Full podcast episode: ua-cam.com/video/Kbk9BiPhm7o/v-deo.html Lex Fridman podcast channel: ua-cam.com/users/lexfridman Guest bio: Elon Musk is CEO of Neuralink, SpaceX, Tesla, xAI, and CTO of X. DJ Seo is COO & President of Neuralink. Matthew MacDougall is Head Neurosurgeon at Neuralink. Bliss Chapman is Brain Interface Software Lead at Neuralink. Noland Arbaugh is the first human to have a Neuralink device implanted in his brain.
This quote is true until such time the AI will achieve Super Intelligence in which case it will be able to seek and find the absolute truths. Lets hope we don't destroy ourselves in the meantime.
@@user-pt1kj5uw3bso now anyone that chooses to enact upon ambition is a sycophant? You should probably take a look closer to home and wonder why you are not moving up before trying to drag others down to your level.
He is also himself very smart and very talented in a lot of different areas, so its easy for lex to move around different types of expertise categories comfortably and not look like he is a fish out of water.
Lex does a great interview. He is smart, understands technical and philosophic concepts and can challenge and push back on his guest all while being super respectful. Great job 👍
Elon Musk Decision Algorithm: 1) Make requirements less dumb 2) Try to delete steps (we need to delete more than we need and then put it back in) 3) Try to simplify or optimise it 4) Any given thing can be sped up 5) Automate it
Not a bad way to go about it. Engineers tend to fall in love with the technology itself, not the problem they're solving or the solution. Over-engineering is one of the most common pitfalls.
Especially about a technology that is not yet(maybe never) complettly understood. A statement several AI experts currently are still holding. We have approaches to remedy that, but it takes time for them to come to fruition and having market pressures urge companies to be the first and win, could turn out counterproductiv. Already economists warn about a bubble where 500 billion were invested in a short period of time while maybe 10 billion in revenues are flowing. So many are stepping up their AI game, which in terms of competion may seem a good thing, but are they all going about it in a safe and secure fashion? The military and its equipment has very tough restrictions on how programmes need to be coded and i think AI companies need to have at least such strict regulations for the building of AI systems.
Many people that interview Elon struggle to extract a coherent and well formatted output from him, or just fail to get him on topic entirely but this is gold. Perhaps it's just Lex unbreakable calmness but I rarely see Elon stay so on topic for this length of time, it makes for great viewing.
"Some people will never learn anything... because they understand everything too soon." ~Alexander Pope "We must all suffer from one of two pains: the pain of discipline or the pain or regret." ~Jim Rohn "For a small reward a man will hurry away on a long journey, while for eternal happiness man will hardly take a single step." ~Thomas à Kempis "Who seeks more than he has hinders himself from enjoying what he has." ~Solomon Ibn Gabirol
Elon, you can eliminate motors from arms by using electric powered magnets versus springs. Springs force fingers to be extended. Increased voltage amplifies magnet to curl the fingers. Entire system connects to wrist via kettle rotator, allowing infinite wrist rotation and delicate finger control simultaneously.
my grandpa lost his pinky and his balance in that hand was severely impacted, the pinky's range of motion and size is extremely crucial when it comes to balancing.
If we armed and equipped the estimated 50 million kangaroos in Australia, they would form a massive force. To put this into perspective, the largest active military force in the world is China, with around 2.3 million active personnel. With 50 million kangaroos, this hypothetical army would be significantly larger than any current military force. In terms of sheer numbers, it would be the largest “army” in the world by a considerable margin.
Lex you are a brilliant interviewer and I’m so grateful to be able to soak in your discourse with my all time hero, Elon Musk. In addition to Elon’s extraordinary intelligence he is a “force for good.” To seek the truth is so important as we move into the realization of SuperAI….wow! Thank you, Lex!
In the movie “terminator 2: judgement day” there were scenes with ample emphasis on the T-800’s robotic arm. Seemingly correlates with Elon’s response in this interview wrt Tesla’s Optimus humanoid engineering focus
What's the 4th step? 1) get the requirements less wrong 2) delete more than you think, so much that you have to put something back in 3) optimize 4)? 5) automate What is the 4th step?
@@arab6745he was asked which he admired Edison or Tesla more? He responded Edison. I was genuinely surprised because I thought of Elon as a dreamer. His reason is because Edison was a better marketer. At first, this I minded, but when I thought of actual achievements, Edison did far more while Tesla was a genius. To humanity, Edison was more helpful. I don't know if Tesla was not stopped. He should have done something to protect himself. IMO. By the above reasoning, I understand why you don't like the I'm, but it made me understand his motivation more. What I find more interesting is that while his short term goals change all the time as technology is always changing, his long term goal never changed. In my opinion, his ability to market is part of him like it or not. I don't think it is a bad thing. I think being able to persuade people to try new is incredibly hard thing to do. I appreciate the skill.
@@arab6745 We all lie, first of all. If you really think about it, who lies more? the media? The government? Even doctors and hospitals. His lies....yes, I guess one would think so. I would like to you to think, are his lies are about things that had never been done? What about the other timelines? Like the ones that have been done and repeated? He is setting unrealistic goals to follow so that he can speed up. He is always late. Perpetual delayer, but always first timer. I forgive him for that reason. I can't call it a lie because he believes that it is not a lie.
Don't think that most people really understand what Tesla is. Most of the folks who listen to Lex do, but that accounts for the minority of the folks who own Tesla stock. I've tried to explain it to investor friends and they tout headlines rather than evidence, claiming anecdotes like "Company X is closer to getting higher level autonomy clearance", or other non correlated data. They don't realize the immense training data that Tesla's AI ecosystem has been building since 2013ish, and that's even if they comprehend that Tesla is a spatial AI and robotics company and not an automotive company in the first place. Elon has somehow been able to get his hands on and innovate on pretty much every area of the global economy.
