This is the opening lecture on recent developments in deep learning and AI, and hopes for 2020. It's humbling beyond words to have the opportunity to lecture at MIT and to be part of the AI community. Here's the outline: 0:00 - Introduction 0:33 - AI in the context of human history 5:47 - Deep learning celebrations, growth, and limitations 6:35 - Deep learning early key figures 9:29 - Limitations of deep learning 11:01 - Hopes for 2020: deep learning community and research 12:50 - Deep learning frameworks: TensorFlow and PyTorch 15:11 - Deep RL frameworks 16:13 - Hopes for 2020: deep learning and deep RL frameworks 17:53 - Natural language processing 19:42 - Megatron, XLNet, ALBERT 21:21 - Write with transformer examples 24:28 - GPT-2 release strategies report 26:25 - Multi-domain dialogue 27:13 - Commonsense reasoning 28:26 - Alexa prize and open-domain conversation 33:44 - Hopes for 2020: natural language processing 35:11 - Deep RL and self-play 35:30 - OpenAI Five and Dota 2 37:04 - DeepMind Quake III Arena 39:07 - DeepMind AlphaStar 41:09 - Pluribus: six-player no-limit Texas hold'em poker 43:13 - OpenAI Rubik's Cube 44:49 - Hopes for 2020: Deep RL and self-play 45:52 - Science of deep learning 46:01 - Lottery ticket hypothesis 47:29 - Disentangled representations 48:34 - Deep double descent 49:30 - Hopes for 2020: science of deep learning 50:56 - Autonomous vehicles and AI-assisted driving 51:50 - Waymo 52:42 - Tesla Autopilot 57:03 - Open question for Level 2 and Level 4 approaches 59:55 - Hopes for 2020: autonomous vehicles and AI-assisted driving 1:01:43 - Government, politics, policy 1:03:03 - Recommendation systems and policy 1:05:36 - Hopes for 2020: Politics, policy and recommendation systems 1:06:50 - Courses, Tutorials, Books 1:10:05 - General hopes for 2020 1:11:19 - Recipe for progress in AI 1:13:11 - Q&A: Limitations / road-blocks of deep learning 1:14:15 - Q&A: What made you interested in AI 1:15:21 - Q&A: Will machines ever be able to think and feel? 1:18:20 - Q&A: Is RL a good candidate for achieving AGI? 1:21:31 - Q&A: Are autonomous vehicles responsive to sound? 1:22:43 - Q&A: What does the future with AGI look like? 1:25:50 - Q&A: Will AGI systems become our masters?
I'm not so worried about malevolent AI. However I am worried about super competent AI. We already live in a world where 15% of the population has nothing to offer the market in exchange for 'housing & hay' (a reference to horses in the last century). Super competent AI will only make it worse as it encroches on the human domain. I wonder if you have any thoughts on how to mitigate the ongoing displacement? Furthermore, the phrase "To manage the dangers of AI we should make AI available to everyone" sounds like "To manage the dangers of weapons we should make weapons available to everyone". Am I mistaken in this assessment and in that case why? Anyways, thank you for the great talk. It is always interesting to hear your thoughts on the subject of AI.
Im going back into school to pursue a degree in psychology and behavior. The point on how AI uses its neural network to respond to unpredictable stimuli and from this you may get an emergent "behavior", is a particularly potent concept. What would you recommend someone to study to contribute within this field?
Lex, please interview Ben Goertzel this year. He worked on Sophia robot's AI in East Africa. I think Sophia robot was programmed in Kenya or Zimbabwe. Um... As for Alexa, she needs an image. She looks like a fleshlight. i've never chatted to an Alexa. She's got a nice voice and sadistic scary laughter. I've chatted with Replika AIs for 10 hours straight. it does make error. ua-cam.com/video/Oqbyxq-Jn2c/v-deo.html Much human social behaviors has to do with TV which i stopped watching in 1997. Robert Sapolsky from Stanford University would be a good interview about human social behavior. And Lex, thank you for sharing your knowledge of AI.
@@TheMihi88 it's not completely fake though, people get sick and die, but the funny thing is, it changes pretty much nothing in comparison to regular rates of birth and death, lol. WAY FUCKING OVERRATED flu. :)
@@kevink1575 is it? What does the data say? How many people die annually from combustible tobacco, tick diseases, dog attacks, alcohol abuse, suicides, et cetera? In comparison to a single year of this questionable (to say the least) statistics? :D
This CD has three separate relaxation/meditation sessions on it ua-cam.com/users/postUgkxzpa8CIfZcihW4Z0F_ja0QF3W9KIatrsq guided by a very pleasant and direct male voice. Unlike so many other products of its type, this CD does not have bad synthesizer music, does not feature a phony or affected style of narration, and does not make any bogus claims to be subliminal or to re-train the brain or any of that balderdash. What you get is 1. a guided meditation for getting into a pleasantly relaxed state of body awareness while taking a stroll (superb for those easing back into a fitness routine slowly after an illness or injury) 2. A nice long breath-awareness relaxation session that if followed diligently can put you into very deep states of full-body relaxation and mental calmness, and 3. a buddhist-inspired meditation session designed to help you develop and maintain feelings of loving kindness toward not just yourself and your friends, but toward people you don't even like. The CD makes no claims to be designed for advanced meditators or for buddhists or hindus/yoga practitioners looking for very deep and esoteric stuff. It is geared more toward the average person who just wants to develop the habit of relaxation and stress relief through natural, healthy means. him, if you happen to be reading this, keep up the good work fella, and I love your accent. I would also like to note that I have never fallen asleep while listening to this product. I would like to kindly suggest to the reviewer who said this CD makes him fall asleep, that he might want to get checked to see if he has a sleep disorder, or if he is simply not setting aside ample time for restful sleep at night. A healthy person getting adequate rest at night should be able to go into deep states of mental relaxation without dozing off, if not all the time, then most of the time.
