie341 my god! Just dont go on comparing everyone to elon and say the're nothing! Elon NATURALLY had a reading habit and had a computer since his childhood. Hkw many people have that NATURALLY? The thing is you're fucking jealous of Lex Fridman, so degrading him by comparing to elon. Get some better job to do.
He's right. The amount of drunks, druggies, retards, idiots, imbeciles and women operating this machinery, is unbelievable. Drivers licence should have much more strict conditions.
@@after_midnight9592I Wasn't going to like the comment because it didn't had to be said, it's so obvious. But then i read women and i had to like it :D
I just imagine God sitting in his heavenly throne, one foot resting on the opposite leg's knee. Chin in his palm, looking bored. Then Elon says that and God doesn't move for a second, then just gives a 🤨 "hmm" and lightly taps the glass dome behind which he sits, more in idle thought than for any structural testing purpose. On the other side of the dome, the tapping causes an enormous tsunami to strike the coast of Japan, killing 14.
What if the notion of outside is only pertinent to the simulation? Also, why do people insist on placing human attributes upon the divine? Isn't that rather arrogant?
As an enginerd myself, I found this video extremely impactful. There is such great deal of insight shared between these two. I will have to watch it a few times to completely understand everything introduced.
@@mike_o7874 for real if I had to choose someone to end it all....I would choose Elon over Mark Suckerberg. Something tells me he would actually take care of us while we die or give us a honorable death, Mark on the other hand would probably find the cheapest of all and make it as sufferable as possible.
@@mike_o7874 You being ok if Elon turned out to be evil all along, simply because you admire him, is pretty much how the most evil dictators in history were allowed to thrive. One of the best quotes I've heard came from the anime villain, Aizen Sousuke: "Admiration is the emotion furthest from comprehenshion."
I loved this podcast. The conversation that started from self-driving cars went all the way to the point of existential crisis in just 30 minutes. The part where they talk about the statistics required to prove to the regulators that self-driving cars are safer than manual cars sounds like that we need a number of deaths and incidents to happen, which of course will happen naturally as we slowly advance towards the point where it is needless to argue about whether autonomous vehicles are safer than human-driven cars. Even at that point, I think we won't be able to let go the fear of being not in control, but this conversation also seems to suggest that our only logical response in that situation would be to admit that we are (or will be) so far more limited compared to the AIs. Too little we can do for being so much of control freaks. No wonder why Elon in that regard is working on Neuralink, or BCI in general, as that is one thing that can bridge the gap between us humans and the machines. Must say thanks to Lex for organizing and sharing such a conversation with Elon.
one issue I have is that just because the AI can outdrive the masses statistically does not mean it can outdrive me with me fully paying attention and all my intuition and reflexes on full capacity. It would not come close to that. All it is doing if it statistically beats the masses is it is proving that it can outdrive granny, teens texting, and people who got distracted at the pretty lady jogging on the sidewalk. Suddenly, the AI "appears" to be a better driver statistically due to these edge cases, but it is not better than someone who is on their A game. That is why I'd struggle to ever trust it with my life because I truly will always be able to outdrive it if I'm at my best.
28:20 Lex has some very good questions that seem to be washed over. Tricking neural nets using adversarial attacks on objects outside it's own environment is surprisingly easy.
That was a very professional interview. At a different level than the other youtube interviews that I've seen so far. I really enjoyed it. Thank you and Elon as well.
You can detect a different level of intelligence here. This MIT kid has a great brain for the workhorse technical stuff he did with his team. Elon, however, has that along with unrivaled common sense and practicality. A powerful combination to say the least.
I actually think the Lex's work on driver/passenger monitoring could be very assuring when teslas are in a robo taxi mode. Instead of it detecting mobile phone use, guitar playing and water drinking it could be detecting and reporting vomiting, littering and vandalism to the owner and Tesla.
Just the idea that the car could circle/drive around/park a mile away, etc etc while you shop, go to the Dr, etc, but arrives when you call it, is game-changing. Think of all the time we've wasted just looking for a space...all that time is no longer lost....Eventually, self-driving cars will be the new ubers and taxis...with no drivers at all....unless legislation precludes it, which is possible....Eventually, we will get completely away from personal ownership of cars..Right now, the major hurdle for auto driving cars is all the cars out there being driven by humans....If all cars were self-driving/autonomous, the accident rate would plummet...
Elons expression as he is thinking of an answer to lex's last question reminds me of doctor strange going through all the different realities and considering all the outcomes.
23:13 25:57 I agree with Musk here. If "automation" is defined by “ letting a machine direct _itself_ all by _itself_ ”, then driver monitoring is actively AGAINST automation.
