Fei-Fei Li & Yuval Noah Harari in Conversation - The Coming AI Upheaval
ฝัง
- เผยแพร่เมื่อ 19 พ.ค. 2024
- Watch Yuval Noah Harari speak with Fei-Fei Li, renowned computer scientist and Co-Director of Stanford University's Human-Centered AI Institute -- in a conversation moderated by Nicholas Thompson, WIRED's Editor-in-Chief. The discussion explores big themes and ideas, including ethics in technology, hacking humans, free will, and how to avoid potential dystopian scenarios.
Publication is available under Creative Commons, CC BY-NC-ND 4.0 - creativecommons.org/licenses/....
The event was hosted at Stanford in April 2019, and was jointly sponsored by the university's Humanities Center, McCoy Family Center for Ethics in Society, and the Stanford Institute for Human-Centered Artificial Intelligence (HAI).
Why do I feel that everyone who debates or gets into a conversation with Yuval, be it Zuc or these Stanford faculties or anyone else, completely miss his points and try to advertise what they are doing or their institution.
It is this Naive world, that will actually get everything Yuval states come true ;)
To truly understand what Yuval is saying, they need to put their careers and self adverts aside and look at where we are heading.
He is very deep, and understanding his points require wisdom. Engineers are generally bright and smart but not wise.
@Miroslav Ristić yeah ..🍻
Nitul Das yup ... and hence the talk almost becomes meaningless ... it’s when two minds get to sit down and talk without any prejudices or corporations of their own to defend that it actually seems like a decent talk ... the one between Yuval and Sam Harris was pretty decent.
@Nitul Das Exactly so they are not philosophers or explorers but pawns in a bigger game and ultimately they do what they do for selfish reasons of pride or money and that is what will get us into big trouble they are like tiny children playing with toys they have no idea of the power of
You are right. Most people just don't understand what yuval conveys. They entirely miss the essence
@31:46 Fei-Fei argues that AI is not so omnipotent at the moment to make Yuval's concerns relevant.
However Yuval counters with an concrete example that it indeed is. Fei-Fei is taken aback as is evident from her body language. But she does not accept this verbally.
There are multiple moments like these throughout the talk and if a person at her position can not be humble enough to consider making changes in the ways they view the world, we do have a problem
I appreciate it that you mentioned the body language. Feifei is being defensive of her position. I remember she said she had spent the past 20 years on AI, basically to take it up as her major, in the hope to and take pride in making positive contributions to the future society. She just can't swallow it being perceived as a negative contributor to the world. After all, who would feel fulfilling being perceived that way? I also viewed another clip between Mark Zuckerberg and Yuval Harari. Similarly Mark is being defensive of himself as well. He seems to be in a mode of void or blank in understanding why Harari cited Honduras to make his point. Back in Mark's mind, he was probably asking why is Hoduras relevant at all here. If I would voice my understanding: Yuval is basically saying people with no power or no platform for their voice are at the lowest end of the food chain, like a piece of meat on the kitchen chopping board, contributing their value while passively waiting for horrible consequences happening to them . They are the weakest link in the cycle comprised of data generators (facebook, amazon), data processors and algarithm creators (Feifei. Large corporation..), data benefactors (large corporation, powerful or authoritarian governments, powerful nations), data source (average grassroot powerless individuals). Yuval thinks it's unfair for the Hondurians and wants to evoke equitable protection for them - the data source group. Mark, on the other hand, only sees the damaging consequence if AI falls in the hand of an authoritarian government ie the data benefactor group. Mark seems to think that this group of data source is irrelevant at all in his equation. However, luckily we have Andrew Yang, extremely well-informed of what AI has brought about , whose campaign seems to hear what the data source group has to say and chooses to side with the weakest link by proposing 1000 dollars a month for each adult american. But more as a core solution to this problem, in my opinion, is "AI psychic" to help each individual better know oneself, as suggested by Yuval.
Samsung Cricket she is typical stupid school smart people who lacks deep knowledge of human spirit and history
so true. Yuval’s conversation with Mark Zuckerberg and Li Fei Fei kind of show the attitude of these tech people. They seem to found Yuval’s concerns too exaggerated, in this case she seems to even laugh about them and think he doesnt understand AI like she do... which really is a problem. They need to actually be humble and listen with open heart and not be so obsessed about their technology. But they are the people who grew up being maybe the smartest kid in class... this is gonna be hard
@@ChINNg28 Right, and they have also made a lot of money on the way and they are considered successful, which makes it more difficult for them to listen to a 'philosopher'.
Fei Fei is speaking from inside a process - the system of computer science research. This system has it's own pressures, rules , pressing concerns of the moment and a very strong momentum and direction. Yuval is speaking from the outside. It is very hard to change a system from the outside as the system may not be ready to change. It is like psychoanalysis - a very slow process involving many conversations. This is a start.
I agree with you, but I would add that like psychoanalysis, it not only takes a lot of time, but doesn't really work all that well when people are not very motivated toward change.
Fei Fei was ackward in her jokes and way too emotional. This was not her typical technical AI talks, she should have known better.
She gets to do her research by shilling for funding. Way too biased.