It's never wasted effort if both sides reinvent and upscale learn adapt then next meeting or cycle is much more understood and appreciated with 10 more time opportunities for both sides. But now everyone is 8n optimum communication . Much like adaptation in Mico and macro sales and utility of that scenarios.
that's why I advocate for an Anabaptist / Old Order Amish lifestyle (or traditional agrarian lifestyle for cultures worldwide) - no cars, no skyscrapers, no supercomputers - just natural families living in community close to the land. Tech creates more problems than it solves.
@@mikekelly3650slippery slope deciding what tech is ok to use vs other tech. What is primitive to us was once not so, but if primitive man never sought or used tech...oooh oooh aaah aaah
In the body modification world some people amputate their little fingers because they think it looks cute. I wonder how many of them regret it? I've got big holes in my earlobes and tattoos but chopping off body parts seems a little unnecessary to me. At least they aren't grooming children into doing it though.
xD in middle aged europe/asia they cut fingers of expected criminals so they couldn't use their weapons. Thats one reason why showing someone the middle finger is still seen as an offence in most countries.
I think certain technological advancements, such as AGI, should be developed outside of the capitalistic structure and should be a globally collaborative and beneficial effort.
Humanity couldn't even begin to debate exactly who should/could work on such a project. Development of AGI is extremely political and deeply militaristic in nature. We would either go to war over it, or just never get anywhere, while everyone develops in secret. This technology can only bring us closer to the abyss.
A partnership could be formed between the best universities in the world, in regards to computer science, physics, philosophy, etc. I think a conflict such as a war would create a great sense of unity for this project. unfortunately/fortunately there is no outside factions for humans to compete with in this pursuit so the competition has to be between ourselves which will lead to inequality, as the group to figure it out will not willingly share it for free. Ultimately this technology, like the exploitation of the atom, will be developed unless we lose the ability to develop it. I just hope precautions are taken to avoid social and economic disasters.
I don't trust anyone as much as I trust Elon in this space, I don't think you want China to have a hand in how world altering technology is implemented in the US
I think the focus should be on deduction rather than "truth". Mathematically speaking there is no absolute truth, all truths are either 1. derived from other truths by implication 2. Tautologies 3. Axiomatic. There only reason there appears to be absolute truths in physics is because there are certain things which are axiomatic to our existence, that is, if they weren't true, you wouldn't exist. So when you ask the questions, there are axioms from which other truths can be derived. But a la Godel, there are also true statements which cannot be proven from the axioms. As Elon says, we can *aspire* for truth and we should but that's the best we can do. And any super intelligence would know that. If we want to avoid long term idiocracy, the general public needs to learn to appreciate the subtleties of logical deduction, and when an AI gives an answer to a question it should list all the assumptions and deductions it used to arrive there. Each individual logical step, just as a mathematician would when presenting a formal proof.
"mathematically speaking"??? Musk points out very specific examples of untruths that are currently being perpetrated in current society that are wrong and harmful to humanity. No need to complicated any further with "mathematics"
@@dcarlisle123 musk plucks out several ideologies that he enjoys fearmongering over in his right wing echo chambers because he doesn't understand the fact that all data has IMPLICIT bias, for which he confuses his own IDEOLOGICAL biases. he literally doesn't understand how data collection and logic work, and it shows.
+ 0:07 you'v also said Play to win or don't play at all 4:37 until you came along and started building Optimus it was through to be an extremely difficult problem 9:35 The mantra
12:37 : It is so clear now why elon decided to fire the Supercharger team. The department head resisted laying off any employees, which to Elon, indicated that the team was immediately overstaffed. As a result, he fired everyone to rebuild up.
*"It is clear now why Elon says he decided" Fixed for ya. The Lex / Elon sections get very aholelicking so it's important we keep some feet in the real world.
@@darin4704 Possibly. But if you are looking to do what Tesla is doing with an extremely competitive and really expensive path to get there, one of the management philosophies is to stay lean and constantly reassess the need for personnel in specific areas and who is valuable and who is not. It is harsh, but if a person can even get in 3 to 5 years of experience at Tesla or SpaceX or any company doing huge things and they get laid off, that is a very strong thing to have on a resume. This is aggressive management 101. He as revolutionized several industries so far and has many thousands of applications to his companies all the time so it must be working
Hi Lex- can you please ask him how the entire waiting line disappeared on the Cybertruck where we are getting invited to pick up the car in October when we were supposed to be #250,000 and how there’s massive sell off of the Cybertruck on CarGurus at basically the purchase price?! This is starting to sound like some kind of shuffle. Good video though as always.
Insights By "YouSum Live" 00:00:13 AI requires powerful training compute to succeed 00:00:21 Training compute improvement must outpace competitors 00:01:41 Data access and human talent are crucial factors 00:02:36 Optimus robots will provide vast real-world data 00:04:36 Humanoid robots could reach production levels of billions 00:08:01 Engineering humanoid robots is a complex challenge 00:10:15 Simplifying processes is essential for effective engineering 00:19:27 Truthfulness in AI programming is critically important 00:23:10 AI must avoid ideological biases in its design 00:24:30 The future of AI depends on rigorous data filtration Insights By "YouSum Live"
@darkness1130 how can outweigh the benefits? Our skeletal structure is capable of many things. It will intact open the door to many improvements. I mean, we are the blueprints. there is no need to waste time on design. Manufacturing would be not a problem at all.
@armadasinterceptor2955 it does make sense. Some fields of physics already use large nets and target them to only understand, turbulence control, 3-Body-Problem, manufacturing at quantum level, or atom-core-clocks. Those nets are not as big as chat-gpt, as mankind is far more interested in chatting than astro or nuclear physics. There are also applications in math, but i don't understand their problems.
Maybe some form of capacitance and reactive mitigation might help with the jitter. Or tune these at least if already on board...? A suggestive question I guess.