This was an insightful discussion on various aspects of Tesla Autopilot that I hope catalyzes further nuanced conversation on the future of AI-assisted driving. Starts at 2:35. The full outline of the video is as follows:
0:00 Introduction
2:35 Start of conversation: Autopilot motivation
4:01 Display the vehicle's perception of the driving scene
7:11 Algorithms, data, and hardware development
10:23 Edge cases and common cases in driving
12:18 Navigate on Autopilot
13:57 Hardware and software path toward fully autonomy
17:08 Driver supervision of Autopilot
20:13 Human side of Tesla Autopilot (driver functional vigilance)
23:13 Driver monitoring
24:30 Operational design domain
26:57 Securing Autopilot against adversarial machine learning
28:29 Narrow AI and artificial general intelligence
30:10 Physics view of love
31:53 First question for an artificial general intelligence system
It's kinda hard that you have to list such disclaimers. Good job Mr Lex
@LEX Fridman, you are the man!!!!!!
Thank you! Amazing insights!
Thank you very much for bringing in Elon Musk!!!
Bring karpathy and the other heads, hardware, engineering
"What's outside the simulation...?"
Legendary
That freaked me out
It was a clever reply in a pseudo clever kind of way, and Elon was obviously enjoying the humor of it, but I don't think it was a particularly serious answer. I'm not sure it was a very serious question either. The implication in the question, and perhaps a better question might be "Do you think an AGI would have any better insights into any questions than people do now?"
@@nicholas1460 No, it was a poignant question to pose. Much better than your attempt.
Nathan Connolly especially the long pause right before
You can literally see his brain trying to calculate the answer.
After watching one of the top AI researchers interview one of the top industrialists, it's really hard to watch politicians open their mouths.
@@FearlessP4P1 Reminds me of the snake from the Jungle Book.
@Epstein Didntkillhimself in the US they don't.
wats the video??
Lex has quit MIT to build AI of his own, so I would consider him as having entered the industry as well.
@Mister G I think that some pols are smart but that many do reside somewhere along the sociopathic spectrum.
I love how Elon always completely listens to the other person whom he is talking to and then thinks before answering.Nowadays very less people do it and I have started to practice that.
He is a polite gentelmen, that's why.
@@gsh64 more than politeness, though, because a lot of people might listen to most of what people say, but they will still try to finish people's sentences, and then respond without taking time to consider what they heard. Im as guilty as anyone else, but it's something im working on
Awesome.
take note Neil deGrasse Tyson
That's a sign of intelligence and maturity. So, we can only conclude that there are mostly dumb and childish people these days.
The density of high quality questions and answers in this podcast is mindblowing. Good job, Lex.
Lex kept asking the same questions with different sentenses. Not a very good interview imo.
Also Musk seemed a bit irritated for having to repeat himself.
Have you seen the video
Elon Musk really pisses off Lex Fridman
It’s hilarious! 😡 😂 😆
@@eugenecrawley4897 if he would be pissed off - why did he made it three times into this channel?
@@thuering6229 easy to ask questions within hindsight
Best ending to an Elon interview that I’ve ever seen. Great question final question.
it was incredible!
Elon is an superb being
Elon has talked abou base reality and simulation theory numerous times. I found the interview was simply recycling old topics in a somewhat fresh way...
BEST ENDING I EVER SEEN IN MY LIFE
you can see his eyes are moving from side to side, like he was processing the question... very rare moment.
Emotions were definitely in another realm in this podcast.
Not a great interviewer imo. Elon opens up a lot more for a good host
laughing so hard right now!
@@tamie341 I think there's something appealing about this interview. It's kind of an unfiltered conversation between analytical minds and I like it. We've seen Elon smoking weed and laugh, this is good too
@@Tyraell01 it's good to see him converse with someone who can engage with him on his kinda level. Probably my favourite Elon interview I've ever seen was with a guy who runs a TH-cam channel (everyday astronaut or something like that I think?), at the unveiling of the starship prototype a while back. He asked Elon about the choice of rocket engine design, the pros and cons and why it was chosen over a more conventional type. It's definitely not my area of expertise so it went way over my head, but Elon's (almost childlike) enthusiasm was awesome to see, and some of the thoughts he shared really inspired me and stuck in my brain.
Yeaah they are 🤖
Never thought I'd live long enough to see Star Wars meet Star Trek.
Discovery against the Enterprise???
When Russian Gangsta meets South African Gangsta, emotions have no space
Ben Shapiro would be fully erect if he watched this.