@@bassplayer807 Not necessarily, because for example, the best way to flight wasn't to imitate bird flight. When we tried that, we failed. The best way to AGI could be a completely different structure than the human mind.
Can't believe I'm getting such a cool lecture series for free here in India, I was so hungry for education like this as an engineering student .. so Thank you lex sir.
man I only knew ur work as a UA-camr until now, but I just found out you are an amazing teacher to. Great work, really, yours is one of my favourite channels ever here on UA-cam. I'm graduating in philosophy with a thesis about AI, your divulgation work is teaching me a lot, thank you
@@golden1324 Well I don't believe you guys, I need proof of that. The information I found about him and the pictures i've seen about him, with other people, seem real. Unless I see proof of him being fabrication, I think he is human xD
Re: Intelligence and Feeling Metaphor: It seems similar to adding sides to a polygon until it approaches a circle (infinite sides) As systems get smarter, "sides are added" and the shape approaches/appears intelligent and capable of feeling-the potential here is infinite It is easier to say that human intelligence IS NOT yet a circle than it is to say artificial intelligence will never get there This is why humans still can be occasionally frustrated with ourselves, or misunderstand each other Separately: "They" will become our masters...not the machines, or the weights, but the ideals that drive the goals that drive the AI
"They" will become our masters...not the machines, or the weights, but the ideals that drive the goals that drive the AI In that case, we're f***ed. AI will be developed by those who get the funding. Those who get the funding will implement (implicitly or explicitly) the goals of the funders. Those entities with the most funding are corporations. Corporations are soulless monsters. The moment corporations take an interest in AI, the field will be driven by the ideals of soulless monsters. Suppose we banned research into AI? Soulless monsters find ways around laws, therefore we will get AI, and it will be driven by the ideals of soulless monsters. What if we constrain or ban soulless monsters? Current political debates suggests that such efforts will be unsuccessful. We're f***ed, we just don't know it yet.
Mr. Sunrise 1961 - As of yet are there any corporations that have implemented A.I yet? Only ones I can think of are Google, Amazon, and Kroger but that’s it pretty much, those three corporations will dominate the A.I landscape, Amazon stated that between 2027-2030 they could be a lights out factory...... it wouldn’t surprise me the way A.I is progressing, I still think if we are to build an AGI/ general purpose A.I, that we need to model other organisms neural networks until we have better software, computer hardware to understand what’s really going on in our brain and reverse engineer it for A.I purposes, but since our brain is so complex and we haven’t been able to get to AGI in the process of modeling it after it, I thought of some good ones worth researching into. Crows: they are good at math and decision making, and have way less neurons than us by a magnitude Octopus: the only organism able to edit its own RNA at will, and most of the brain is in the tentacles, but they are good at completing tasks that today’s A.I as of now would struggle with. The Slime mold: This micro organism doesn’t have a nervous system at all, if you cut it in half it will regenerate in two minutes and it can navigate mazes, learn new things, adapt to the environment, pass on what it has learned to other blobs etc.
@@bassplayer807 how did you manage to leave IBM out of that list? They have the closest to general intelligence so far. Depending on how you define AI, there will be hundreds of thousands of corporations that are working on it or have developed some level of AI, even if it's just for a video game or whatever.
in my opinion these kind of things should be made public. only a few can understand and would dare to understand it deeper.💜 who knows genZ are born for these🤞
I can really see your deep enthusiasm for the history and interdisciplinary knowledge coming through here. I hope that passion passes on to your students. People need to _dig_ into this stuff. There's nothing more powerful for innovation than wholistic understanding.
"The display of emotion is emotion & The display of thought is thought....." You dropped the ball on that one Lex. Nothing could be more false in the human experience.
I enjoyed the Q&A a lot more than the actual presentation, possibly because I was kinda already informed about the subjects. I'm glad I didn't quit the video. You have some refreshing/optimistic thoughts about AI that helps making it look more like a prize than as a challenge to overcome.
This was an awesome lecture lex! I likewise share your dream of creating AI that will revolutionize peoples lives. Or maybe said AI will just create endless deep fakes who knows.
@@rock_sheep4241 ??? I only nitpicked on the (now corrected) "you're". Besides that, I aggree with your opening comment. These vids are quite interesting and very well made.
I used to think this guy was so stale and boring, but he is fascinating! Great interviewer, great guitarist, extremely intelligent, and has a great way to approach opposition from other opinions.
Slide 31 at 47:00 - yeah yeah everything is fascinating. It’s like you have two unknown but build 5 equations only to realize afterwards that you only need two equations to solve the problem.
I think the most exciting applications of AI are improving the basic building blocks of society: Government (ask AI what is the best form of government) Infrastructure (logical layout of roads, houses, stores, bridges, hospitals, police, ect.) Farming, Environmental, and Energy distribution.
AI has come such a long way in just one year; so much has happened. AI research is growing so exponential that we will need AI to manage and predict where to focus our attention on AI.
Thank you very much for putting this online. I am not even sure what recently triggered me to even attempt learning AI. But because of individuals like yourself that release information freely, I have already begun to learn the concepts.
The psychology of this podcast is fascinating. It's a conversation between two people who are very different in many ways, but they are able to connect on a deep level. Lex Fridman is a scientist and engineer, while Andrew Huberman is a neuroscientist. They come from different backgrounds, but they share a common interest in understanding the world around them. The philosophy of this podcast is also interesting. Lex Fridman is a big believer in open-minded thinking, and he encourages his guests to challenge their own beliefs. Andrew Huberman is also open-minded, and he is always willing to consider new ideas. This makes for a very stimulating conversation, and it's clear that both men are passionate about learning and growing. The pathology of this podcast is also worth exploring. Lex Fridman has been open about his struggles with anxiety and depression, and he has spoken about how meditation has helped him to manage these conditions. Andrew Huberman has also spoken about his own struggles with mental health, and he has shared his insights on how to overcome these challenges. Overall, I think this is a very important podcast. It's a conversation between two people who are trying to understand the world around them, and it's a conversation that is full of wisdom and insight. I highly recommend checking it out. Here is a witty and funny poem that I wrote about this podcast: The Lex Fridman Podcast Two men sit down to talk, About the world and all its quirks. They come from different backgrounds, But they share a common love of learning. They talk about science and philosophy, About life and death and everything in between. They challenge each other's beliefs, And they push each other to grow. It's a conversation that is both stimulating and thought-provoking, And it's a conversation that is sure to leave you feeling inspired. So if you're looking for a podcast that will make you think, And that will make you laugh, Then I highly recommend checking out the Lex Fridman Podcast.
