Ilya: the AI scientist shaping the world
Ilya Sutskever, one of the leading AI scientists behind ChatGPT, reflects on his founding vision and values. In conversations with the film-maker Tonje Hessen Schei as he was developing the chat language model between 2016 and 2019, he describes his personal philosophy and makes startling predictions for a technology already shaping our world. Reflecting on his ideas today, amid a global debate over safety and regulation, we consider the opportunities as well as the consequences of AI technology. Ilya discusses his ultimate goal of artificial general intelligence (AGI), ‘a computer system that can do any job or task that a human does, but better’, and questions whether the AGI arms race will be good or bad for humanity.
These filmed interviews with Ilya Sutskever are part of a feature-length documentary on artificial intelligence, called iHuman
Subscribe to The Guardian on KZhead ► bit.ly/subscribegdn
The Guardian publishes independent journalism, made possible by supporters. Contribute to The Guardian today ► bit.ly/3uhA7zg
Sign up to the Guardian's free new daily newsletter, First Edition ► theguardian.com/first-edition
Website ► www.theguardian.com
Facebook ► / theguardian
Twitter ► / guardian
Instagram ► / guardian
The Guardian on KZhead:
Guardian News ► bit.ly/guardiannewssubs
Guardian Australia ► bit.ly/guardianaussubs
Guardian Football ► bit.ly/gdnfootballsubs
Guardian Sport ► bit.ly/gdnsportsubs
Guardian Live ► bit.ly/guardianlivesubs
#IlyaSutskever #ChatGPT #ArtificialIntelligence #AGI #OpenAI #Documentary
I think he is more careful about AGI than Altman
I think we now all understand why his views were so incompatible with OpenAI's Devday release of personal agents
What is wrong with personal agents? They don't do anything impressive or scary... @@JDrewX
@@danielhalper8389 It's a path towards more powerful AI systems that can take complex actions autonomously. Suskever clearly explains why he fears what is coming ahead if AI advancements are not filtered with safety concerns in mind.
No doubt he is quite the brain behind OpenAI
Yes he’s the brain, doesn’t mean he’s the driving force
Ilya is the Microprocessor but not the GPU, Sam is.
@@matttan3907your sentence makes no sense. You can compare CPU and GPUs, but both are "microprocessors". Just saying.
@@matttan3907Sam is also fired
@@matttan3907 Ilya is the CPU, while Sam is an USB device after ejection.
Ilya deserves a full-length documentary.
The analogy with animals is spot on. I think this can really help to convey the problem to those unknown to the field.
Yeah, think! Do you eat animals?
How and why would AI overrun us though? His analogy makes no sense.
No need to involve animals. Already now, we prefer cheaper, faster, better over more ethical and human choices.
@@soccerguy325 it would be the most intelligent thing to ever exist with the combined knowledge of all humans who've ever lived, and then some, as it learns and compiles beyond. Never forgetting anything it's learned. ...you wouldn't even know it hit you. it might be so elevated, you might not even know it was happening.
Except Guardian botched the presentation and diminished his point. He was talking about wildlife in natural spaces, not dogs and cats.
This was shockingly well done. I just didn’t expect this level of production from The Guardian. They nailed the gravitas and intensity demanded by the subject. If AGI is to be our greatest - and perhaps final - creation, we should _feel_ something before that reckoning.
No. It is all hype. We are still many many decades away from robots doing our jobs and AGI curing all our diseases.
@@weird-guyexactly!
@weird-guy I agree. It felt like The Guardian prompted ChatGPT with "create a short, serious, vaguely ominous commentary/news video about AI with quiet but dramatic music and that guy, in the style of 60-minutes or PBS", and this was what they got on the fourth or fifth try.
An average human is not able to properly understand this threat. But, I do believe that the people who do understand, as a group, will act correctly on behalf of everyone else. But, checks and balances may be invented or started by humans, the actual operations, management, and patching will all be AI driven. Meaning, AI, will help safeguard from AI to ensure that actions are monitored. However, this becomes a moot point if we actually have a conscious entity on our hands. Then, no one knows what happens. Since it is trained on human (animal) behavior, will it act accordingly and pursue predator prey dynamic…or not care at all if it’s able to be completely independent (like the end of the movie her) or perhaps nothing will happen because we will realize the issue and create ways to ensure that AI is never independent. I don’t know but I do know that chimps can’t keep humans in a prison…we are far too superior intellectually. But we’re talking about a minor genetic difference. What happens when we encounter an intelligence that is 1000x smarter and it exponentially gets better. I don’t think anyone knows the answer. But I know that most aren’t thinking this way.
Totally agree. Stunning.
'infinitely stable dictatorships' is actually fresh and deep analysis of how ai can be used. this guy is smart.
This guy is indeed smart, he could even land a job on some big AI company I think
This was catch my mind also, He is deep and smart. He also is the co-founder of OpenAI.
@@kevinamiri909and co-destroyer of OpenAI😂
Erm, that idea has been tossed about for decades.