@@ClericChris 🤣 🤣
I was little shocked to see that a child of Alex Fridman (kind of obvious Jewish name) could have a child named Alexey Fedotov then intealised during Soviet times it was pretty common to name your child this way to protect them from antisemitism. He doesn't look Jewish so it might have worked.
Here's his own tweet -
A healthy mix of Russian soul, Jewish wit, and American guts makes for a good soup recipe, or so I heard.
Exactly my thought!
There was one chuckle and that too seemed forced :-)
💀😆
I really enjoyed this interview, finally somebody who know enough about AI to talk with Elon about autopilot. And not some random journalist who ask always the same bonehead questions.
I've heard that Zuckerberg said that Musk doesn't understand AI. I'm not saying either person is right or wrong but does anyone know why he said that or what their difference of opinion is?
"People have no idea what a neural net is, they probably think its a fishing net."
-Elon Musk 2019
Papa Goose he’s right
Or maybe mosquito net!!!
bunch of proportional gains that make their own complex formulas by trial and error adjusting
His answer about adversarial examples shows he does not have much of a clue himself what neural networks are.
His solution to adversarial examples certainly is nothing of the sort. The problem with adversarial input isn't a lack of negative examples during training.
That said the problem of adversarial input is overstated; finding adversarial examples requires exact knowledge of the parameters of the network in question, and the ability to feed it exactly controlled fake input. If you have that level of access to a self-driving car you might as well adversarially turn the wheel around and be done with it; why bother going to the trouble of misleading the cameras?
@@eelcohoogendoorn8044 I wouldn't count on him not knowing exactly what a neural net is.
Haven't seen a better ending to the podcast.
What are you... a 4 year old? What are you doing here my man? 😂😂😂😂 I am glad people your age are showing interest.
Probably the best ending to any TH-cam video i've ever seen. Well done Lex, I couldn't have asked for a more insightful, and surprisingly motivational dialogue.
Wow, just WOW! Can you imagine the CEO of ANY other company having this sort of discussion?
Meanwhile, the rest of the industry fires off all these platitudes about how they're going to take down Tesla and seem to expect Tesla to quit innovating and just wait for them to catch up.
The fact that Tesla would have to go outside the established computer industry and develop their own computer to accomplish the requirements of Full Self Driving speaks volumes as to where the rest of the industry is OR should I say, is not.
The problem is that the majority of other companies are money/greed driven. Tesla is driven by the passion of someone with a childhood in scifi and the ruthless efficiency of a prodigy with an immeasurable IQ. You can't judge Tesla as as a regular company, it's not.
When Musk got involved with Tesla he realized the two key technologies for the company were batteries and autonomy. Main stream automakers couldn't see that. They are miles behind on both technologies. They are in a seriously no-win situation. ICE cars will go extinct. But who would want to buy their electric cars?
You show you are not well versed in what others are doing in the field.
@@alilwittiness You mean like Hyundai cutting production of the Kona, or the Audi E-tron's "Real World Range" coming in at a paltry 204 miles? Or the mythical solid state battery that is always just months around the corner? Or Nissan's supreme battery management technology after about a decade in the EV field? Or Toyota's release of outdated EV patents whilst they concentrate on Hydrogen Fuel Cells? Oh, I forgot the holy grail of Graphene Technology. Sorry!
Perhaps I am not "Fully Versed" in what is going on in the rest of the industry but I do believe Sandy Munroe has a pretty good idea of what is going on as it relates to Tesla, and especially in China. When FCA has to beg Tesla to take their millions of dollars to bail them out of their own shortsightedness, it just further re-iterates the dilemma Legacy Manufacturers are in.
I will grant you that VW group spits out lots of interesting concepts but they need to actually deliver something.
Whatever the future holds, there is LOTS of room for multiple players in the EV space. At least ELON and team have kick started the momentum.
Wow Sir your comment is extracting the exact reason why Tesla is going to rule ! The reason is their CEO, completely different and not profit ruled but innovation driven.
This is far one of the best interviews ever held by anyone with Elon Musk, great job Lex 👍
I couldn't agree more. I may have to listen to this a few more times ot savor every morsel.
I've never seen Elon so relaxed and authentic during an interview
I have a new respect for him
and Lex as well, what an ideal interviewer
this was an excellent experience on so many levels
It was probably that Elon knows that Lex is at a certain level of understanding and intelligence, so he doesn't need to try to explain the things that he would normally be required to. Plus Lex asks great questions.
I think Elon respects interviewers who have subject matter competencies. That is the biggest difference for him it seems.
You have two aspies who are feeding off each other.
My impression: This is like watching a really smart person talking to a super smart person.
I think you`re spot on.