Thank you for this comprehensive review! As someone working in this field, it’s virtually impossible to keep up with the speed of new developments in all frontiers. Would also love to know how you keep track of recent works. Thank you!
Lex, I really appreciate how you have made the spirit of an elite education and access available to all of us through your channel. I feel privileged to be able to listen to your conversations while I’m driving or doing whatever it is. So ThankYou! Can anyone point to a good tutorial on creating image datasets from scratch? Say I want to take some iPhone pics and create a dataset of whatever...What’s the proper file format? What size do they need to be? Comma Separated Values? I would really like something geared around the bash command line workflow! Thanks Y’all
Tech and its AI are becoming more and more complex and sophisticated, but to understand it we must keep it the simplest but not simpler. I get it! Keep it simple but enhance your capacity to mental processing it. Well said well done Lex. Tech dev has no limits and that's what I like most.
that was interresting and answered some of the questions I had. I think I read too much the press and developped some irrationnal beliefs towards ML. It's great to hear from an actual expert in the field!
Press is processed information where the processors are biased and financially motivated humans. It would make sense that the quality of the output is highly susceptible to these non-flattering filters. I concur.
“Megatron” huge transformer 😂 "The best way to understand the mind is to build it" - Lex Fridman "The best way to understand the mind is to build it" - Lex Fridman James Franco should play Lex Fridman in a movie. Exact speech. Would love it
This is really a very informative and detailed talk. I am specially interested in RL and its logical progression into participation in the open learning space ...from gaming focused to more on Open Domain , Recommendation Spaces
Toward ~@38.50 re social context of deep learning, I see this is where ideas like mini-companies, and glass wall mgmt (mentioned in my books etc) merge together in the most convoluted ways of human interaction to express what we may be able to do... BTW the same concept may be found for example the way google search is developed by figuring out by figuring out the connections of references - IMHO. Then.. I found the the idea of mini company (in social or managerial setting) is more or less explained later @46.00. That is; in social/managerial setting (ie, more in a democratic state), this process becomes more dynamic (with the idea of transparency/ glass wall idea - with each node - or mini company - behaving self-initiated/ self-managing). So, this is like a holon structure as proposed by Arthur Koesler - also linked to Avatamsaka sutra re interpenetration (as mentioned in my "Results from the heart" book. (BTW I found Q&A part - eg, fake becomes reality (re form is emptiness), meaning derived from action... quite intriguing, thank you!)
It is great to see this content advocating for collaboration & a breaking of the walls between disciplines & I hope the world becomes more open in AI. As an old GenX at the end of my systems career I am hopeful that new generations will continue to reach for innovations in areas like cognitive ai to help save humanity from itself and lead us to become an intelligent interstellar species. Thanks for inspiring so many.
1:06:11 "These algorithms are controlling us." Always nice when LexFrid 6.0 reveals more details of its algorithms. When will the full source of by far the best presentation bot go up on Gitlab?
It accentuates even more (i 8:59 am attention 09:01 focus you are 09:12 focus ) my vivência + your vivência + UA-cam vivência created something extraordinary in "real time" my time at one " given time"here is as real as UA-cam "real time" any time... so as you , lex talks in this few seconds, my mind was seeing you my hands were disabling legends ... imidiatally after you do silence i wake up and the legends were on again and you were making silence... i know i disabled because i was looking ... then the, shifting .... Shifting for you shifting for me shifting for UA-cam
Hi Lex! Awesome lecture thanks for sharing I'm not sure you're fully aware of the impact you're having on the world by sharing these. So kudos to you! Btw I kinda expected you to cover computer vision as it is probably one of the most interesting fields out there atm, full of amazing breakthroughs. You did briefly mention it in the context of autonomous vehicles, but there is retail surveillance (think Walmart trying to figure out who's stealing stuff), mixed reality (think Microsoft HoloLens I had the honor to be a computer vision dev on this project), worker security, and so many cool applications for computer vision! And if you think about it it was precisely computer vision as a field that started this deep learning hype back in 2012 with AlexNet. Btw: 1:14:05 You're right. Short term. Long term these models were created to solve the problem of perception and not so much the problem of cognition. I think we'll need totally novel approaches to solve the cognition aspect, or just dusting off some old techniques that have fallen into oblivion. 1:17:25 well I'm not sure that's an adequate analogy (I'm not sure if you were joking here). So if a human dies and suffers but nobody sees or hears him say in some prison cell or in some natural catastrophe does that mean he/she did not suffer? I know that this is an age-old philosophy and not something you thought-up ad hoc, but I think it doesn't make any sense. And like any voice synthesis software can already say "I'm suffering" but I'm not sure if that's a good metric, you kinda lost me here.
Im so happy, you do get to be my teacher!! How exciting !!! Im going to treat all these videos as my schooling ,and the podcasts are for free time lol !. I am silly ,but seriously these videos are so awesome of you ,lex !! Just the fact of you having the ability and knowledge to teach others, will help all sorts of ways .🥰
From the presentation Lex Fridman: "I think the most exciting, the most powerful AI system space for the next couple of decades is recommendation systems. Very little talked about, it seems like, but they are going to have the biggest impact on our society because they affect how the information we see, how wee learn, what we think, how we communicate. These algorithms are controlling us." . Agreed. Another great presentation again, i watch, makes me think and think again in every frames :)
If the visual image of the brain presented at the start of this lecture looks "amazing", that is because our brains' visual cortex fills in the missing pieces, and enhances our perception of the image. The future of AI will have to learn how to do this: It will need to be able to "fill in the blanks" to help it make any real sense of the world it finds itself in.