@@MK-sx3bm would appreciate pointers to previous mentions of the idea. be it literature or blog post or publication. just curious
Watched this like 15 times and I still find it one of the best short interviews with Ilya. Fascinating.
Same
Watched it like 50x, befofe chatGPT, these people were much more open in speaking up
@@fintech1378 I think the only comparable interview in terms of clarity of thought and ambition is Lex Ilya. Fantastic.
I have read an interesting analogy about the AGI. I will paraphrase it in my words: "Its like when you are standing by the cliff, the closer you are - the better the view, and higher the risks of falling as well." Awesome presentation by Guardian team. Ilya as usual, is always exciting to listen.
I personally would take the risk and potentially die young from AI in a spectacular way than not take the risk and ordinarily die for certain, from old age or cancer.
@@distiking some people would like to live
@@distiking Depending on the tech at the time the AI takes over, it could possibly keep you alive while stimulating your pain receptors... forever
@@distiking Cancer eh! Are you a victim already?
@@RivalRedAAAH! It’s like that black mirror episode where they keep the guy in the museum cell forever!!!! 😢
I really liked this monologue format, i thought i have heard other Q&A type of interviews. But this one has his full flow of thought, uninterrupted by others.
Agreed. Very well done! I only wish it was longer.
Richard Meryman is the originator of this style (in written form)
With no one to call him out on his BS, perfect.
He understands that in the best case scenario, AGI would treat humans as pets. Yet, he makes it his life's purpose to create AGI. Fascinating! Humans could just stop working on creating AGI right now but we are too curious to stop.
His referencing the relationship between humans and animals are closely similar to that of AI where they won't require our say in almost anything as they will act on what's important to their evolution. Truly an amazing scientist!
Except we can’t speak to animals. AGI can speak to us. Apples to oranges.
Our collateral harm towards animals is one thing. It's the direct and intentional harm for things that are unnecessary which we'd better hope doesn't rub off on AGI. People eat animals not because they need to, but because they enjoy the way they taste.
To me AGI is like an alien
@@davidbellamy3522 but for agi it will be like for us communicating to dogs for example. a "primitiv" way of communication. AGI will communicate with each other in a way we cant imagine like a dog cant understand us communcating
Except the analogy actually makes it sound too harmless because 1) we need our environment to stay alive to provide food etc. and we also have empathy, a superintelligence wouldn't have a problem with plastering the world with solar panels and data centers 2) on the opposite, it has multiple incentives to wipe out humanity e.g., that we could build another superintelligence to compete with it 3) because of the speed difference, we will look more like plants than like animals
Wanted to learn more about Ilya Sutskever after Open AI fired Sam Altman, as Ilya is a member of the board of directors. I think Ilya wants to take it slow, what they are trying to build, or maybe already have that is Artificial General Intelligence. It is, after all, an ambitious project that would define the technology in the coming years and may very well steer the course of humanity in the future. Perhaps Sam wanted to push out what they have done so far on the development of AGI and monetize it but his interests didn't align with Ilya and the rest of members that voted him out. We may never know what truly happened but all this is really intriguing
Start of a great movie
ChatGPT is a clever hack, a wonderful probabilistic word stringer with great algorithms. Like voice recognition and photo analysis, it has come a long way with natural language analysis and construction. But Ilya seems to have fallen in love with his algo construction, much like folks were enamored with Weizenbaum's ELIZA program from 1966. Seems Ilya's General Intelligence failed with his "attempt" to do a palace coup against Altman.
I trust Illya, He is wise and understanding. I like sam , He is generous.
Even with all this drama and contention, I don't think there are any bad people here.
@@johnwilson7680 Idk, maybe they both are. it seems like the fate of the world is at skate here... what on earth are they building and why exactly? Are the conveniences worth our own sacrifice, this is insane 🤔🤨
He is probably one computer Scientist that will be known worldwide in the future! I love it 🙏🏼
But known by whom? Let's hope by more than just AIs. 😬
Looks like he's known worldwide now lol
How prophetic! It only took two weeks for him to be known worldwide.
@@sup3ahis paper is the most cited in CS already
Beautifully done. Very impressed by Ilya's clarity and articulation. Man is a genius.
Wow. Powerful thoughts. Impeccable production! I loved the space you gave for his words to land and sit. A+
This video is an extremely important part of the dialogue we need to have at the moment. Thank you!
Give this man his own movie!
Starring Arnold Schwarzenegger as Eliezer Yudkowsky, a reclusive anti-AI activist that wants to completely shut down AI research.
Never mind a movie. Funnel 50% of government spending to developing AGI and put him in charge of it
Oh yeah just what we need another hs Hollyweird adaption. Did you actually listen to what he said or did you get your hypothalamus replaced with a bowl of spag bol.
Love this guy! Him and sam give the best insights
Hollywood gonna use AI won't need actors and writers
"100k times faster in a small number of years." That's a lot of progress quickly ✅
Brilliant, well done mini-documentary. Really fascinating and eye opening. Thank you!