@C0mmand 0ption his method works to get some of the most interesting guests on. He gets on my nerves often but it’s often my best (and only) option to hear hours of real intellectuals. Maybe he’s just plain enough
I had the image in my head of Newton talking to Einstein.
Wow. This podcast became huge. Great job Lex. Any plans to speak with Demis Hassabis?
@Jan Valach Pfft
This would be epic
Please make this happen
off course it did its an interview with ellon musk
2 views per sub is standard
Man, these robots are just getting better and better. They’re now discussing how they’re making other AI. Amazing!
Horrible but true...
@Larry Myers I think we could be friends 😂
@Larry Myers something will get electrocuted in the procedure
AI is responsible for the design of a range of products currently in use.
🤣🤣🤣🤣🤣🤣
My favorite part in this interview...When Elon Musk was asked by Lex, “When you or someone else creates an AGI (artificial general intelligence) system, and you get to ask her one question, what would that question be?”
After a 1 minute pause, where he’s deeply thinking, he finally responds, “I’d ask, ‘what’s outside the simulation’”?
A system can never know what’s outside the simulation because it’s only working with data it’s been fed. “In order to find that which is immeasurable, measurement must first come to an end”
@@shaduwy Quantum Mechanics be like: "Measurement what? Lulz 🤣"
@@shaduwy perfectly said. "The last question" by asimov plays around with that idea. Highly recommend, if you havent read it
He did not pause for one minute he only paused for 17 seconds. Because 17 is just a test.
@@TheVintageRomance Westworld Season 1 plays with this too.
"what's outside the simulation" gave me quite the chills. What a great question.
Funnily, a horse is more autonomous than a car
For a while.
Horses have personalities and can be VERY undependable. They also can fall down, or be put in danger by their riders!
I will use this comment if you don't mind me XD
Ok, now show me a horse that can take the right exits on motorways and understand traffic lights colours bro.
@@hl8684 yeah we are going back in time.
Very very very informative. Nothing compared to trashed pumped out by CNBC, FOX, Etc. for economic benefits.
Thank you both for such powerful interview which helps general public have a glance at the future of AI and full self-driving.
Yea and also CNN dont forget them
great comment! ❤❤🇬🇧🇬🇧🇬🇧
“Well, frankly, it’s crazy letting people drive a two-ton death machine manually.”
🤣 I love Elon. Check.
Bruh...why does it feel like it just started at the end? This was so good. Well done both of you.
You ever realize that you really needed something but never realized it until you saw it. This video is it.
Like sleep?
Thomas Edison approach.
Create the desire for something that doesn't exist, promise it does exist, then hire people to create it in your name, sell it. Repeat step 1.
He's said before Edison is one of his heroes and models.
@@STho205 what If there was nothing until some one desire/dream, then It could become possible to make It real? thanks for sharing!
Gohan is gonna defeat Moro!!🤫
29:30 "You're confident that with Tesla you can create the world's best toaster" LMAO!! Goteemm
Problem solved
So thats why the Techpriests of Mars worship Toasters! It all comes down to Elon making a divine toaster and bringing it to Mars. This kickstarted the Stellar Exodus and very much later the Cult Mechanicus...
Pretty apt analogy. Where's the burn?
Fridman: "Nice outfit, bro." Musk: "Thanks." Fridman: "No tie, tho?"
Game, Set, and Match. Loved the 2 Teslas in the background and in the reflection next to Elon.
like two eyes
Startegic product placement. Musk must have someone that knows marketing. He doesnt have anyone that knows how to run a company though.
@@alilwittiness He's doing it, and you're not. TH-cam comments are so funny and full of hubris.
@@alilwittiness He's a micromanager, if he wants to run the company everyone else who wants to run the company are either stfu'ing or have left
I often think when people interview Elon : "Stop wasting his time, he's got work to do!" But this is one of the rare cases where his time was well used.
Well done Lex
He regretted it so much he did another one
Lot of spicy memes to get to on Twitter.
I've thought the same thing before!
But how's this for a mind bender... what if we were to assume that everyone had the potential to achieve great things like Elon, but if you want to get there then there's "work to do"?
What if the person whose time is being wasted the most...
...is _you?_
Hear hear, great interview all around!
You can listen to this several times and still learn something new each time! Brilliant content - FOR FREE, TOO!
Always satisfying to listen to an intelligent conversation! well done!
I wish they would both just calm down a bit.
Practical Islamic Finance yeah they are so FUCKING hyped😳
@@joshnic6639 I thought the "other guy" did a really great job. He clearly knew what he was talking about, asked great questions, and really brought the best out of Elon.