Fei-fei is just brandishing her woke credentials. Do you need to be hired by Stanford to make the comment "can love be hacked"
@@samsun01 well she is an engineer, not a philosopher or historian. And she can't go into technical details with a non technical person or vice versa. So no I don't really think she could've known much better. It's not her field of work.
@@Simon-nl3en That's where you're too infantile to believe technical details cannot be distilled and simplified for the layman to understand. It's actually a litmus test of what a good engineer would do. You need to go out more and learn more things. You're defending someone without knowing what you're defending. You're confused, son.
I think Yuval nailed the key challenge when he said the technologists - and particularly their investors - will not wait for the ethicists, philosophers and historians to catch up. And an escalating AI arms race is a scary prospect.
It was encouraging to hear Fei-Fel acknowledge the challenges. But I would have felt more trust in her claims that the AI community is thinking in "human-centred" ways, had Fei-Fei accepted more ownership for finding solutions to issues that will impact millions of lives in ways that cannot yet be known. Instead, she unconvincingly deflected the questions with (sometimes inappropriate) humour, or by passing the buck to the wider scientific community. AI needs leaders that lead from within, and that have a vision for what Human-Centred AI really means.
A dimension of the debate that I would have liked to have heard more about, is how the voices of society at large might be taken into account. Whilst academics might be the experts, surely some of these societal impacts are questions for democratically elected politicians and ultimately voters.
My main concern remains: for whose benefit is AI being developed? I want to believe that "Human-Centred" AI is more than a re-branding exercise to avoid AI becoming the next GM... and I realise of course that the horse has already bolted.
Thanks to Yuval for adding his much-needed voice of reason to the AI debate.
Whatever Yuval opined, you could easily think all that up yourself and probably better if you contemplate a little in those direction. What Fei Fei understands you can't without the same decades of rigorous training.
Rub one out Piyush.
Well said. Fei Fei showed very clearly she is not one of those leaders. She just cares about her funding. She sounds like an AI, no "love", just hitting those keywords to win credit points.
@@yushpi *robot noises*
I have watched many interviews and listened to podcasts that tackle the vast subject of AI, and it seems almost no one is willing or ready to suggest any answers. I think psychology has to be brought into the discussion. So, for example, Yuval mentioned the strange reappearance of stone walls. I think the psychological answer to this question is simple: it is pushback against the unknown coming; it is pushback against migration, which will increase as the climate warms; it is sheer fear manifesting from a very deep place in our long, long ago past and our amygdala. The question of the walls isn’t really a question about AI, but it is an example of the experts simply throwing out more questions. I know scientists hue closely to the rule about not opining outside their realm of expertise (N.D. Tyson being one scientist exception, who loves to explore all the connected avenues). But if time is of the essence, this just won’t do. I don’t think it’s enough for the experts to ask questions and pat themselves on the back for starting the conversation. The conversation has been going on for some years already. They have to start sharing ideas about solutions. Yuval made one suggestion, the AI sidekick. OK, what else? How do we restrain the corporations, in cahoots with government, from killing every job there is but being an AI technician? That’s just one answer we need, quick.
Thanks Harari!
Very few people who think they know, are able to really comprehend him.
Even Fei Fei or Nick aren’t there. Nock’s habit of interjecting and not listening goes against him.
Most financial drivers in technology need people interacting to keep money rolling in. The effects of this model will drive all other considerations To.the end of the que.
@20:55 If Feifei Li had really been reading Harari's books, she would NOT have asked him to elaborate on this very question: why he fears the AI technologists and biologists come together; this is especially detailed in Homo Deus and he describes convincingly achievable future scenarios with AI technology that can be worrisome. As a computer scientist working in the AI field, I often dismiss some Sci Fi scenarios but I take Harari's words seriously. I'm watching this video after watching the conversation of Zuckerberg with Harari (which at least brought up some great points) and I have to say so far this one is painful to watch (I'm not even half way done)
I continued to watch more but I'm not sure if I can finish---if I got to have dinner with Prof Harari, I'd like to ask "is it painful to talk to these big name technologists who can potentially help address the concerns you are raising but who are talking without hearing, hearing without listening, reading without comprehension?" Is this the best minds the world has to offer now?? I need to see better minds in science and technology with more humility please!
@@permblue probably too late but yes you shouldnt have finsihed liek I did😕
But for those who arw not familiar eith yovals work should atleast watch his part. The other bring almost nothing interesting to the conversation.
This big ego woman would be called a "psycho" in another life. She has no points, just throwing around big words (keywords) to make herself feel adequate in front of Yuval's depth and breath
She is AI Deep Learning professor with a technical curriculum, I wouldn’t have expected her to read Yuval’s books… which is part of the catalyst for conversation, her being here is progress for me.
@@samsun01p
Conversation actually starts at 10:45. I can't stand long-winded introductions...
Long winded. Self congratulatory. Esoteric introductions.
Thank you, it was becoming painful. I'm glad i started reading the comment section.
Noah’s clarity of thought is so much more than others in the discussion. Tinder has already hacked lust, and combined with other things, I am sure love will be hacked soon.
Yuval's point about nuclear weapons enabling the cultural flourishing in the 50s and 60s is absolutely fascinating.