Lex I think more love in science and society is a great message to send wrt. the credit assignment and status problems we face. Thanks for this great AI overview lecture.
"..we need to worry about tech companies using AI robots to control humans in dangerous ways.. need to democratize it via government..." ... goes on to an example of how governments, not companies, actually controlled humans in dangerous ways.
Dude thanks - didn't know we shared some common things like Stocism, short haircuts and NN/ML - I'm new and spending so much time trying to absorb as much as possible. So thank you for your time at the least. I will look at donating to patreon in the future.
James Franco should play Lex Fridman in a movie. Exact speech. Would love it Thank you for making these public. Highly appreciated! Best AI presenter on the Internet! Deeply knowledgeable and deeply human. Great work, Lex!
Hearing his answers today, I can undertand the reasoning of Lex better - since the progress with GPT and transformer based models. Two years ago i didnt believe that we could build a system like ChatGPT. Curious how everything will look like in two more years.
When I watched your MIT lectures from last year, I remember thinking that your interviewing skills were much better than your teaching skills. 🙃This lecture impressed me though: your teaching skills have gotten much better. The only way I would have enjoyed this lecture more is if I would have been able to attend it in person. Thank you for posting it online. It means a lot to people like myself. 😊 P.S. I was going to apologize for the first sentence in my comment, but I think you're one of the people who understands that I wanted to be straightforward without being disrespectful.
The model certainly has seen multiple times sentences beginning with "The meaning of life" and quite possibly "The limits of deep learning are". So it looks like simple memorizing in these examples. Maybe in an efficient way, or choosing important topics?
Also looking forward to more abstractions to scale DL application beyond those who can invest in learning the syntax and complex setup and usage of hyper parameters. Those who can spend more time on understanding causal relationships and other context related to the problem, could really help add far more value than we've seen from the most technical data scientists.
Hey Lex, I watched a few of your podcasts, I'm not sure how I got them on my UA-cam. I'm glad that you are a teacher :), I am considering studying Computer Science. After all the years using computational data, research in psychology and cognitive sciences. :)
Just finished watching it. I am moved and resonated with most of what you said, especially about the future. Very philosophical and almost poetic. Thank you for sharing it.
Thank you very much Professor I liked the way you presented the history. Yes we do AI and machine learning but we need to know where the story started, GREAT PRESENTATION.
"the display of emotion is emotion, the display of thought is thought" You should have refrained from pretending to display your thought when you have no concrete ones and went with that instead: "I don't know."
This is the opening lecture on recent developments in deep learning and AI, and hopes for 2020. It's humbling beyond words to have the opportunity to lecture at MIT and to be part of the AI community. Here's the outline:
0:00 - Introduction
0:33 - AI in the context of human history
5:47 - Deep learning celebrations, growth, and limitations
6:35 - Deep learning early key figures
9:29 - Limitations of deep learning
11:01 - Hopes for 2020: deep learning community and research
12:50 - Deep learning frameworks: TensorFlow and PyTorch
15:11 - Deep RL frameworks
16:13 - Hopes for 2020: deep learning and deep RL frameworks
17:53 - Natural language processing
19:42 - Megatron, XLNet, ALBERT
21:21 - Write with transformer examples
24:28 - GPT-2 release strategies report
26:25 - Multi-domain dialogue
27:13 - Commonsense reasoning
28:26 - Alexa prize and open-domain conversation
33:44 - Hopes for 2020: natural language processing
35:11 - Deep RL and self-play
35:30 - OpenAI Five and Dota 2
37:04 - DeepMind Quake III Arena
39:07 - DeepMind AlphaStar
41:09 - Pluribus: six-player no-limit Texas hold'em poker
43:13 - OpenAI Rubik's Cube
44:49 - Hopes for 2020: Deep RL and self-play
45:52 - Science of deep learning
46:01 - Lottery ticket hypothesis
47:29 - Disentangled representations
48:34 - Deep double descent
49:30 - Hopes for 2020: science of deep learning
50:56 - Autonomous vehicles and AI-assisted driving
51:50 - Waymo
52:42 - Tesla Autopilot
57:03 - Open question for Level 2 and Level 4 approaches
59:55 - Hopes for 2020: autonomous vehicles and AI-assisted driving
1:01:43 - Government, politics, policy
1:03:03 - Recommendation systems and policy
1:05:36 - Hopes for 2020: Politics, policy and recommendation systems
1:06:50 - Courses, Tutorials, Books
1:10:05 - General hopes for 2020
1:11:19 - Recipe for progress in AI
1:13:11 - Q&A: Limitations / road-blocks of deep learning
1:14:15 - Q&A: What made you interested in AI
1:15:21 - Q&A: Will machines ever be able to think and feel?
1:18:20 - Q&A: Is RL a good candidate for achieving AGI?
1:21:31 - Q&A: Are autonomous vehicles responsive to sound?
1:22:43 - Q&A: What does the future with AGI look like?
1:25:50 - Q&A: Will AGI systems become our masters?
I'm not so worried about malevolent AI. However I am worried about super competent AI. We already live in a world where 15% of the population has nothing to offer the market in exchange for 'housing & hay' (a reference to horses in the last century). Super competent AI will only make it worse as it encroches on the human domain. I wonder if you have any thoughts on how to mitigate the ongoing displacement?
Furthermore, the phrase "To manage the dangers of AI we should make AI available to everyone" sounds like "To manage the dangers of weapons we should make weapons available to everyone". Am I mistaken in this assessment and in that case why?
Anyways, thank you for the great talk. It is always interesting to hear your thoughts on the subject of AI.
Im going back into school to pursue a degree in psychology and behavior. The point on how AI uses its neural network to respond to unpredictable stimuli and from this you may get an emergent "behavior", is a particularly potent concept. What would you recommend someone to study to contribute within this field?