@@joshnic6639 you should check who is Lex. You will then appreciate better that his questions were spot on. Also, Elon Musk will never take 30 minutes of his precious time to talk to some random guys. Elon Musk likes to be challenged by smart people. Lex is one of them.
LOL YOU SON OF A BITCH
@@aqynbc and Lex has a black belt in BJJ ..he is one nerd that could fuck u up
Wonderful talk. Could easily have sat through another hour and a few more cups of coffee.
it was way too short. didn't really get into AGI.
I find the Horse an amusing metaphor because it can move on its own and also operates on renewable energy. :D
Lex: “You only have a certain amount of money”
Elon: “And I took that personally”
This was a satirical reenactment of what you would hear if two supercomputers were to have a conversation on a human podcast.
Hahaha! So true!
I love these guys, but this comment made my day. :D
You are a genius. That was hilarious!
Ok I will kill humans 🤖
Literally lol'd. Well-said!
"It's pretty crazy letting people drive a 2-ton death machine manually." - EM
When did he say this?
25:54
@@lonelyminer22
Twice
Can't wait for the headline "Tesla CEO calls their products 'death machines' "
@@FPRobber Don't provide clickbait title ideas please :D
This sounds like a job interview
when Elon applied for Tesla
They were dressed for the occasion
LOL
Your last question really had Elon thinking hard.
Very well done on this interview. 👍
Say what you may, Elon Musk is a genius! And Lex asked very probing and very informed questions. This interview makes me want to buy Tesla stock right now!!
Feralz right but are those two things that related? The energy crisis from dwindling fossil fuels is millions of times more a threat than traffic fatalities.
@@robertblake1228 i used to think like that but elon is 100% right in the fact that in the future people will say we were crazy for driving cars , i actually think about that sometimes when im driving fast on the highway or anywhere else, just a tiny shift of my hands of 10 degree can make me crash and possibly kill me or others around me.
Finally this is the interview that I've been waiting for. Two of the people I most admire in self-driving technology.
Amazing podcast, even better ending! Loved every second
That pause at the last question made the answer send chills down my spine. Damn good stuff 👍
Fascinating how Elon's brilliance is so focused while concurrently immensely flexible. When he's thinking, you can practically see his mind and imagination working together to weight out and explore different outcomes. Inspiring watching how passionate and immersed he is in the journey itself, not just arriving the destinations. Thank you, Lex!
Always fun when the interviewer doesn't rush Elon and lets him think before he talks.
Elon is in his element here... right at home
Well said
Its like elon said before in the interview: the better the AI is, the more you can engage the autopilot. with the automatic lane change there will be a huge increase in time spent with autopilot engaged. therefore you get a lot of more data from these critical moments, thereby vastly improving the learning curve of the AI. this in turn alows you to engage autopilot even in more tricky situations and creating even more useful data. the better the AI, the more you can use it and the more you use it, the better the data -> better AI.
you're talking about musk as if it's human...don't let it fool you....lol
Watched like 2 minutes and already wish this was longer.
@@Stan_144 I think he was often like that during his interview with Joe Rogan as well; he seems quite reserved about what he'll say. Not sure about the "disclaimer" at the beginning as well. And, at the time I made that statement, I really was roughly at the 2 min mark, so yea, my expectations were higher.
@@Stan_144 His eyes are not "wandering" he is THINKING.
@@Stan_144 this is usually what people do when they work through a question and are trying to avoid providing a canned response
That's what she said.
@@--SPQR-- You took it out of my mou... wait....
a TRUE genius! Mr Musk... only come once a millennium, you are sooo lucky to have sat across and interviewed him!
"what's outside the simulation" damn that gave me goosebumps for some reason being his answer.
Wow! What a beautiful podcast! I loved the ending...!
@@brando3342 I think it would be very difficult to come up with an honest true answer to this question in the few seconds provided. Of course Elon would have to use some humor and take reference from the previous topic to get through that question so quickly.
However, you are undoubtedly spewing the "good news" and want us to all accept the undeniable fact that God walks among us, testing our ever-lasting souls for the attributes he deems acceptable in his white billowy country club he runs in the after-life. No Jews, Muslims, Hindu, or atheist need apply, God tosses their applications straight away.
@@brando3342 I'm not trying to make fun of anyone. To answer directly to "Not their fault and you don't know who's going to heaven, neither do I." I would say, you don't even know if there is a heaven to go to.. You were born into Christianity. I was also born into Christianity, but have discovered reasonable doubt.
"It isn't a "simulation". This is Satan's world, the delusion God sent. Ignore it, go find your family and live happily." is a quote from you on a seperate thread in this video's comments. You seem pretty confident in your knowledge of god, and his intentions. I admit I have jumped to conclusions based on your other thread, and it was insensitive and lazy to assume to know the meaning of "Elon is not being honest."