Wow, you did it again, Lex. You revealed the genius of the greatest engineer of all time, Elon Musk, in a way we can all understand. Thank you, Lex. You are the best.
You’re supposed to overcorrect when you call for indirect fire too. We called it bracketing. It’s cool to find multi-disciplinary principles. I’m going to have to remember this. It reminds me of the Pareto principle. Which kind of seems to apply just about anywhere you look.
0:34 When Musk says "by the end of this year", it's like those cartoons where the criminal leaves their little notes and clues for the detectives. Facing fraud charges and still uttering the exact same phrase is legendary
Wow Hate much? Dude’s companies have connected a guys brain to a computer. His other company invented the electric car industry. His other company has Rockets that land on barges in the ocean and take astronauts to space. And you’re here casting shade like he’s such a loser. Your panties twisted because he supports free speech and maybe doesn’t agree with your political opinions? He’s allowed his opinion just like you and if you knew anything about him you would understand that he’s simply trying to push humanity into a better future.
Are you sure about that? What about power, metal, silicon, deleting human so the world stays colder and every electrical systems rus more efficient ... Well last point is a bit extreme, but power and silicon production sure is relevant.
I still not able to understand how they're struggling to construct a robot duplicating the functioning of human body yet they discovered a electronic device that can able to merge along with one of the most mysterious, not fully studied organ in human body BRAIN.
Our muscles work a bit different, then the motor devices we have now. weight, power, precision...still hard for a machine, even with a price as big as Tesla for one prototype.
When he explains that safety is a function of alignment to truthfulness I think it is slightly different. I personally believe that it should aligned with factfulness. The truth depending on the perspective or opinion carry a connotation of being good or bad. Whereas facts are simply stating a immutable non opionated conclusion
11:27 Three years from now… Elon: Can you solve climate change? Use my engineering principles to solve this problem. AI: Ah, the main issue seems to be humanity. I will move forward to delete this as per you request. Elon: Big oof..
As for now, we still haven't complettly figured out the inner workings of LLMs, but sure giving more compute, more abilities, internet access for agents and what not, that sure carries no risks at all, which seemingly doesn't matter in the persuing of winning.
@@Art-is-craft well differance is, that they are less likely to cause an uprising or otherwise fatal accurance of mass deaths than a system used on large scale that may only later become faulty. Like say due to a bad patching process as recently witnessed with Microsofts OS Windows and that security company. It is on a different abstraction level, physics, chemistry, biology, technology ... maybe AI neurology will be a subject for studies more and more. Point being, there are ways to understand the underlaying principles i am quite certain, it just will take time to find and proof them. Even then though, depending on the advances meanwhile are made and i think there is equall effort as well as chance at times that lead to those advances, objectivly being able to know and think the same as others do was never something we could achiev between the same species as in us humans. We may find the ruling principles, but truely understanding then in the end where this leads to ... a different type of entity and intelligence, that seems like a big ask anyways. So if we can not rely on that happening, the alignment and control of an ever increasing intelligence seems hardly even worth considering before that. Therefor i would guess we are in for a wild ride.
The system-wide RDMA is quite something. IIRC Frontier has that within a compute node with 4 h100s, which is neat and Fugaku, while not RDMA took pains to have 6D interconnects. Elon’s engine is quite something.
I am unsure, but isn't removing ideological bias also a subjectiv task in itself and thereby contradicting. I further assume there is no living being able to hold itself in its judgment to Kant's imperativ, but always have at the very least its own perspective bias, not being able to be objective even though if trying to be. The statistical amalgamation that are training data sets, some may see as the closest approximation to Kants imperative, but seeing that certain biases then spill over ignoring huge sections of people at best, activly discriminating at worst ... i may give a point to his 5 point step program to avoid unnessary functions or maybe at some point stumble upon better ordering mechanisms.
Toyota, KUKA, Miso, Philips Medical, Hanson, Shadow Robot. There are at least 2 dozen companies holding very strong know-how about humanoid systems like the hand. And most of them you haven't even heard of even if you are working in the field.
I hope that when I watch the whole interview there is mention of the 3 laws of robotics. Ive thought many years on those and still can not find a flaw in them.
I'd say 10 years. Many companies have already unveiled robots that do most household chores with demos. They just need to be trained more extensively before being fully released.
public personals tells what people want to hear, he tells things as they are with great understanding of any topic he speaks of. Too much of a meme lord for basic personas, he truly believes being himself is allowing him to win so that's what he does (and it really is working tho lol)
The sensory control distortion always _changes_ too. Before I was typing this... my _elbows_ were hightently distorted, even above my face and fingers. They were manipulating the arms on my swivel chair. Thanks. Elon and Lex for "making" me notice 😃
So one of the most innovative companies the world has ever known is struggling to reengineer the human hand by “intelligent design” yet we teach our kids the actual human hand evolved thru random genetic mutations.
Every model xai has put out has been subpar. Vastly undertrained. Haven’t solved any of the inherent limitations. Pure scale isn’t all you need. This is pure ego and cap.
I wouldn't write them off. The company was hobbled together very, very quickly and the models they've released so far weren't trained with that much compute. I think k we'll have a clearer picture of things when Grok 2 and 3 come out later this year.
this is not zero sum game. elon sometimes is incredibly out of phase. he wants to win or to prove to everybody, that he is megasuper duper smart, that his solution is the best or his vision is the right one, while this is only showing off nothing more.
13:01 #4 any given thing can be done faster 24:01 there’s so much ai polluted material in just searching for things 24:20 “is this data most likely correct?”
Full podcast episode: ua-cam.com/video/Kbk9BiPhm7o/v-deo.html
Lex Fridman podcast channel: ua-cam.com/users/lexfridman
Guest bio: Elon Musk is CEO of Neuralink, SpaceX, Tesla, xAI, and CTO of X. DJ Seo is COO & President of Neuralink. Matthew MacDougall is Head Neurosurgeon at Neuralink. Bliss Chapman is Brain Interface Software Lead at Neuralink. Noland Arbaugh is the first human to have a Neuralink device implanted in his brain.