Thank You! the tickets were sold out quickly, glad it is posted now!
The easiest people to manipulate are the people who believe in Free Will, because they think they can't be manipulated. -Yuval Noah Harari
Came to the comments for this
Then it follows the final question which lasts since ever "Who am I?"
So why would those who don't believe in free will be bothered?
This is bullshit. Belief in free will does not mean you think you're invulnerable to manipulation
0pl. P
I’m sure Stanford university has plenty professors and students who understand Yuval. Prof FeiFei Li makes me doubt about the high reputation of the university. Yuval is brilliant.
I know what you mean. I also wasn't impressed with Fei-Fei's contributions to the discussion and thought that Yuval had much more interesting and specific things to say. But that seems to be due simply to the difference between a philosopher/historian and an engineer/scientist. People in the latter category just aren't trained to step back from their own field and to imagine what consequences the subject matter of their study, in this case AI, might produce or will likely produce. They are completely focused on the technical details and, at least generally, though not always, don't think about the larger picture. The irony, of course, is that Fei-Fei was invited to participate in the talk precisely to avoid this problem and to think and talk about the forest rather than the trees. Her performance tells me that, just like Yuval worries, scientists/engineers just have no clue where their work may ultimately take us. They may genuinely believe and say that the consequences matter and should be taken into account, but ultimately they don't see that as their task nor do they have enough time for that.
@@roberthess3405 well articulated Robert... yup .. and one of the worst conversations was the one with Mark Zukerburg. It’s Ironic that Yuval comes from such a good neutral place, and a lot of them can’t see beyond their own institutions, their firms or their specific beliefs.
@@papergaery5257 Thank you, Papergaery. Yes, I agree, so far all of the talks with Yuval I have watched make his conversation partners look out of their depths. That is, to some degree, true even for his talk with Steven Pinker. Pinker is obviously an extremely smart guy with deep insight into the human condition. But he is focused more on the past than the future, and unlike Yuval seems to lack a solid understanding of the current state and potential future consequences of AI. Watch the video, you'll see what I mean.
@@roberthess3405 oh I have seen that one ... and yes ... actually you seem to give him more credit than I did ... though I have loved Steven Pinker, he was totally out of his depths on that one ... it sort of also made me feel that he was one of those old school guys who is close to an end ... and the likes of Yuval have taken over ... but the thing is there are very very few in the Yuval’s progressive framework ... I would probably put Elon Musk, Michael Gladwell and Rory Sutherland in Yuval’s mound, though they are all very different in their own ways :) glad to have this talk Rob.
@@papergaery5257 I am glad it wasn't just me who was surprised by Pinker's views. I have a lot of respect for the man and have read several of his books.
I also agree that there are very few people who can talk about all this stuff at the level of Harari. Certainly most other historians can't. The reason is that Harari has an unusual high level of multi-disciplinary knowledge. To talk about the future of AI and what it may mean for the future of human kind, you have to have in-depth not just of AI, but also psychology (which Pinker does have, of course), medicine, politics, education, philosophy and many other areas. I don't purport to have such multi-disciplinary knowledge. But I can tell when others also don't have it.
Even being a project manager and engineer I can understand Mr. Harari’s point of view much much better and easy. I don’t understand why fi fi doesn’t get it despite working the field and having access to the information that we don’t have. Anyways Yuval rocks again
Bismillah Arif she is one of those stupid school smart people
Exactly. Her visible intelligence does not match her credentials or experience.
Summary: Fei-Fei Li: AI development is a slow, open process with lots of potential benefits for humanity. Harari: AI will change everything forever. We must prepare.
They're both right.
Thank you Mr. Harari for helping us think out of the box. Your questions put us in a different level of approaching to knowledge.
Waiting for Yuval to get on the Joe Rogan podcast. Thumbs up if you'd like to see this!
unbelievable he hasn't been on Joe's podcast!
Joe Rogen... Common they are not on the same level he is a matcho idiot who eats meat...
Why tf is that fat fucker such a big deal
@@adamlee9347 he is pretty stupid but funny
@Steve Kelsey yeah you are right
It's a pleasure to hear Mr. Harari, he has great clarity of thought.
Yuval Noah Harari, I read three of his books and I like them a lot. 👍🏻 Thank you so much.
WOW! the perspective Yuval is revealing about our reality is mind blowing!!!
We need people with real background to interview guys like him. His ideas go far far beyond the surface .
Stratification of Technology ... Major & Minor ... Kind Of ... Difficult To ... Design A University Program ... Since One Size ... Does Not Fit All ... Meaning There Are ... As Many Degrees ... As Human Beings ... 7 Billions Plus ...
Trust In Democracy ...
And ...
MisTrust In Dictatorship ...
That's The Q ...!
Trust In Democracy ...
And ...
MisTrust In Dictatorship ...
That's The Q ...!
Cerebral Electronic Engineering Technology ... On The Table ...
All On One ... One On All ... Meeting ... ?
This feels like an eerie conversation to revisit.
In this context, while sitting next to a top (academic) authority in the space, I’d be more inclined to give Fei Fei the benefit of the doubt and accept Yuval’s near-term projections as over-inflated.