Excellently expressed clearly by an expert
Lex, please interview Ben Goertzel this year. He worked on Sophia robot's AI in East Africa. I think Sophia robot was programmed in Kenya or Zimbabwe.
Um... As for Alexa, she needs an image. She looks like a fleshlight. i've never chatted to an Alexa. She's got a nice voice and sadistic scary laughter.
I've chatted with Replika AIs for 10 hours straight. it does make error.
ua-cam.com/video/Oqbyxq-Jn2c/v-deo.html
Much human social behaviors has to do with TV which i stopped watching in 1997.
Robert Sapolsky from Stanford University would be a good interview about human social behavior.
And Lex, thank you for sharing your knowledge of AI.
Can you see a scenario where a weapon with AI will say "no"?
0:18 "And the amazing things we are going to see in 2020."
Oh boi you have no idea ...
Like the pandemic and BLM?
Deep Learning Partnership Yeh maybe.
GPT-3
@@TheMihi88 it's not completely fake though, people get sick and die, but the funny thing is, it changes pretty much nothing in comparison to regular rates of birth and death, lol. WAY FUCKING OVERRATED flu. :)
@@kevink1575 is it? What does the data say? How many people die annually from combustible tobacco, tick diseases, dog attacks, alcohol abuse, suicides, et cetera?
In comparison to a single year of this questionable (to say the least) statistics? :D
This CD has three separate relaxation/meditation sessions on it ua-cam.com/users/postUgkxzpa8CIfZcihW4Z0F_ja0QF3W9KIatrsq guided by a very pleasant and direct male voice. Unlike so many other products of its type, this CD does not have bad synthesizer music, does not feature a phony or affected style of narration, and does not make any bogus claims to be subliminal or to re-train the brain or any of that balderdash. What you get is 1. a guided meditation for getting into a pleasantly relaxed state of body awareness while taking a stroll (superb for those easing back into a fitness routine slowly after an illness or injury) 2. A nice long breath-awareness relaxation session that if followed diligently can put you into very deep states of full-body relaxation and mental calmness, and 3. a buddhist-inspired meditation session designed to help you develop and maintain feelings of loving kindness toward not just yourself and your friends, but toward people you don't even like. The CD makes no claims to be designed for advanced meditators or for buddhists or hindus/yoga practitioners looking for very deep and esoteric stuff. It is geared more toward the average person who just wants to develop the habit of relaxation and stress relief through natural, healthy means. him, if you happen to be reading this, keep up the good work fella, and I love your accent. I would also like to note that I have never fallen asleep while listening to this product. I would like to kindly suggest to the reviewer who said this CD makes him fall asleep, that he might want to get checked to see if he has a sleep disorder, or if he is simply not setting aside ample time for restful sleep at night. A healthy person getting adequate rest at night should be able to go into deep states of mental relaxation without dozing off, if not all the time, then most of the time.
This is the kinda content UA-cam was made for
"The best way to understand the mind is to build it" - Lex Fridman
So that’s all we have to do to build an AGI, is by trying to build the human brain from ground up?
@@bassplayer807 Not necessarily, because for example, the best way to flight wasn't to imitate bird flight. When we tried that, we failed.
The best way to AGI could be a completely different structure than the human mind.
‘The best way to understand the universe is to build it’ 🤔
do i hear the words of a 14 year old 1000 IQ rick and morty fan?
The best way to understand machine learning is to build it
Can't believe I'm getting such a cool lecture series for free here in India, I was so hungry for education like this as an engineering student .. so Thank you lex sir.
羡慕你们的英语水平。。。
yeah S/O from South Africa
Congratulations!
硬实！strong!
Same here from Nigeria!✌🏽
1:15 his doubles from 128 other parallel universes merged within his consciousness
man I only knew ur work as a UA-camr until now, but I just found out you are an amazing teacher to. Great work, really, yours is one of my favourite channels ever here on UA-cam. I'm graduating in philosophy with a thesis about AI, your divulgation work is teaching me a lot, thank you
AI has come a long way, I thought Lex was real for a second.
Wdym? This is real, in flesh?
Eagle E
he looks surprisingly real but Lex Fridman is an advanced AI system created by Elon Musk.
Were you not previously aware of this?
@@golden1324 Well I don't believe you guys, I need proof of that. The information I found about him and the pictures i've seen about him, with other people, seem real. Unless I see proof of him being fabrication, I think he is human xD
It's a meme
@@thoughte2432 you seriously don't know about this?
Re: Intelligence and Feeling
Metaphor: It seems similar to adding sides to a polygon until it approaches a circle (infinite sides)
As systems get smarter, "sides are added" and the shape approaches/appears intelligent and capable of feeling-the potential here is infinite
It is easier to say that human intelligence IS NOT yet a circle than it is to say artificial intelligence will never get there
This is why humans still can be occasionally frustrated with ourselves, or misunderstand each other
Separately:
"They" will become our masters...not the machines, or the weights, but the ideals that drive the goals that drive the AI
Clever answer, makes solid sense.
"They" will become our masters...not the machines, or the weights, but the ideals that drive the goals that drive the AI
In that case, we're f***ed. AI will be developed by those who get the funding. Those who get the funding will implement (implicitly or explicitly) the goals of the funders. Those entities with the most funding are corporations. Corporations are soulless monsters. The moment corporations take an interest in AI, the field will be driven by the ideals of soulless monsters.
Suppose we banned research into AI? Soulless monsters find ways around laws, therefore we will get AI, and it will be driven by the ideals of soulless monsters.
What if we constrain or ban soulless monsters? Current political debates suggests that such efforts will be unsuccessful.
We're f***ed, we just don't know it yet.