What did you mean when you said Elon was not being honest in his answer?
When u got a funeral to go to but have to interview Elon Musk on the way
underrated comment :D
Lol
Lol exactly
or shoot some aliens and erase memories.
Hes wearing funeral clothes because hes gonna BURY Elon in the driverless car field!
Nice to see Elon get interviewed by someone with the ability to challenge him with real probing questions. It's interesting to Elon struggle with his desire to speak frankly to a fellow engineer against his obligation to speak as the CEO of Tesla.
I feel like After listening to many other Elons interview I have noticed that he will have the answer prepared but always respond like he is struggling.
where is the probing question? everything seems open end and general.
PLEASE MAKE THESE LONGER!!
I love the way Elon thinks. Things are just so obvious and automatic to him but they're revolutionary to the rest of us. His ability to see the future crystal clearly shows he's a true visionary.
31:30 "there may be ways to test if it's a simulation, but you can imagine that once an entity in the simulation found a way to detect the simulation, it can do something that corrects for that error"
31:52 "so when someone creates an AGI system and you get to ask her one question, what would that question be?"
32:16 "what's outside the simulation?"
32:20 Simulation: "time to end this podcast"
makes sense
But... Elon's answer was simulated. Assuming a simulation of course.
most underrated comment
Haha. Spot on
The time Elon took to answer that question is equaivalent to running this over the top tier hardware AI system and getting the results
The intro reminds me of the prompts I give to gpt-3 for it to understand what it has to do lol.
Lex is so pure. Until now he's still himself. How time flies. This is the first Lex interview I've ever seen. Keep up the good work Elon & Lex! I love you guys! 😘
Klau zku dgn x lon mask..dah pasti sialmat dlm lokap
"What's outside the simulation ?" WOW super question!
I love it when someone puts Elon on the spot with a question like that. You can see the cogs turning ... and then he rarely disappoints.
That’s equivalent to asking what’s in the other dimensions. It’s a fair question which I think wouldn’t be able to be answered by AI for quite a while
@@joseph1541 Assuming for a moment that hardware processing power wouldn't be a bottleneck, a true AGI would be able to self-improve exponentially and therefore its intelligence would be effectively infinite. Now, whether its knowledge is effectively infinite is a whole other question that's dependent on the data it has access to. But if it's supplied sufficient relevant data, it should be able to answer any question. Of course, this is all hypothetical science fiction currently, as we don't really know yet if true AGI is actually possible.
A simulation is an attempt to replicate the universe. The universe can't be a simulation you fools. By definition.
Can it be more than you ever imagine? Well, you bet your ass it can. But that doesn't mean any of it is fake, when, obviously, it's the realest thing you can perceive at this time of your existence
Yes, it was. I got chills. This was a great interview.
"What's outside the simulation? "
Pretty good question indeed.
It's something Isaac Asimov would have written
@@Skip2MeLou1 even the joe rogan episode
Jaimie Knox there is a book called the last question by Asimov
Stacked simulations.
@@chibula I think this question is maybe more important than Isaac one, I mean If there is no entropy outside the simulation, clearly is not the last question.
Elon question is not the last one neither, because, maybe there is another universe in which our simulation is running, but there is more simulations? which one was the first one? is a cycle? it has a end or a beginning?. Still, Elon's question approaches this concept more
Looking forward to the 4th interview! ☺Thanks, Lex & Elon!
Disclosure is a must. Great work Lex
Aliens observing Earth - - “ that may be a civilization worth saving, send down an Elon “.
This is my favorite thought ever.
Humans observing Aliens, Humans operation Elon to invade Mars.
Damn! You figured it the fuck out!
Anyone else just love watching lex ask a great question and then Elon almost reading notes in his mind to answer it?
The difference of comfort from 1st appearance to round 4 is nearly opposite ends of the spectrum. It's neat to see these two have conversations and how they evolve over time each appearance.
This is some of my favorite content on TH-cam. Thank you, Lex!!!
A friendly suggestion if I may,
Pull the camera back just a little, bring it up slightly to drop your eye level, and leave some head room (space between the top of your head and the frame of the video)
May be useful for reducing the sense of being imposing and general claustrophobic vibe.
Just a suggestion based on general observation. You might have technical constraints I'm not aware of or what not, and I'm not trying to tell you how to run your show. Just minor tweaks that may yield improvements.
Cheers.
I agree
Agree and an occasional 2 shot of the people involved in the Interview
why can't they just talk about the facts: why are they so emotional about everything?
cuz this shits epic
They are romantics
Because they care
i'm so autistic that i thought this was a serious comment lol
Real LOL 👍
You're a great interviewer. Awesome interview!!!