❤
Im old enough to remember when Elon tried to force competitors to stop their research due to safety concerns.
there has to be 2 types of AI:
1. one for our trivia searches
2. one to self-learn so that it gets so smart it can teach us
@@LexClips good job 🫡
22232
"If you train AI to lie, you're asking for trouble, even if that lie is done with good intentions." Based.
What about training yourself to exaggerate and lie constantly?
This quote is true until such time the AI will achieve Super Intelligence in which case it will be able to seek and find the absolute truths. Lets hope we don't destroy ourselves in the meantime.
Google's Gemeni is already a liar, and they have deliberately trained it to be so.
And let's hope that Ai won't gravitate to "lord of this world thinking" after a few milliseconds of conscious initiation...(even though it will be).
Like leftists? @@sn1000k
You can tell that these two are friends at this point. Lex's hard work and perseverance really moved him up the social ladder. I'm proud of you, Lex.
Not sure about the social ladder,but certainly the intellectual ladder plus he's very easy to listen to..like the best lecturer..
Yes hardwork, definitely not his sycophantic tendencies
He's always a heavy hitter. sleeping giant for sure
@@user-pt1kj5uw3bso now anyone that chooses to enact upon ambition is a sycophant? You should probably take a look closer to home and wonder why you are not moving up before trying to drag others down to your level.
He is also himself very smart and very talented in a lot of different areas, so its easy for lex to move around different types of expertise categories comfortably and not look like he is a fish out of water.
Lex does a great interview. He is smart, understands technical and philosophic concepts and can challenge and push back on his guest all while being super respectful. Great job 👍
Even Elon calls it Twitter LOL
It’s sentimental by all. ☔️
When the AI companies scraped the platform it was still called Twitter.
This interview is very old
No it's not @@777VOID-1
@@777VOID-1 If you consider two days very old?
Elon Musk Decision Algorithm:
1) Make requirements less dumb
2) Try to delete steps (we need to delete more than we need and then put it back in)
3) Try to simplify or optimise it
4) Any given thing can be sped up
5) Automate it
Thank you, sir!
1. Just engineer it as best as possible.
2. Delete rest of Elon’s stupid rules
Yes thats why starship is so overly complicated 🤣🤣🤣
Not a bad way to go about it. Engineers tend to fall in love with the technology itself, not the problem they're solving or the solution. Over-engineering is one of the most common pitfalls.
@@sirmiba elon isnt an engineer
"play to win or don't play at all" is very powerful statement
This is for Big Boys! Fastness relative…
Especially about a technology that is not yet(maybe never) complettly understood.
A statement several AI experts currently are still holding.
We have approaches to remedy that, but it takes time for them to come to fruition and having market pressures urge companies to be the first and win, could turn out counterproductiv. Already economists warn about a bubble where 500 billion were invested in a short period of time while maybe 10 billion in revenues are flowing. So many are stepping up their AI game, which in terms of competion may seem a good thing, but are they all going about it in a safe and secure fashion? The military and its equipment has very tough restrictions on how programmes need to be coded and i think AI companies need to have at least such strict regulations for the building of AI systems.
Nobody likes to suffer. That is 2+2
Don't do unto others that which you wouldn't have done to you. That is 2+2
2+2=🙂
Project 2+2:
Resist D.U.M.
UTS ALECM
Objectives/Alignment
Avoid. M.U.D.
sensAwewunda
The Age of Wisdom
UTS ALECM:
right Understanding
right Thought
right Speech
right Action
right Livelihood
right Effort
right Concentration
right Mindfulness
Release the video of Optimus walking in the park
Many people that interview Elon struggle to extract a coherent and well formatted output from him, or just fail to get him on topic entirely but this is gold. Perhaps it's just Lex unbreakable calmness but I rarely see Elon stay so on topic for this length of time, it makes for great viewing.
That's because the topics are designed to make him look bad.
"Some people will never learn anything... because they understand everything too soon." ~Alexander Pope
"We must all suffer from one of two pains: the pain of discipline or the pain or regret." ~Jim Rohn
"For a small reward a man will hurry away on a long journey, while for eternal happiness man will hardly take a single step." ~Thomas à Kempis
"Who seeks more than he has hinders himself from enjoying what he has." ~Solomon Ibn Gabirol
You can definitely have knowledge and no understanding, sxx bravo Juliett
nice quotes, I like the 2nd one a lot
It amazes me that people have so much hate towards this man.
Humanity is better off because of Elon Musk.
You don't read enough
@@lenny9548 I read plenty.
Elon is an amazing human.
@@lenny9548 You don't understand enough
I believe you are Genius. You are inspirations.... Good Man get Good Luck Elon.
Cobalt miners agree.
Elon, you can eliminate motors from arms by using electric powered magnets versus springs. Springs force fingers to be extended. Increased voltage amplifies magnet to curl the fingers. Entire system connects to wrist via kettle rotator, allowing infinite wrist rotation and delicate finger control simultaneously.
I like how Lex quietly has a mini Dimebag guitar behind him.
I like how Musk casually has one up his nose.
What!? hahaha
Why would it be making a noise?
Dimebag Darrell RIP
I did noticed that too.😂
my grandpa lost his pinky and his balance in that hand was severely impacted, the pinky's range of motion and size is extremely crucial when it comes to balancing.
Really?? That’s FASCINATING
I’ll forever be conscious of peoples pinky’s
If we armed and equipped the estimated 50 million kangaroos in Australia, they would form a massive force. To put this into perspective, the largest active military force in the world is China, with around 2.3 million active personnel. With 50 million kangaroos, this hypothetical army would be significantly larger than any current military force. In terms of sheer numbers, it would be the largest “army” in the world by a considerable margin.
no, no its not.