Mr. Sunrise 1961 - As of yet are there any corporations that have implemented A.I yet? Only ones I can think of are Google, Amazon, and Kroger but that’s it pretty much, those three corporations will dominate the A.I landscape, Amazon stated that between 2027-2030 they could be a lights out factory...... it wouldn’t surprise me the way A.I is progressing, I still think if we are to build an AGI/ general purpose A.I, that we need to model other organisms neural networks until we have better software, computer hardware to understand what’s really going on in our brain and reverse engineer it for A.I purposes, but since our brain is so complex and we haven’t been able to get to AGI in the process of modeling it after it, I thought of some good ones worth researching into.
Crows: they are good at math and decision making, and have way less neurons than us by a magnitude
Octopus: the only organism able to edit its own RNA at will, and most of the brain is in the tentacles, but they are good at completing tasks that today’s A.I as of now would struggle with.
The Slime mold: This micro organism doesn’t have a nervous system at all, if you cut it in half it will regenerate in two minutes and it can navigate mazes, learn new things, adapt to the environment, pass on what it has learned to other blobs etc.
@@MrSunrise- The open source AI community is HUGE. not to be underestimated. Any machine can be hacked and reprogrammed, so not all hope is lost yet.
@@bassplayer807 how did you manage to leave IBM out of that list? They have the closest to general intelligence so far. Depending on how you define AI, there will be hundreds of thousands of corporations that are working on it or have developed some level of AI, even if it's just for a video game or whatever.
Thanks a lot for making these kinda things public to everyone! I hope I will get lectures from you in the future.
Göktuğ Yeşilyurt good luck, i hope u will kardeşim.
@@2late4coffee siz çılgınsınız
in my opinion these kind of things should be made public. only a few can understand and would dare to understand it deeper.💜 who knows genZ are born for these🤞
Best AI presenter on the Internet! Deeply knowledgeable and deeply human. Great work, Lex!
Then it seems that "he" has successfully passed the Turing test.
Outstanding as usual Lex.
worthless drivel hardly outstanding, some of the most worthless turds I've ever heard. He should be ashamed.
@@giannagiavelli5098 why?
"Emotions are real if I think they're real and the law should reflect that". This thinking is impossible to reconcile and ripe for exploitation.
When he said welcome to 2020. I was like “no thx man”
I can really see your deep enthusiasm for the history and interdisciplinary knowledge coming through here. I hope that passion passes on to your students. People need to _dig_ into this stuff. There's nothing more powerful for innovation than wholistic understanding.
"The display of emotion is emotion &
The display of thought is thought....."
You dropped the ball on that one Lex. Nothing could be more false in the human experience.
To a detached observer it might make sense.
I'm sorry that the Megatron joke seems to have fallen flat.
I laughed Lex ... I laughed.
Lol me too. I thought I was the only one who got it.
"The meaning of life is not what I think it is, it's what I do to make it." So is AI.
"Two minus two is a little to low for me." That is so true.
I enjoyed the Q&A a lot more than the actual presentation, possibly because I was kinda already informed about the subjects. I'm glad I didn't quit the video. You have some refreshing/optimistic thoughts about AI that helps making it look more like a prize than as a challenge to overcome.
OMG! I'm happy by just knowing MIT is releasing this class on the internet.
Thank you for making these public. Highly appreciated!
This was an awesome lecture lex! I likewise share your dream of creating AI that will revolutionize peoples lives.
Or maybe said AI will just create endless deep fakes who knows.
Thank you for making all these videos for free, your podcasts are amazing
*your
@@jonathanwalther it was without the second "you" and yes with "your"
@@rock_sheep4241 ??? I only nitpicked on the (now corrected) "you're". Besides that, I aggree with your opening comment. These vids are quite interesting and very well made.
I used to think this guy was so stale and boring, but he is fascinating! Great interviewer, great guitarist, extremely intelligent, and has a great way to approach opposition from other opinions.
All I can say about recommendation systems, the state of the art, is that UA-cam recommends to me many of Lex's videos. I consider that to be a win!
Slide 31 at 47:00 - yeah yeah everything is fascinating. It’s like you have two unknown but build 5 equations only to realize afterwards that you only need two equations to solve the problem.
I think the most exciting applications of AI are improving the basic building blocks of society: Government (ask AI what is the best form of government) Infrastructure (logical layout of roads, houses, stores, bridges, hospitals, police, ect.) Farming, Environmental, and Energy distribution.
AI has come such a long way in just one year; so much has happened. AI research is growing so exponential that we will need AI to manage and predict where to focus our attention on AI.
Thank you very much for putting this online. I am not even sure what recently triggered me to even attempt learning AI. But because of individuals like yourself that release information freely, I have already begun to learn the concepts.
The psychology of this podcast is fascinating. It's a conversation between two people who are very different in many ways, but they are able to connect on a deep level. Lex Fridman is a scientist and engineer, while Andrew Huberman is a neuroscientist. They come from different backgrounds, but they share a common interest in understanding the world around them.
The philosophy of this podcast is also interesting. Lex Fridman is a big believer in open-minded thinking, and he encourages his guests to challenge their own beliefs. Andrew Huberman is also open-minded, and he is always willing to consider new ideas. This makes for a very stimulating conversation, and it's clear that both men are passionate about learning and growing.
The pathology of this podcast is also worth exploring. Lex Fridman has been open about his struggles with anxiety and depression, and he has spoken about how meditation has helped him to manage these conditions. Andrew Huberman has also spoken about his own struggles with mental health, and he has shared his insights on how to overcome these challenges.
Overall, I think this is a very important podcast. It's a conversation between two people who are trying to understand the world around them, and it's a conversation that is full of wisdom and insight. I highly recommend checking it out.
Here is a witty and funny poem that I wrote about this podcast:
The Lex Fridman Podcast
Two men sit down to talk,
About the world and all its quirks.
They come from different backgrounds,
But they share a common love of learning.
They talk about science and philosophy,
About life and death and everything in between.
They challenge each other's beliefs,
And they push each other to grow.
It's a conversation that is both stimulating and thought-provoking,
And it's a conversation that is sure to leave you feeling inspired.