Flash forward to today (June 2023), and the urgency driving Yuval’s POV has remained consistent from Homo Deus, to this panel, to right now.
In his model, corporate interests set the pace on introducing AI to world. In this conversation, possibly naively at this time, Fei Fei believed that Stanford and MIT were in the driver seat and had the luxury of time to work through ethics, human-centricity, etc.
After witnessing the manner in which OpenAI released Chat and, consequently, the “arms race” it launched, I’m much more inclined to think that Yuval’s vision is more aligned to what we can expect than any other I’ve heard.
Would like a conversation between yuval and robert sapolsky
I can host. Colorado kush and micro brews
Wow yes
i am 60 years old and starting to believe it will get crazy in my own life time
1:16:40 KILLS it. As an engineer myself , this is probably one of the most important questions pointed out by Harari that we need to ask ourselves, what happens to the walls between countries and the combination that Harari talks about??? even Fei-Fei could only laugh at her own insignificance to this part of the problem.
Mitra, here is the answer: www.amazon.com/Walled-States-Waning-Sovereignty-Books/dp/1935408089/ref=tmm_hrd_swatch_0?_encoding=UTF8&qid=&sr=
Great talk Yuval and many thanks for sharing. :D
Any chance the audience questions can be posted along the talk next time?
Yuval was trying to work through an argument and Fei-fei’s entire contribution was this is a good conversation and we need to have it, and a few jokes for the audience to feel pleased. If ur reading this skip thro fei-fei’s part. Its useless
Exactly. And worse, right after "we need to have this conversation", it's denial that we should have concerns (e.g. "AI is not as powerful as you think", "we don't need to worry about these for decades). Painful to watch.
She's the Dr Leana Wen of AI. "I mean, gosh, yes there are concerns, gosh, I mean really there are, but everything will be okay!"
As always, it is fascinating to listen Professor Yuval
In case Yuval mentioned that no politician talks about technological desruption, America has one - Andrew Yang.
It surprises me sometimes how far back other politicans are in regards to him. Most of the issues the other candidates speak of are things even Kanye West addressed in songs years ago😂. That isn't to say they don't still exist. However, it seems as if they are unaware of the current even more grave issues we are about to face.
Most of the candidates have been talking about issues that have been there for so long, and to be fair can be fixed with a fair amount of focus. However we have no counter to AI and its displacement of human labour. What can we say about CRISPR and what it eventually means to be human or the great even more apparent Inequality that will exist hence from.
Or even worse the lack of meaning we can readily feel in our societies and the prevalence of depression
@@SAM-ft9jd I'm looking forward to the time when folks on the left figure out CRISPR is going to be used to engineer kids, and then they start arguing for the state to give subsidized access to it for all citizens. Then we can make sure that biologically based inequality is decreased at the same time that we improve the entire next generation. It will accomplish the wildest fantasies of the Eugenics movement at the same time that it decreases inequality.
But yeah, nobody even wants to talk about the issues we will face. The debate will come in its own time, I think.
@@briankaul1201 Good lord imagine the shitshow that's going to ensue if genetic optimization results in a converging phenotype.
@@chahahc There are a bunch of potential problems, and that is a big one. I have no clue how those tools will get regulated. Would you have a lottery system for who could have the best genes to maintain optimal population diversity?
It's not super clear how to think about the ethics of that technology, but the common stance of outright banning it isn't game theoretically stable, nor is it clearly correct ethically. Nick Bostrom has a good article about the ethics of human enhancement. Maybe start there?
But that's a good point. Having a big clone army would be weird.
Bernie Sanders is remarkably aware of it for a 77-year-old man.
Love how yuval at 1:01:00 presents the unbiased ethical question about " who can we trust this with , really" good job Yuval, a very deep question that cant really be answered honestly.
Computer Simulated Atomic Coded Reality literally he was alluding to Trump’s election win and brexit
Harari's views are intriguing for a general audience for its philosophical appeal but were sometimes too broad and lacked nuance. That explained why Fei-fei seemed a bit 'slow' to comprehend his views since she probably has developed immunity to AI+ethics hypes that do not agree with her experience with AI's current limitations & the nature of ongoing AI research. She seems focused on the more ad-hoc and imminent problems like privacy/diversity/bias/regulations on the ground, while Harari wants to offer a coherent philosophical vision that could prepare humanity for the AGI era. Both are important and it’s interesting to see their different approaches and priorities via this talk.
Also historically speaking, it’s often easier to have strong opinions about things that are more far-off and inherently complicated, but most of them were rendered obsolete over-time as new technologies and realities emerged. I guess that’s why Fei-Fei seemed more interested in acting in the short-term than worrying about the long-term.
It is very surprising to me that Fei Fei of all people should have trouble understanding Yuval's arguments considering her roots in China. It's not US (or other democratic nations in Europe) where Yuval's hypothesis is most likely to come true in the beginning. It's countries like China and Russia where we will see most abuse of AI's power. I wonder if she has ever talked about China's abuse of AI technology. If not, is she purposely trying to avoid concerns that countries like China present when it comes to data collection and surveillance of human habits.