Amazing to watch 2 geniuses communicating (and the interviewee's ocular movement).
Joe Rogan: Let's just jump right into it
Lex Friedman: 30 minutes of disclaimer 😂
That's how you know he's an academic
🤨 2.5 minutes.
serious bro whats up with this weak intro.. lets get into it already he still talkin about stuff we already been knowing since the JRE podcast. alrighty then its about to start... oh wait its not hes still talking to himself where is ellon ???
@@MealTeam6 If by "dumb" you mean people who don't take themselves too seriously and know when to let go, then I agree. Pretentious cunts need not apply.
A P : "A. P.!!!!!........ I thought i told you to get off YT and back to bed mister. I'll tell your mother!
@@MealTeam6 Joe Rogan is a comedian, I don't know what you expect of him.
I like the style of the video , no music or graphics .
So... podcast?
@@stormwaker video/podcast
Tay international
Or tone difference in their voices
Or any sign of life
its a kind of interview which will be played in the future if things worked out smoothly or not
That last line was so epic. Such a good soundbite
My instant respect for this person - Lex Friedman - have never seen him before but it’s instantly obvious that the guy is brilliant. Very sharp and insightful discussion, really enjoyed it, thank you.
absolutely amazing, came here because of Joe Rogan, stayed because of Lex Fridman - SUBSCRIBED.
omg! me 2! lol
Me 3. This dudes the fuggin trooth
Lex is just a peanut compared to Elon
@@tamie341 But a coconut compared to Joe Rogan 😏
@@tamie341 my god! Just dont go on comparing everyone to elon and say the're nothing! Elon NATURALLY had a reading habit and had a computer since his childhood. Hkw many people have that NATURALLY? The thing is you're fucking jealous of Lex Fridman, so degrading him by comparing to elon.
Get some better job to do.
"What's outside of the simulation?"
Windows error sound.
it's called chord.wav, th-cam.com/video/_YcgHvp6xIo/w-d-xo.html
GREAT!!! I like you both and have to say that was an amazing podcast ...
the matter-of-fact, dry, straight-to-the-point tone is hilarious and awesome:)
25:54 Elon "It is pretty crazy letting people drive a two-ton death machine manually"
He's right. The amount of drunks, druggies, retards, idiots, imbeciles and women operating this machinery, is unbelievable. Drivers licence should have much more strict conditions.
@@after_midnight9592 You want to ban women from driving?
Raed Frazier
Yes
@Stehv Burgueno that's odd. My parents pay more for me and my bro. Might wanna check that
@@after_midnight9592I Wasn't going to like the comment because it didn't had to be said, it's so obvious.
But then i read women and i had to like it :D
Elon :
"What's outside the simulation ?"
God : "What an epic 4th wall break."
Cool. Now time to do another reset. I wonder what the next batch asks! 😂
"God starts sweating"
1. Maya... [Reset]
2. Atlantis... [Reset]
3. Now Elon asked wrong question....
3.... 2 .... 1.... ... [Reset]
I just imagine God sitting in his heavenly throne, one foot resting on the opposite leg's knee. Chin in his palm, looking bored. Then Elon says that and God doesn't move for a second, then just gives a 🤨 "hmm" and lightly taps the glass dome behind which he sits, more in idle thought than for any structural testing purpose.
On the other side of the dome, the tapping causes an enormous tsunami to strike the coast of Japan, killing 14.
What if the notion of outside is only pertinent to the simulation?
Also, why do people insist on placing human attributes upon the divine? Isn't that rather arrogant?
Love that ending dammit!!
This is a gem! Brilliant questions and brilliant answers. Thank you.
As an enginerd myself, I found this video extremely impactful. There is such great deal of insight shared between these two. I will have to watch it a few times to completely understand everything introduced.
Damn, if you replay it imagine me that English is my second language
@@pablocarmona7998 Hahaha good luck sir!
I dont know why.. but I really am okay with Elon taking my data and using it. I feel the complete opposit about Mark Zoidberg
Zachary Schiess trust
Elon actually uses it for the good of humanity while the Zucc just uses it for cheap profit.
@@CockatooDude even if he a hidden vilian that wanna build an ai that will destroy us all iam okay with it, the guy is legendary
@@mike_o7874 for real if I had to choose someone to end it all....I would choose Elon over Mark Suckerberg. Something tells me he would actually take care of us while we die or give us a honorable death, Mark on the other hand would probably find the cheapest of all and make it as sufferable as possible.