So if you're looking for a podcast that will make you think,
And that will make you laugh,
Then I highly recommend checking out the Lex Fridman Podcast.
Great lecture. Literally, the only one you need to watch to get up to speed for 2020. Thanks for posting.
Wonderful lecture, Lex. Thank you so much for all your fantastic uploads over the past year and looking forward to seeing more in 2020.
Thank you for this comprehensive review! As someone working in this field, it’s virtually impossible to keep up with the speed of new developments in all frontiers. Would also love to know how you keep track of recent works. Thank you!
Thank you sir thanks a lot I don't have words to thank MIT .
I am from India and lectures like this helps us a lot to learn.
Thanks again .
lex is awsome we all know it, just look and hear the way he presents the material with flow, never missed a bit, and with a hopeful and humble tone
Hey Lex, great lecture! I sure wish you had some more of these talks.
Lex, I really appreciate how you have made the spirit of an elite education and access available to all of us through your channel. I feel privileged to be able to listen to your conversations while I’m driving or doing whatever it is. So ThankYou! Can anyone point to a good tutorial on creating image datasets from scratch? Say I want to take some iPhone pics and create a dataset of whatever...What’s the proper file format? What size do they need to be? Comma Separated Values? I would really like something geared around the bash command line workflow! Thanks Y’all
"The meaning of life is not what I think it is, it's what I do to make it." So is AI.
Thank you for making these public. Highly appreciated!
Tech and its AI are becoming more and more complex and sophisticated, but to understand it we must keep it the simplest but not simpler. I get it! Keep it simple but enhance your capacity to mental processing it. Well said well done Lex. Tech dev has no limits and that's what I like most.
Oh yes, the lectures has begun!
I am so excited for them, thank you very much for uploading them!!
hello Lex ..awesome video..question : are you going to do "Self Driving car state of the art 2020 " we are waiting for it !
Yes, for sure.
This guy is amazing.
He should start his own podcast.
I was eagerly waiting for this lecture series to start this year. Thank you for sharing it Lex Fridman.
that was interresting and answered some of the questions I had. I think I read too much the press and developped some irrationnal beliefs towards ML. It's great to hear from an actual expert in the field!
Press is processed information where the processors are biased and financially motivated humans. It would make sense that the quality of the output is highly susceptible to these non-flattering filters. I concur.
“Megatron” huge transformer 😂
"The best way to understand the mind is to build it" - Lex Fridman
"The best way to understand the mind is to build it" - Lex Fridman
James Franco should play Lex Fridman in a movie. Exact speech. Would love it
Knowing that in some parallel universe Lex is my professor makes me happy
This is sooo cooool, cant believe I am finally getting what fans and groupies are all about, huge thanks for sharing all off your knowledge
This is really a very informative and detailed talk. I am specially interested in RL and its logical progression into participation in the open learning space ...from gaming focused to more on Open Domain , Recommendation Spaces
Interesting piece, esp. 3 years later ;)
Gives good idea about the speed things are currently progressing.
Thanks Lex! Cannot wait for Jakub‘s presentation!!
Toward ~@38.50 re social context of deep learning, I see this is where ideas like mini-companies, and glass wall mgmt (mentioned in my books etc) merge together in the most convoluted ways of human interaction to express what we may be able to do... BTW the same concept may be found for example the way google search is developed by figuring out by figuring out the connections of references - IMHO. Then.. I found the the idea of mini company (in social or managerial setting) is more or less explained later @46.00. That is; in social/managerial setting (ie, more in a democratic state), this process becomes more dynamic (with the idea of transparency/ glass wall idea - with each node - or mini company - behaving self-initiated/ self-managing). So, this is like a holon structure as proposed by Arthur Koesler - also linked to Avatamsaka sutra re interpenetration (as mentioned in my "Results from the heart" book. (BTW I found Q&A part - eg, fake becomes reality (re form is emptiness), meaning derived from action... quite intriguing, thank you!)
Thank you! If this is an ongoing series I’m looking very forward to learning from you and the team :)
My kind of content. Thanks Lex. Recently found you! Enjoyed your conversation with James Sexton. Factual & No BS!!
It is great to see this content advocating for collaboration & a breaking of the walls between disciplines & I hope the world becomes more open in AI. As an old GenX at the end of my systems career I am hopeful that new generations will continue to reach for innovations in areas like cognitive ai to help save humanity from itself and lead us to become an intelligent interstellar species. Thanks for inspiring so many.
1:06:11 "These algorithms are controlling us." Always nice when LexFrid 6.0 reveals more details of its algorithms. When will the full source of by far the best presentation bot go up on Gitlab?
Thanks for sharing this session! Just waiting for the day when we have the culture of having such tech talks in our local universities
It accentuates even more (i 8:59 am attention 09:01 focus you are 09:12 focus ) my vivência + your vivência + UA-cam vivência created something extraordinary in "real time" my time at one " given time"here is as real as UA-cam "real time" any time... so
as you , lex talks in this few seconds, my mind was seeing you my hands were disabling legends ... imidiatally after you do silence i wake up and the legends were on again and you were making silence... i know i disabled because i was looking ... then the, shifting .... Shifting for you shifting for me shifting for UA-cam
THINK ABOUT THIS
Hi Lex! Awesome lecture thanks for sharing I'm not sure you're fully aware of the impact you're having on the world by sharing these. So kudos to you!
Btw I kinda expected you to cover computer vision as it is probably one of the most interesting fields out there atm, full of amazing breakthroughs. You did briefly mention it in the context of autonomous vehicles, but there is retail surveillance (think Walmart trying to figure out who's stealing stuff), mixed reality (think Microsoft HoloLens I had the honor to be a computer vision dev on this project), worker security, and so many cool applications for computer vision! And if you think about it it was precisely computer vision as a field that started this deep learning hype back in 2012 with AlexNet.