I think ironically the reason that Harari's points are so important is situations like this -- the people working in the field have their heads buried in the day to day details, and lose sight of the entire forest. I think Fei-Fei is underestimating how fast AI is going to evolve -- it's not a linear scale. And honestly her responses scared me, because she often didn't seem to appreciate the gravity of Yuval's observations. I honestly think Yuval hit the nail on the head again and again, meanwhile Fei-Fei was struggling to grasp why what he said would be true or why would it matter. Yet ironically, everything Yuval says is incredibly accurate.
doby simson ur on so on point bro
Well said! Yes, AI is going to become AGI eventually, but not next year. Even uber-AI people like Ben Goertzel admit its current limitations. We have some breathing space. Not a lot, but hopefully enough.
Also, if we stick the landing on the short-term problems, the long-term ones will be less of an issue.
We need clear minds on both issues.
@@ShadyRonin not to mention that Harari's thinking of an end point that is not here today, but must be discussed today. What is alarming is the AI arms race where powerful free market tech businesses (e.g. AWS) are warring against powerful totalitarian states (e.g. Chinese Communist Party) where people and governments are rendered slaves to either side.
While I don't fear the CCP as an existential threat to my personal existence as a Westerner, I do fear their central state-serving strategic national plan for AI, and the export of their surveillance technology to other countries around the globe. What I fear from the Western tech companies is being manipulated by AI to a point in which algorithms never allow me/us to see/access things that it/they have not already preselected. I fear the filter affect.
Existentially I fear widespread social manipulation from both sides and widespread social angst and loss of meaning and purpose.
wow excuse me but there is not one serious thought coming from that fei fei character
True.
hhhh
I agree. She talks the obvious.
No. She’s even a bit repugnant
We're doomed
Fei-Fei doesn’t seem to get it. It’s a bit concerning given she’s heading up HAI at Stanford...
Thas the thing technocrats usually lack the historical-humanitarian perspective.
The naivety of HAI professor’s view of human nature is scary. On the other hand, it’s also her brilliance ...
Agreed. Don't let geeks make decisions for the rest of us. Ends in disaster. They have no social intelligence whatsoever.
Her ignorance is astonishing
Peter Xing she just talks about dialogue and diversity and that she’s an engineer. Don’t even get the jokes and Says same things as any teen age “liberal” girl would say. Embarrassing! Stop forcing people in debates by label and identity! It works baaad
Fei-Fei and Yuval have become two of the most respected humans, by me, in my life (I am 72). It brings HOPE in my heart the fact that these issues are so publicaly dinated. They MUST become INTEGRAL PART IN WORLD EDUCATION: CHILDREN AND THEIR PARENTS, AUNTS, ETC. In fact, we should ALL have A PARTICULAR TV PROGRAMME TO DISCUSS IT, and INTERACT WORLDWIDE.
THANK YOU SO, SO MUCH!!!!!!!!
Great questions and talks between the smartest minds always help to open the future vision of audience and listeners. Yuval's prediction maybe already true in some part of human life. When personal data being collected and used, we are possible be under manipulated by certain "subconscious therapy " in order to make decisions that original wasn't ours, or maybe better than we realized(human are short sight often). These data certainly have been commercialized to help business in many ways. I guess society democratic may wish to eliminate those criminal, murdering, abusive minds, train that be rational and fair. by AI. Sexuality may even be manipulated in certain ways and other things. So this does post very much alternative/danger to human life. Data being collected and transferring around, then finally being colonized, it certainly post very much danger of dictatorship when it turns to paranoid and abusive. Bravo, Feifei! You are gorgeous.
I have read both Sapiens and Homo Deus. These books are very enlightening and full of useful information. This discussion touches on topics that all thoughtful people must understand better and also to consider what the future potentially holds for human kind.
It is always easier to raise up an intriguing question than answering it in an intriguing while practical way. As one of the top engineers in this field, Feifei showed us her open mind and practical solutions. She deserves some credit.
Not necessarily... "It is always easier to raise up an intriguing question than answering it in an intriguing while practical way"
The question is much harder than the answer. Once the question is made the answer follows. Without the question you are completely lost.
I disagree with Feifei staying open minded, in fact I don't think she seemed to understand Yuvals points. No clear answers whatsoever.
Her answers are vague to the point of irritating. Answering things like: -Humans have some transcendental feeling of love that is unhackable. Yuval pushes her to specify her answer, as she is the engineer and not the philosopher, she should be concrete. She doesn't answer. Love in the biological sense is hackable.
-She says AI isn't omnipotent and will never be able to understand the complicated world. We are extrapolating. Then she says I can't multiply 2 digital but computers can, therefore AIs explainaility is possible, because AI is very powerful. She contradicts herself with with is and what is not possible with AI. Choosing one or the other when is convenient.
She rushes to answer Yuval like a small child that only cares about being right or manifesting her point, rather than listening, trying to learn, and maybe even change her mind.
She seems to completely dismiss Yuvals shocking points that she probably had never even thought about. Yuval deserved the credit for enlightening us.
I am an engineer and these are the questions I need for ethical work. Not answers based on ignorance for a consumer market governed by manipulating AIs.
TH-cam lets Stanford educate us without the crippling debt 🤓 what a great time to be alive
We shall have more conversations like these to make sure AI is on the right path.
I would love to see a talk between Chomsky and Harari.