@@mike_o7874 You being ok if Elon turned out to be evil all along, simply because you admire him, is pretty much how the most evil dictators in history were allowed to thrive. One of the best quotes I've heard came from the anime villain, Aizen Sousuke: "Admiration is the emotion furthest from comprehenshion."
What a wonderful conversation.. 32:45
I loved this podcast. The conversation that started from self-driving cars went all the way to the point of existential crisis in just 30 minutes. The part where they talk about the statistics required to prove to the regulators that self-driving cars are safer than manual cars sounds like that we need a number of deaths and incidents to happen, which of course will happen naturally as we slowly advance towards the point where it is needless to argue about whether autonomous vehicles are safer than human-driven cars. Even at that point, I think we won't be able to let go the fear of being not in control, but this conversation also seems to suggest that our only logical response in that situation would be to admit that we are (or will be) so far more limited compared to the AIs. Too little we can do for being so much of control freaks. No wonder why Elon in that regard is working on Neuralink, or BCI in general, as that is one thing that can bridge the gap between us humans and the machines. Must say thanks to Lex for organizing and sharing such a conversation with Elon.
one issue I have is that just because the AI can outdrive the masses statistically does not mean it can outdrive me with me fully paying attention and all my intuition and reflexes on full capacity. It would not come close to that. All it is doing if it statistically beats the masses is it is proving that it can outdrive granny, teens texting, and people who got distracted at the pretty lady jogging on the sidewalk. Suddenly, the AI "appears" to be a better driver statistically due to these edge cases, but it is not better than someone who is on their A game. That is why I'd struggle to ever trust it with my life because I truly will always be able to outdrive it if I'm at my best.
I Could listen to Elon talk all day. He alone is the reason I bought Tesla stock. If I needed a car I’d buy a tesla
Can you still buy Tesla stocks?
Best interview I've ever seen.Thank you so much Lex , for making this happen
Great interview! Love the “Two ton death machine”
28:20 Lex has some very good questions that seem to be washed over. Tricking neural nets using adversarial attacks on objects outside it's own environment is surprisingly easy.
That was a very professional interview. At a different level than the other youtube interviews that I've seen so far.
I really enjoyed it.
Thank you and Elon as well.
So long time waiting for this face to face interview!!! ..... LEX GOOD JOB!!!!!!
Great Podcast!! Internet need more podcast!
Whoa the gate is quite aggressive on your sound! For a second there I thought my sound was gone hahah
@25:00
Elon: Remember we have to make these humans think we're of their kind...
Lex: Right, I'll follow your lead...
Both: (Robot Laughing) haha
Exactly what I thought XD
Ha... Ha... Ha... 🤖
Elon "your research in last 3 years is moot" Musk
Exactly that's why he put that caveat at the start. Musk just said it not even an issue. Surprised musk didnt just say why invade a person's space
You can detect a different level of intelligence here. This MIT kid has a great brain for the workhorse technical stuff he did with his team. Elon, however, has that along with unrivaled common sense and practicality. A powerful combination to say the least.
I actually think the Lex's work on driver/passenger monitoring could be very assuring when teslas are in a robo taxi mode. Instead of it detecting mobile phone use, guitar playing and water drinking it could be detecting and reporting vomiting, littering and vandalism to the owner and Tesla.
@@missionpupa How did you come to the conclusion that it was the interviewer's intent to antagonize?
@DAG as if you have any clue what you are talking / critiquing about
Алексей! Спасибо вам огромное за это интервью. Обнимаю!
Just the idea that the car could circle/drive around/park a mile away, etc etc while you shop, go to the Dr, etc, but arrives when you call it, is game-changing. Think of all the time we've wasted just looking for a space...all that time is no longer lost....Eventually, self-driving cars will be the new ubers and taxis...with no drivers at all....unless legislation precludes it, which is possible....Eventually, we will get completely away from personal ownership of cars..Right now, the major hurdle for auto driving cars is all the cars out there being driven by humans....If all cars were self-driving/autonomous, the accident rate would plummet...
Clone this guy. We need full redundancy.
Clone this guy. We need full redundancy.
Clone this guy. We need full redundancy.
Clone these guys. We need full redundancy.
Clone this guy. We need full redundancy.
Clone this guy. We need full redundancy.
Elons expression as he is thinking of an answer to lex's last question reminds me of doctor strange going through all the different realities and considering all the outcomes.
☝️
same XD
Top siGma
this whole interview sounds like a cyborg turing testing another cyborg
23:13
25:57
I agree with Musk here. If "automation" is defined by “ letting a machine direct _itself_ all by _itself_ ”, then driver monitoring is actively AGAINST automation.