@@jjg1501 you can use google to not look like a fool
@@jjg1501 a simple google search can keep you from looking like a fool
Elon on Lex is one of my fav pods
❤❤same❤❤
It’s also the title of a completely different video on a different platform.
@@unnamedchannel1237 what model are you?
Elon and his Simp are also my favorite people
@@unnamedchannel1237link please, unless it's porn...
Can Elon run the US on Fridays ? Asking for a friend.
😂
When Trump wins he has promised Elon to lead the proposed government efficiency commission
Pretty sure we won't have a deficit if he wins trump that is
Seriously
Elon Musk is not like the rest of us... He has so much knowledge about so many areas
Whoa. Didn't see the link between Grok training sets and the unique data sets from Optimus runtime
Lex you are a brilliant interviewer and I’m so grateful to be able to soak in your discourse with my all time hero, Elon Musk. In addition to Elon’s extraordinary intelligence he is a “force for good.” To seek the truth is so important as we move into the realization of SuperAI….wow! Thank you, Lex!
In the movie “terminator 2: judgement day” there were scenes with ample emphasis on the T-800’s robotic arm. Seemingly correlates with Elon’s response in this interview wrt Tesla’s Optimus humanoid engineering focus
Love you talks, keep up the awesome work. Asking the real question!!
What's the 4th step?
1) get the requirements less wrong
2) delete more than you think, so much that you have to put something back in
3) optimize
4)?
5) automate
What is the 4th step?
Seems step 4 is to do it over and start from 1st principles and leave behind preconceived notions and constraints.
It’s refreshing to see that Elon at the end of the day is incredibly rational and naturalistic about everything he says.
He exaggerate for marketing too (Tesla's self-driving, for example). I used to be an Elon fan, but not anymore.
@@arab6745he was asked which he admired Edison or Tesla more? He responded Edison. I was genuinely surprised because I thought of Elon as a dreamer. His reason is because Edison was a better marketer.
At first, this I minded, but when I thought of actual achievements, Edison did far more while Tesla was a genius. To humanity, Edison was more helpful. I don't know if Tesla was not stopped. He should have done something to protect himself. IMO.
By the above reasoning, I understand why you don't like the I'm, but it made me understand his motivation more. What I find more interesting is that while his short term goals change all the time as technology is always changing, his long term goal never changed.
In my opinion, his ability to market is part of him like it or not. I don't think it is a bad thing.
I think being able to persuade people to try new is incredibly hard thing to do. I appreciate the skill.
@@arab6745who do you admire instead? If he doesn’t impress you the rest of us must be unbearable.
@@kjshy I didn't want to be seem aggressive, but by "exaggerate" I meant lying. But I see your point, and it has its validity.
@@arab6745 We all lie, first of all. If you really think about it, who lies more? the media? The government? Even doctors and hospitals. His lies....yes, I guess one would think so. I would like to you to think, are his lies are about things that had never been done? What about the other timelines? Like the ones that have been done and repeated? He is setting unrealistic goals to follow so that he can speed up. He is always late. Perpetual delayer, but always first timer. I forgive him for that reason. I can't call it a lie because he believes that it is not a lie.
1:06 Elon hasn't seen checo drive 😂
Don't think that most people really understand what Tesla is. Most of the folks who listen to Lex do, but that accounts for the minority of the folks who own Tesla stock. I've tried to explain it to investor friends and they tout headlines rather than evidence, claiming anecdotes like "Company X is closer to getting higher level autonomy clearance", or other non correlated data. They don't realize the immense training data that Tesla's AI ecosystem has been building since 2013ish, and that's even if they comprehend that Tesla is a spatial AI and robotics company and not an automotive company in the first place. Elon has somehow been able to get his hands on and innovate on pretty much every area of the global economy.
That's how it works in markets. 99% follow the headlines and are therefore months behind the 1%
It's never wasted effort if both sides reinvent and upscale learn adapt then next meeting or cycle is much more understood and appreciated with 10 more time opportunities for both sides. But now everyone is 8n optimum communication . Much like adaptation in Mico and macro sales and utility of that scenarios.
Musk...."The biggest error of most design engineers is they try to optimise something that doesn't need to exist!" 😂
that's why I advocate for an Anabaptist / Old Order Amish lifestyle (or traditional agrarian lifestyle for cultures worldwide) - no cars, no skyscrapers, no supercomputers - just natural families living in community close to the land. Tech creates more problems than it solves.
@@mikekelly3650slippery slope deciding what tech is ok to use vs other tech. What is primitive to us was once not so, but if primitive man never sought or used tech...oooh oooh aaah aaah
@@mikekelly3650 if you advocate for throwing away tech then why are you on your phone watching a tech podcast?
"Speeding up something that shouldn't exist is absurd" 😂
Grok, a good name. I read Stranger in a Strange Land when I was a teenager. It had a huge influence on me.
In the body modification world some people amputate their little fingers because they think it looks cute. I wonder how many of them regret it? I've got big holes in my earlobes and tattoos but chopping off body parts seems a little unnecessary to me. At least they aren't grooming children into doing it though.
your just not rad enof
damn phone keeps slipping from my hands..
They chop off more than there Little pinkies and they do let children do it.
No hate or judgement, but large holes in your earlobes didn't seem a little unnecessary?
xD in middle aged europe/asia they cut fingers of expected criminals so they couldn't use their weapons.
Thats one reason why showing someone the middle finger is still seen as an offence in most countries.
I think certain technological advancements, such as AGI, should be developed outside of the capitalistic structure and should be a globally collaborative and beneficial effort.
Humanity couldn't even begin to debate exactly who should/could work on such a project.
Development of AGI is extremely political and deeply militaristic in nature.
We would either go to war over it, or just never get anywhere, while everyone develops in secret.