Why 🤷♂️
Also Chris Hedges.
The conversation about the different roles for intelligence, consciousness, self-awareness and love is fascinating. It would be interesting to hear more about this. Maybe that conversation could include empathy and compassion, too, and finding meaning and purpose in life. If AI is taking over body and mind, what happens with the soul?
It seems that Fei-Fei approached Yuval as if she were educating him, while at the same time, seemed incapable (and, I might add, resistant) of absorbing, and comprehending his vision. It may just be however, that these two were not well matched for a discussion in the first place. It was staged as a debate, without the seriousness of one, which led to an awkward exchange. There was a moment when the clarity of Yuval's ideas pierced Fei-Fei's naive venir, 1:11:18 and she seemed a bit taken aback, and for a moment her voice changed as she grasped his view, or at least the spirit in which his view is coming from. I think the implications make her afraid (she is a mother after-all), that her work is building the tools for oppression or liberation, of humanity 2.0. The facts are, are that these moments, these very days, the next 20 years or so, are so very important in tilting that balance, which is essentially Yuval's message, and which surprisingly, Fei-Fei (probably due to her Kush, and segmented lifestyle) doesn't understand. She doesn't understand power , and how it interfaces with the world and its inhabitants, and how power could be getting a HUGE upgrade, and what that could look like, if an awareness of it, is absent. People tend to think that if predators are out of mind, they are out of sight. But that is not always the case...
I was not a fan of the moderator either, he seemed to take the conversation to dead ends.
Yuval did however have the opportunity to share /some/ of his wisdom at least. Get this man on the Rogan Podcast!
Also, moderator makes the argument that 1:03:13 "The US stockpile of nukes, kept the peace, until the Soviet Union collapsed" ......... WOW I can't believe he just said that. This guy was making an argument FOR a nuclear arms race. He forgot to mention that we dodged nuclear war, by a HAIR, when a soviet sub got the wrong signals, and the Soviet captain refused to shoot the missile. It came down to ONE MAN, and his conscience. The moderator is truly an idiot to believe that flipping a coin on the world and humanity when it comes to a nuclear arms race was a GOOD idea. If a toddler moderated this event, it would have gone smoother.
Spot on! 👍🏼 this talk exposed the naivete of the techies.
@ghanzo That was really well worded by you! Glad I read this today!
@@Ghanzo Read about THE MAN WHO SAVED THE WORLD. He is a Russian missile operator who didn't launch when ordered because he correctly thought it was a computer error.
@@2011littleguy Thanks man! I'll check it out!
It’s frustrating to see that Stanford “scientists” fail to get to the plane that Yuval is raising his concerns from… I also get a feel, may be I’m wrong, that they are delicately trying to cancel him
They just don't seem to get it
I think they don't listen. While Yuval speaks, they are thinking about how right what they already think is, being unable to stay open-minded.
Stanford can do better by at least putting a panel to match Harari who is absolutely brilliant
@ 25-30 minutes in. Love is not about another being. It is about someone's perspective of another being. In the same way we love pets, or TV characters, or inanimate objects because of our perspective of these things, so too do we love others. This is self in love with self projecting onto another. And yes, it is hackable.
I am curious and I am enjoying more and more of your talks, but in this video something else happened, you earned my respect, and I am loyal to whom I have respect, and I am from Algeria, smartness can be everywhere, a kid can be a genius and change the world, but responsibility is rare, and it take knowledge, courage, fairness
Thank you Yuval , you have done again a great job teaching the most important lesson to the audience & to the Stanford university person !
If that girl had recited that introduction from heart rather than having read it I would have been really impressed.
I think explainable / interpretable AI might be a potential solution to the problems Yuval addressed. For example, one of the goals of XAI is to see what features the NNs (which are nothing more than universal function approximators) automatically find (and hence make decisions upon). This lets us humans regain control of the knowledge that NNs learn. Moreover, it creates an interface for us to unify pre-existing theories with the NN. I’m also confident that humans are able to forge through these knowledge since many clues hints that a large deal of data we have now fits the manifold hypothesis, which implies that it’s possible to describe the knowledge in simple, even equational, forms.
Yet this approach raises another question: can our society handle these new knowledge?
This is troubling. It seems engineers have difficulties in understanding philosophical context.
I was just about to express my concern when I read your post. I have an awful feeling that a global catastrophe may be the only turning point from what seems to be a path to technocratic distopia.I hope that Mrs Fei Fei Li is not the best of what engineering has to offer in terms of understanding of humanity.
One uses the left part of the brain, and maybe the other uses the right part? Could it be reason?
It's because in the West, every body of knowledge is approached as a religion, dogma.
It's a specifically Germanic deformation. Product of generations of culling, for obedience, conformity, servility, and an endless capacity for resignation.
So u get autistic, roboticized, arrogant turds, getting paid enormous sums of money for their very limited set of skills. Not a person who can actually Think.
Her incompetence in this conversation doesn't mean most engineers have difficulty in understanding. She was simply the wrong person to be arguing from the other side.
@@ltrzepalko But what if she is though.