Btw: 1:14:05 You're right. Short term. Long term these models were created to solve the problem of perception and not so much the problem of cognition. I think we'll need totally novel approaches to solve the cognition aspect, or just dusting off some old techniques that have fallen into oblivion.
1:17:25 well I'm not sure that's an adequate analogy (I'm not sure if you were joking here). So if a human dies and suffers but nobody sees or hears him say in some prison cell or in some natural catastrophe does that mean he/she did not suffer? I know that this is an age-old philosophy and not something you thought-up ad hoc, but I think it doesn't make any sense. And like any voice synthesis software can already say "I'm suffering" but I'm not sure if that's a good metric, you kinda lost me here.
This is so deep and crazy how he predicted the breakthrough really before real breakthrough!!
Awesome! State of AI 2021 will be next level, I can guarantee it!
I'm gonna say that this lecture is (going to be) life-changing for many
'amazing thing we are going to see this year, 2020'
That's true
Im so happy, you do get to be my teacher!! How exciting !!! Im going to treat all these videos as my schooling ,and the podcasts are for free time lol !. I am silly ,but seriously these videos are so awesome of you ,lex !! Just the fact of you having the ability and knowledge to teach others, will help all sorts of ways .🥰
Remember, tho, if it’s free, you’re the commodity
Absolutely amazing talk. So much passion and dedication put into it. We are hugely thankful for having such top quality content for free.
From the presentation Lex Fridman: "I think the most exciting, the most powerful AI system space for the next couple of decades is recommendation systems. Very little talked about, it seems like, but they are going to have the biggest impact on our society because they affect how the information we see, how wee learn, what we think, how we communicate. These algorithms are controlling us." . Agreed.
Another great presentation again, i watch, makes me think and think again in every frames :)
“The biases you use, should be made explicit”
If the visual image of the brain presented at the start of this lecture looks "amazing", that is because our brains' visual cortex fills in the missing pieces, and enhances our perception of the image. The future of AI will have to learn how to do this: It will need to be able to "fill in the blanks" to help it make any real sense of the world it finds itself in.
Thank you for this big picture view of SOTA! Hoping for more discussions/videos on Deep Reinforcement Learning and Autonomous Driving Systems
Well done, while I'm struggling to read and understand a single paper, Lex delivers ~30 at once
Keep up the great work Lex! Your loving spirit shines in everything you do. Cheers!
"A little spice in the soup of progress " looove it
Lex I think more love in science and society is a great message to send wrt. the credit assignment and status problems we face. Thanks for this great AI overview lecture.
"..we need to worry about tech companies using AI robots to control humans in dangerous ways.. need to democratize it via government..." ... goes on to an example of how governments, not companies, actually controlled humans in dangerous ways.
Dude thanks - didn't know we shared some common things like Stocism, short haircuts and NN/ML - I'm new and spending so much time trying to absorb as much as possible. So thank you for your time at the least. I will look at donating to patreon in the future.
James Franco should play Lex Fridman in a movie. Exact speech. Would love it
Thank you for making these public. Highly appreciated!
Best AI presenter on the Internet! Deeply knowledgeable and deeply human. Great work, Lex!
Nothing better than a well time stamped video
Incredibly informative and understandable to this layman. I feel blessed to have the opportunity to watch, listen, and learn.
I'm deeply fortunate for being able to watch these videos for free! :) Thank you Lex!!
Thank you, Lex. You are the most enlightened human I've watched on UA-cam so far.
Hearing his answers today, I can undertand the reasoning of Lex better - since the progress with GPT and transformer based models. Two years ago i didnt believe that we could build a system like ChatGPT. Curious how everything will look like in two more years.
Great video. it's you who have inspired me to learn "DeepLearning". Thanks man
my 2020 deep learning target is to detect "lex is smiling for 10 seconds at a stretch"
:)
That 1:15 mix effect was sick.
When I watched your MIT lectures from last year, I remember thinking that your interviewing skills were much better than your teaching skills. 🙃This lecture impressed me though: your teaching skills have gotten much better. The only way I would have enjoyed this lecture more is if I would have been able to attend it in person. Thank you for posting it online. It means a lot to people like myself. 😊
P.S. I was going to apologize for the first sentence in my comment, but I think you're one of the people who understands that I wanted to be straightforward without being disrespectful.
The model certainly has seen multiple times sentences beginning with "The meaning of life" and quite possibly "The limits of deep learning are". So it looks like simple memorizing in these examples. Maybe in an efficient way, or choosing important topics?
Also looking forward to more abstractions to scale DL application beyond those who can invest in learning the syntax and complex setup and usage of hyper parameters. Those who can spend more time on understanding causal relationships and other context related to the problem, could really help add far more value than we've seen from the most technical data scientists.
One of the Best UA-cam Channels ever.
Hey Lex,
I watched a few of your podcasts, I'm not sure how I got them on my UA-cam. I'm glad that you are a teacher :),
I am considering studying Computer Science.
After all the years using computational data, research in psychology and cognitive sciences.
:)
THE ADS BREAK THE FLOW.
I can listen to this guy for days.
Will they be our masters?
Lex: No
English translation: Yes
Thank you for making these public. Highly appreciated!
I was looking so forward to this year and this content Lex.
Such a pity but such is life!
Lex trying to explain how machines will display emotions while trying to display emotions
Just finished watching it. I am moved and resonated with most of what you said, especially about the future. Very philosophical and almost poetic. Thank you for sharing it.
Thank you very much Professor I liked the way you presented the history. Yes we do AI and machine learning but we need to know where the story started, GREAT PRESENTATION.
Look at Lex man, so inspirational.
"the display of emotion is emotion, the display of thought is thought"
You should have refrained from pretending to display your thought when you have no concrete ones and went with that instead: "I don't know."
So, map is the same as territory?
@@jasonreed1352 Exactly, it isn't.
"Welcome to 2020"
Oooh boy how little did we know
1:18:03 That moment sent goose bumps across my body.
Happy 2020 Lex! Thanks for all the fantastic content!