This technology can only bring us closer to the abyss.
A partnership could be formed between the best universities in the world, in regards to computer science, physics, philosophy, etc.
I think a conflict such as a war would create a great sense of unity for this project. unfortunately/fortunately there is no outside factions for humans to compete with in this pursuit so the competition has to be between ourselves which will lead to inequality, as the group to figure it out will not willingly share it for free.
Ultimately this technology, like the exploitation of the atom, will be developed unless we lose the ability to develop it. I just hope precautions are taken to avoid social and economic disasters.
@@AlteredVisioned Man we need some Aliens to unite us.
I don't trust anyone as much as I trust Elon in this space, I don't think you want China to have a hand in how world altering technology is implemented in the US
@@MIbra96 Bro it might happen haha. It would definitely give governments and military more control over people
great clip, thanks Lex
Thank you for your hard work and talent! I will always stand by you.
I think the focus should be on deduction rather than "truth". Mathematically speaking there is no absolute truth, all truths are either 1. derived from other truths by implication 2. Tautologies 3. Axiomatic. There only reason there appears to be absolute truths in physics is because there are certain things which are axiomatic to our existence, that is, if they weren't true, you wouldn't exist. So when you ask the questions, there are axioms from which other truths can be derived. But a la Godel, there are also true statements which cannot be proven from the axioms. As Elon says, we can *aspire* for truth and we should but that's the best we can do. And any super intelligence would know that. If we want to avoid long term idiocracy, the general public needs to learn to appreciate the subtleties of logical deduction, and when an AI gives an answer to a question it should list all the assumptions and deductions it used to arrive there. Each individual logical step, just as a mathematician would when presenting a formal proof.
"mathematically speaking"???
Musk points out very specific examples of untruths that are currently being perpetrated in current society that are wrong and harmful to humanity. No need to complicated any further with "mathematics"
@@dcarlisle123 musk plucks out several ideologies that he enjoys fearmongering over in his right wing echo chambers because he doesn't understand the fact that all data has IMPLICIT bias, for which he confuses his own IDEOLOGICAL biases. he literally doesn't understand how data collection and logic work, and it shows.
this is a great comment. it's sad that most people, including musk, will never understand it.
“Jesus saith unto him, I am the way, the truth, and the life: no man cometh unto the Father, but by me.”
John 14:6
+ 0:07 you'v also said Play to win or don't play at all
4:37 until you came along and started building Optimus it was through to be an extremely difficult problem
9:35 The mantra
12:37 : It is so clear now why elon decided to fire the Supercharger team. The department head resisted laying off any employees, which to Elon, indicated that the team was immediately overstaffed. As a result, he fired everyone to rebuild up.
I think this is accurate
*"It is clear now why Elon says he decided"
Fixed for ya. The Lex / Elon sections get very aholelicking so it's important we keep some feet in the real world.
@@darin4704 Possibly. But if you are looking to do what Tesla is doing with an extremely competitive and really expensive path to get there, one of the management philosophies is to stay lean and constantly reassess the need for personnel in specific areas and who is valuable and who is not. It is harsh, but if a person can even get in 3 to 5 years of experience at Tesla or SpaceX or any company doing huge things and they get laid off, that is a very strong thing to have on a resume.
This is aggressive management 101. He as revolutionized several industries so far and has many thousands of applications to his companies all the time so it must be working
@@darin4704how’d that work out for him/them
@@darin4704capitalism is ruthless. Socialism is kind until it fails
Fourth time on your show because you’re a superior interviewer. Lemonhead is crying on his unemployment check. 😂
Elon Musk's laugh reminds me Beavis and Butthead
Hi Lex- can you please ask him how the entire waiting line disappeared on the Cybertruck where we are getting invited to pick up the car in October when we were supposed to be #250,000 and how there’s massive sell off of the Cybertruck on CarGurus at basically the purchase price?! This is starting to sound like some kind of shuffle. Good video though as always.
who thought of the terminator hand reveal in the movie when they started talking about the hand...
They had the same design
Insights By "YouSum Live"
00:00:13 AI requires powerful training compute to succeed
00:00:21 Training compute improvement must outpace competitors
00:01:41 Data access and human talent are crucial factors
00:02:36 Optimus robots will provide vast real-world data
00:04:36 Humanoid robots could reach production levels of billions
00:08:01 Engineering humanoid robots is a complex challenge
00:10:15 Simplifying processes is essential for effective engineering
00:19:27 Truthfulness in AI programming is critically important
00:23:10 AI must avoid ideological biases in its design
00:24:30 The future of AI depends on rigorous data filtration
Insights By "YouSum Live"
It shouldn’t be about winning. it should be about collaborating with others and innovating to create the best possible result for humanity.
I don't think that's possible for a man like Elon.
Should but isn't
Competition is what drives innovation. You want companies to be vying for first place.
Exactly. For anything else competition is good. For Ai it is utterly dangerous. Scary
You've already lost.
0:07 Elon with Half-Life t-shirt!!
Half life shirt. Freidman > Freeman. Grok 3.
Half life 3 confirmed
@@denjamin2633 I hope HL3 is VR. HL;Alyx is phenomenal, the very end reveal is incredible (in VR).
@@denjamin2633 Clever :)
@Elon Musk, why not use our skeletal structure on Optimus and then build around that?
No, dude. L take.
Interesting biomimicry concept..but the manufacturing complexity probably might outweigh the benefits..
@darkness1130 how can outweigh the benefits? Our skeletal structure is capable of many things. It will intact open the door to many improvements. I mean, we are the blueprints. there is no need to waste time on design. Manufacturing would be not a problem at all.
Now you see why AI image generators struggled with hands.
Does anyone know around when this talk was?
this week i think
45 years ago
😂👍🙏
There is so much synergy going on here. The genuine respect for the truth brings out the highest level of information. Together pure magic
there has to be 2 types of AI:
1. one for our trivia searches
2. one to self-learn so that it gets so smart it can teach us
Your idea doesn't make sense yet.