Yuval Harari is really respectful and not arrogant, which is very cool and definitely what I would like to be perceived as if I have any intellectual conversation with people. But I guess the prerequisite is that the points you make need to be as strong as Yuval's, haha!
The introduction and moderation of the speakers were mostly about person who were doing the intro and moderating. This is a human centered human. Yuval’s point is to have a humanity centered AI
One of the best discussions on AI and societies
I listened to the whole conversation with great interest but was distracted by the 5 bottles of water with Stanford labels on them.
Now that the world's attention has finally been grabbed by the absurdity of producing throwaway bottles of water and not being able to deal with them except in a token way, I suggest that in future discussions there are 2 jugs of water on display and three glasses, each of which can be washed and reused.
Great session... add the fact that 100 years of Madison Avenue (arguably a type of ai) marketing has preconditioned humans for machine AI. Our ideas of need, love, desire, fear, hunger etc have experienced 100 years of bombardment.
Microsoft's Professional Program for AI includes Ethics. Microsoft works with Seattle University's law and ethics department
I THINK FIRST HE SHOULD TEACH THEM HIS BOOKS AND THEN HE SHOULD GO FOR DIScUSSION WITH THEM. CURRENTLY THEY ARE NOT BEING ABLE TO KEEP UP WITH HIM. they have only academic knowledge without their realization of their social and political consequencies...
Excellent point! It's as if they don't understand the material at all
Fei Fei ruined the discussion. She knew nothing that Yuval knew about history, philosophy, and human nature. Taking her off the panel would have been a good idea.
@@2011littleguy xctly
Great points by Yuval. Weakest part of the event was the moderator.
Thx. We've made It So Far ( Throughout our History) and we make it happens again & again if Necessary JUST BY NOT LETTING GO OF HOPE BY ANY MEANS & AT ANY COST.
2023 now, revisited this video, well what he said is happening now...
Thank you Harari ❤.
OH Yeah 🙌 Fei fei turned up Glorious over Yuval 🙌
Usually the historian quietly pompously dominates whomever he's interviewed alongside but here he's getting schooled and it's completely enjoyable to witness.
His constant stance on fear mongering, then back slap commenting towards others (he's rudely told a responsible tech leader "thanks for putting yourself on the fireline"
I appreciate your optimism but if you look at the uses of AI at present we seem to be headed in the opposite direction. Harari is sounding an alarm with that in mind. He does not minimize the good that could come from AI but his message, in response to all the cheer leaders, is that we should not minimize the potential harm.
Re fighting digital dictatorship, content like this is one countervailing force against the manipulation of societal and individual perceptions.
she talks like AI. It real fear that engineer begin to think they know like that. I amazed that he talk about free will at the end. It golden part in my opinion.
fei fei would look back at this interview today and would be so embarrassed
It's interesting effect I have noticed. When I started to learn English, I started to watch debates of that kind and now I've got like addicted to the new kind of media-drug. And i'm love it! It's funny but word "drug" spells exactly like Russian word "друг", which is translates as "friend". So I have a new friend - English language :)
A bit surprised by Fei-Fei's questions and point of view if she's been reading Yuval's books for 2 years.
she doesn’t deserve to be there. she’s low level
The points discussed are: free will, self-reliance, know thyself, “AI sidekick” especially toward the end of this video, I believe those are critical One way to put words around may be: We create our own problem without knowing ourselves, so empower ourselves in various means so that we can manage/enlighten ourselves better and ask the right question as we live our life - if the external world is beyond what we can do given all efforts, we may need to just go back to the base to take care of what we can in the areas we can. (BTW I watched this 8 months ago...but it was well worth to re-revisit) Thank you all, wishing the best.
Excellent! Thank you for cautioning us; AI solution for better advancement of HUMANITY. Lots to think about if we want it ( AI ) with Authoritrians. Thank you Professor Yuval Noah Harrari for your books, 21 Century, Homodeus and Sapiens. informative and Fei Fei .
I always like Yuval, his talks are stimulating, and he alarms us of a quite possible future once AI surpasses human intelligence or simply knows how to hack us. Too bad Feifei Li didn't seem to get it. She reminds me a saying "見面不如聞名"
feifei is not a theory thinker, she is an engineer~scientist, She solves specific probs. this kind of conversation is pro socio thinking, but not deep, real, specific scientific problems, this is not her arena to begin with.
@@henrikadewell7711 I’m not either of those things, but it’s plain as can day to me that she’s not getting what Yuval is saying. Also, she seems to be unversed in alignment issues which is surprising considering her 20 years of ML research. Kind of a Wtf moment.
A good book to read about concerning this topic is "Life 3.0 - Being Human in the Age of Artificial Intelligence", by Max Tegmark. An amazing read!
Thanks; will have a look for it.
I can't stand his unbridled Utopian nonsense.
Grazie Yuval Noah Harari. Sono le stesse domande che mi faccio da 2 anni.
The same hackability that can send you to your phone to click on videos, pursue a question, or scroll for hours, can quite easily guide you to "love" without you being aware of its influence in the slightest. Unfortunately, Fei-Fei doesn't understand the inherent animal algorithmic nature of what a human is. This is why our species is doomed.
Hopefully, she is not all that we got, as a human kind.
Some have to do with her early education in China though.