@@armadasinterceptor2955 Or ever.
@armadasinterceptor2955 it does make sense. Some fields of physics already use large nets and target them to only understand, turbulence control, 3-Body-Problem, manufacturing at quantum level, or atom-core-clocks. Those nets are not as big as chat-gpt, as mankind is far more interested in chatting than astro or nuclear physics.
There are also applications in math, but i don't understand their problems.
Maybe some form of capacitance and reactive mitigation might help with the jitter. Or tune these at least if already on board...? A suggestive question I guess.
Also a suggestive statement
@@Dontspreadnegativityit called tact
Wow, you did it again, Lex. You revealed the genius of the greatest engineer of all time, Elon Musk, in a way we can all understand. Thank you, Lex. You are the best.
Are you okay dude?
Nice ,,,,,,,,,,,,,,,,,,,,,, you used!
Bot
Bot
@@nitrozack3062bot activity detected 🚨
You’re supposed to overcorrect when you call for indirect fire too. We called it bracketing. It’s cool to find multi-disciplinary principles. I’m going to have to remember this. It reminds me of the Pareto principle. Which kind of seems to apply just about anywhere you look.
Musk seems like an AI to me..
How far did you get in school? 😅
He definitely “hallucinates” facts far too often the way modern LLMs do.
He’s on the spectrum . You would know this if you were around this type
He put Neuralink in first 🤣
This pod looked and felt like AI to me off the get go. Very odd
My question is aside from statistics how does ai learn sensations ? If so can it develop a taste in things ?
0:34 When Musk says "by the end of this year", it's like those cartoons where the criminal leaves their little notes and clues for the detectives. Facing fraud charges and still uttering the exact same phrase is legendary
Remember when he announced he was building a robot and he brought out a guy in a robot costume on stage dancing to Skrillex? 😂
when elon gives a end date you know its almost always wrong.
He is always late to the party but the party almost never starts without him.
Wow Hate much? Dude’s companies have connected a guys brain to a computer. His other company invented the electric car industry. His other company has Rockets that land on barges in the ocean and take astronauts to space. And you’re here casting shade like he’s such a loser. Your panties twisted because he supports free speech and maybe doesn’t agree with your political opinions? He’s allowed his opinion just like you and if you knew anything about him you would understand that he’s simply trying to push humanity into a better future.
Reminds me of when my brother was renovating his house and he said for the good part of the whole year just two more weeks. Lol.
The way to deal with extreme power jitter is to use something of a check valve to regulate the flow.
Why is it a competition? AI and humans are not competing for the same resources.
Are you sure about that? What about power, metal, silicon, deleting human so the world stays colder and every electrical systems rus more efficient ...
Well last point is a bit extreme, but power and silicon production sure is relevant.
I'm incapable of watching Optimus walk without picturing Joe Biden. I just can't separate the two.
Optimus still can't trip going up the stairs. That level of sophistication can only be described as "presidential".
I still not able to understand how they're struggling to construct a robot duplicating the functioning of human body yet they discovered a electronic device that can able to merge along with one of the most mysterious, not fully studied organ in human body BRAIN.
Our muscles work a bit different, then the motor devices we have now. weight, power, precision...still hard for a machine, even with a price as big as Tesla for one prototype.
@@danielchoritz1903
But still I feel neurones are highly complicated than muscles and tendons.
You need about halve a brain and the whole nervous system to control your body. Compared to that a interface seems rather simple.
what is that short key under the Microphone. lol. You two are amazing. Thanks for all you two do.
Usually Lex Fridman is the awkward-but-overly-intelligent one...
LOL
It is far more awkward commenting that type of stuff faceless
It’s the beard it brings out the animal in lex lmfao
@@suisumainichi Yes, good catch. I have achieved more awkwardness levels than both of those guys put together. 💪
Intelligent? Let’s say above average 😅
When he explains that safety is a function of alignment to truthfulness I think it is slightly different. I personally believe that it should aligned with factfulness. The truth depending on the perspective or opinion carry a connotation of being good or bad. Whereas facts are simply stating a immutable non opionated conclusion
He’s the world’s best engineer. Period.
No hes the worlds best salesman
The Cybertruck sux
He's a scumbag
11:27 Three years from now…
Elon: Can you solve climate change? Use my engineering principles to solve this problem.
AI: Ah, the main issue seems to be humanity. I will move forward to delete this as per you request.
Elon: Big oof..
Elon looks... tired. I hope he's taking care of himself. Protect this man at all costs
unlike most lazy useless scumbags he actually works hard every day
This was super fascinating!
Thank you for this, Lex!
why dont they use ai to design a better power grid
They probably do already.
They probably do. Implementation is a challenge AI cannot solve now
As for now, we still haven't complettly figured out the inner workings of LLMs, but sure giving more compute, more abilities, internet access for agents and what not, that sure carries no risks at all, which seemingly doesn't matter in the persuing of winning.
Same with hammers and cups. Their molecular structures are still mysterious as is Space, Time and gravity.
@@Art-is-craft well differance is, that they are less likely to cause an uprising or otherwise fatal accurance of mass deaths than a system used on large scale that may only later become faulty. Like say due to a bad patching process as recently witnessed with Microsofts OS Windows and that security company.
It is on a different abstraction level, physics, chemistry, biology, technology ... maybe AI neurology will be a subject for studies more and more. Point being, there are ways to understand the underlaying principles i am quite certain, it just will take time to find and proof them. Even then though, depending on the advances meanwhile are made and i think there is equall effort as well as chance at times that lead to those advances, objectivly being able to know and think the same as others do was never something we could achiev between the same species as in us humans. We may find the ruling principles, but truely understanding then in the end where this leads to ... a different type of entity and intelligence, that seems like a big ask anyways.
So if we can not rely on that happening, the alignment and control of an ever increasing intelligence seems hardly even worth considering before that.
Therefor i would guess we are in for a wild ride.
Agi can’t happen until we figure out gi.
Mind Begs the Question:
▪︎If those who operate,support Apartheid
▪︎Possess AI,Digital Money,Nukes
▪︎Likely to be used for - Good or Evil?
@@HumanBeingsRThinkingBeings
Apparently good or evil depends on the interpretation from a doer or a victim
@@HumanBeingsRThinkingBeingsits the jews
- ye
The system-wide RDMA is quite something. IIRC Frontier has that within a compute node with 4 h100s, which is neat and Fugaku, while not RDMA took pains to have 6D interconnects.
Elon’s engine is quite something.
90% of us are getting deleted lol
Lololol 💯💯
nah, more like 99%
Thanks for sharing the conversation.
elon is very humble. he knows his competition is not to be joked with
Unless it comes to politics
You should lookup the meaning of humble in the dictionary
@@paulm3969 I was more speaking to the second sentence in that comment Paul
What a cool convo
Diversity is the new religion. And in the name of religion anything can be sacrificed. Truth, first.
I am unsure, but isn't removing ideological bias also a subjectiv task in itself and thereby contradicting.
I further assume there is no living being able to hold itself in its judgment to Kant's imperativ, but always have at the very least its own perspective bias, not being able to be objective even though if trying to be.
The statistical amalgamation that are training data sets, some may see as the closest approximation to Kants imperative, but seeing that certain biases then spill over ignoring huge sections of people at best, activly discriminating at worst ... i may give a point to his 5 point step program to avoid unnessary functions or maybe at some point stumble upon better ordering mechanisms.
How about first actually delivering full self driving before making any other promises.
How about you worry about your OWN money and the way you spend it, and let him do with what he will with his MONEY..Fair???
@@1flash3571 nooo daddy elon 😭
@@1flash3571 bro is in love with elon musk
@@1flash3571hahahahsha
@@1flash3571 no, he is right
Create tiers of importance. And lines not to cross in order to accomplish a goal.
The leaders right now are Boston dynamics, openAI and Nvidia.
Toyota, KUKA, Miso, Philips Medical, Hanson, Shadow Robot.
There are at least 2 dozen companies holding very strong know-how about humanoid systems like the hand. And most of them you haven't even heard of even if you are working in the field.
Regards, thank you.
Elon: AI training data center lead to high stress on the electric grid and black outs.
Elon: Let’s build more AI training data centers 🎉
I hope that when I watch the whole interview there is mention of the 3 laws of robotics. Ive thought many years on those and still can not find a flaw in them.
How long until I can get my robot slave to do all my housework
Or have sex with it?
Add at least 20 years to whatever date Elon says
I'd say 10 years. Many companies have already unveiled robots that do most household chores with demos. They just need to be trained more extensively before being fully released.
2 years. But you won't be able to afford it.
The first 10 years of any product like that will be full of injuries, amputations and deaths.
Just don't be in the room when it is running.
Is there a possibility to invest?
Is this just his public persona image, I wonder?
public personals tells what people want to hear, he tells things as they are with great understanding of any topic he speaks of. Too much of a meme lord for basic personas, he truly believes being himself is allowing him to win so that's what he does (and it really is working tho lol)
The sensory control distortion always _changes_ too. Before I was typing this... my _elbows_ were hightently distorted, even above my face and fingers. They were manipulating the arms on my swivel chair.
Thanks. Elon and Lex for "making" me notice 😃
So one of the most innovative companies the world has ever known is struggling to reengineer the human hand by “intelligent design” yet we teach our kids the actual human hand evolved thru random genetic mutations.
They have been friends for quite awhile. We missed their whistling tune today.
deletegineering
Aspiring to the "truth" in our litigious society is quite the challenge, especially when one can't afford to be sued and/or cancelled...
Every model xai has put out has been subpar. Vastly undertrained. Haven’t solved any of the inherent limitations. Pure scale isn’t all you need. This is pure ego and cap.
He’s gonna cap for the clout but those who know what’s up, understand that OpenAI have likely already reached some form of AGI.
You all understand the headstart the other companies had, right?
I wouldn't write them off. The company was hobbled together very, very quickly and the models they've released so far weren't trained with that much compute. I think k we'll have a clearer picture of things when Grok 2 and 3 come out later this year.
@@MrSlyBadger stop it
You don't even know the word inherent..... and you think you know what's going on at Tesla?? 😂
Power fluctuation... variable frequency drives needs to be tuned
this is not zero sum game. elon sometimes is incredibly out of phase. he wants to win or to prove to everybody, that he is megasuper duper smart, that his solution is the best or his vision is the right one, while this is only showing off nothing more.
You can still lose though. Win might just mean be amongst the winners and not the losers.
He is deeply insecure.
@@PentUpPentatonicsdoubt it
@PentUpPentatonics I think you’re right. He hasn’t found the true way. However, he is on the right path when he mentions “the truth.”
Curious what his take is on Palantir
xAI is not even a competitor lmao. I forgot about them.
Not yet
@@eduardtarniceriu102 Not ever
@@awsmith1007 lmao keep underestimating
@@awsmith1007 i bet youre one of those guys who was mocking tesla back in 2010.
Both of these men are incredibly awesome humans!
CON
MAN
Yep, but Lex not Elon@@Techtalk2030
Snake oil salesman
The sensory tactile will not be replaced
A funny thing happened on the way to Elon figuring out how robots should be built... he learned that G_d is real.
Wow ….
It’s interesting to hear you guy talk ❤️
13:01 #4 any given thing can be done faster
24:01 there’s so much ai polluted material in just searching for things
24:20 “is this data most likely correct?”
Little Finger-What else would you pick your nose with?