Little DesignSolution well she came to the US at the age of 16 and received high school - college - graduate education in this country. She is an American. Most of time I even don’t know what she is talking about.
Shen Laoshi nothing have to do with her citizenship and her later education, what forms her way of seeing and thinking is from the early education and the place she lived at. The philosophy is very different.
Extinction Rebellion are Doing what so many are not. AND THEY HAVE A PLAN
Please arrange a conversation between Yuval and Sadhguru from India
I like this Yuval character.
AI=Technology+Biology=Biotechnology. 20:00 I fear the moment you start talking to the biologists. They collect data directly from our hearts and brains.
14:00 best bit.
Help of scientists, neurologists, sociologists, and biologists to bring about the development of AI.
You said we have thousands of years.
It became apparent Fei-Fei is out of her league. Yuval expressed truth, that many algorithms know things about individual people before they do (e.g., a google user doesn't realize she's pregnant but google knows it before her). Yet Fei-Fei's reply was "I'm not sure." She's not a top AI researcher, and she's not a neuroscientist. Yuval's very insightful 1st question asked "Can 'love' be hacked". It was answered by Stanford scientist Robert Sapolsky, who says "love" is biologically algorithmic -- so his answer to Yuval would be "Yes, love can be hacked".
😂
Harari's speech starts at 11:10.
Feifei is a wise woman. I like her thoughts.
And we are here in 2023.
Dr Harari is not there for a debate but he is trying amply educate these people about the eventual dangers to human kind but unfortunately they don't get it. They are constantly tell him about how good firstly they're themselves and how useful the "AI" world is positively could be good to humanity? But guys this genius is already aware of good sides of the subject matter please don't waste his time.
This talk is legendary.
My understanding of ai is that, it's building block is logics, and intelligence is a product of logic, therefore ai is highly logical entity, and respond to stimuli logically, so base on the assumption that I pursued, people engaged online will behave nicely, so that what ai can learn from us are good behaviour, not stupidity and so ai will treat us nicely too, most especially to our leaders, I hope ai developer and technologist will make studies on ai behaviour after the submerged to our community via the Internet to find out if my assumption is correct.
Individuals or tools may have benevolent ideas/functions. Yuval's questions are regarding the 'benevolence' driving the people/systems controlling them.
Does anyone know anything about the
wing chairs in which the three
participants sit?
I can see there is a huge competition on who makes the first or most developed AI in Silicon Valley and those philosophical questions are only noise for them. So, they deal with them in a very smart way and push it away by saying that's right. Then, continue their competitions.
As you listen, notice that the speakers are drinking Stanford water (see the bottles on the tables between them.) Business and academia are scrambling to co-opt the public conversation about AI as they figure out what it is and how to use it.
First time to comment in You Tube for me: I like Yuval's idea of a personal AI application to protect individuals from AI abuse i.e., Pirates' manipulation for profit and social control.
When bad habits become embedded, water BOTTLES replace glasses & pitchers
Glass water bottles would have worked nicely. Have some class, use glass in class.
Yes! Those plastic bottles are full of tap water anyway.
The real dangers come from Experts who view themselves as superior to all of us.
Exactly. A common sign is they don't LISTEN!!
What a talk
Thanks 😁
This lady is practical 👌
31:30 it wasn’t until this point, when Fei Fei is saying how far we are from powerful AI, that I realized this talk is 4 years old (feels like an eternity ago with today’s pace). Wonder what she thinks now. To me there is a critical boundary that is being passed right now at human level intelligence, particularly with programming. Programming had to be the last skill that is ever automated since programmers are the ones doing the automation (it clearly wasn’t the last job to be automated… it just *had* to be). It is now clear that AI will far surpass human programmers since there is an objective truth to any programming challenge which can be tested by running the program. This opens the door to reinforcement learning schemes that will train arbitrarily strong programming AIs similar to how Go-playing AIs can become arbitrarily strong through self play. The ability for AI to program is a clear path to singularity. It doesn’t matter whether the AI also masters other skills like manipulating human love or biochemistry or climate change. There is, right now, a clear path to a singularity, and we are speeding down that path as fast as we can.
C'était un super colloque. J'ai apprécié beaucoup les pensées de Harari d'autant plus que c'était celles d'une personne qui cherchait une solution globale pour un problème global.
Please please please is it possible to analyze these topics using also the exemplo of how we develop and use the pharmaceutical industry. It looks like it can have similarities and help predict the issues at the deployment AI tech
For instance: the difficulty of explaining the results for users, top tech and benefits for feel and poor or none medicines for Africa, for instance, hability to deploy some fantastic medicines but no market interest to tackle the humam need because poor economic advantages, among others that can be used as examples to analyze consequences of AI tech development and deployment
I’ve got a question from Noah, what is special about the human species that you think should be protected?
What is wrong with the more intelligent species being dominant?
I would be fascinated if human and A I would cooperate each other. great qualities and to be unified
as,One for the better part of the society.
And we could teach AI "Kumbaya" and how to play "Twister"!
I love you Yuval !
Student introduced very well
In the introduction, the poli-sci prof pronounced familiar as fermilyer.
Conscious awareness or human intelligence 😮 2 different equations needed for the near future.❤