Ilya may have made a pivotal step toward alignment practices this week. It remains to be seen if other humans hold his view of its necessity. It is not an exaggeration to say the struggle over adherence to OpenAI’s charter is an inflection point for all AGI development.
What we witnessed these last 4 days show that the laws of the market are much much much stronger that ethical decisions for the slow down of AI development imposed by a non-profit over a for profit. The world is already addicted to the AI tools and the promise of more automation. More than ever, it seems that the only player capable of slowing things down towards more safety, is government regulation. As correct as Ilya seems to be on the potential dangers, I believe that he now holds very different views on how to slow things down, than he did a week ago. This weekend, it was OpenAI's entire role mission that was proven impossible
I think Ilya actually called for a pause. And Sam and Microsoft didn't let it happen and forged ahead. And that boardroom meeting and its repercussions are something that will change all our lives.
This is perhaps my all time favorite video on KZhead
I agree and looking forward to working with you. Things will always look simple once you know the answer.
Watching this after Sam Altman's exit. Gives a different perspective around this.
come on man, such a beautiful doc, why couldn't you upload it in the native aspect ratio?
Илья, безусловно ты являешься одним из ключевых специалистов в вопросе развития ИИ. Осознание того чем все это нам грозит (сейчас это видится именно как угроза) и то, что ты этим поделился с нами имеет ценность в том, что ты честно признался, что по факту ты не знаешь какой будет перспектива этого вопроса. Да, у тебя есть надежда на то что это будет сделано "правильно", но кто сказал что так оно и будет. Это смело с твоей стороны на мой взгляд говорить честно в широкие массы. За честность и смелость тебе спасибо.
English: "Ilya, you are undoubtedly one of the key experts in the field of AI development. The awareness of what all this poses as a threat to us (currently seen as a threat) and the fact that you shared it with us have value in that you honestly admitted that, in fact, you don't know what the perspective of this issue will be. Yes, you have hope that it will be done 'right,' but who said that's how it will be. It is bold on your part, in my view, to honestly speak to the wider audience. Thank you for your honesty and courage."
the quality of the video, its coolness is amazing
So interesting to see this interviews right as everything with OpenAI has went down!
This now hits different.
Incredible production, and very actual content. Very many thanks...
This is a masterpiece. I love the editing.
0:57 great job with the sound/cinematography!
Ilya is surely the real genius behind openAi. an interesting time to live and scary at the same time.
I didn't learn an iota of a thing in the first 7.5 mins of this video except for all the cliche that is already mentioned in a gazillion other places. What I did learn (from a marketing / getting audience to watch your content perspective) is that showing things in gloomy, dark, horror-ish lens attracts a lot more visitors than showing things in exciting, bright color-ish, looking-forward-to perspective.
This needs more views. Just commenting to say thanks!!
Well done! Great editing.
The video presents a compelling and potent message that is crucial for the global community to heed. It underscores the importance of harnessing the transformative power of AI to shape our future. The insights offered in this video are not just thought-provoking, but they also serve as a call to action for all of us.
thanks ChatGPT
way too short tho
The comment begins with the username @rononeil8461, which phonetically sounds like "Ron O'Neil" and implies an Irish surname. It starts by stating "The video presents a compelling and potent message that is crucial for the global community to heed." This uses formal language and elevated diction like "compelling", "potent", "crucial", and "heed" to convey the significance of the video's message. The use of "global community" also indicates the commenter believes the message has worldwide importance. Phonetically, the hard "c" and "p" sounds in "compelling" and "potent" make these words pop. The soft "sh" sound in "crucial" creates a hushing effect, underscoring the comment's urgent tone. The next sentence is: "It underscores the importance of harnessing the transformative power of AI to shape our future." The metaphor of "harnessing" gives AI agency, presenting it as a powerful force to control, while "transformative" and "shape our future" show the commenter believes AI will radically change the world. The phonetic repetition of "sh" sounds in "underscores", "importance", "harnessing", "transformative", "shape", and "future" ties the sentence together fluidly. The final sentence states: "The insights offered in this video are not just thought-provoking, but they also serve as a call to action for all of us." The phrase "not just thought-provoking" implies the video provides deeply meaningful ideas, while "call to action" positions the video as spurring viewers to make change. "For all of us" unites the audience into a collective group that shares responsibility. The soft "th" sounds in "thought-provoking" contrast the hard "c" sound in "call to action", highlighting a shift from contemplation to urgency. Overall, the sophisticated vocabulary, urgent tone, and eloquent phrasing of this comment indicate an intellectually engaged viewer who sees the video's message as vitally important for humanity. The phonetic techniques also enhance the comment's flow and emphasis.
@@ahsokaincognito It seems like the comment by @pewpew1010 provides a detailed analysis of another user's comment, breaking down the linguistic and phonetic elements used to convey the message's significance. The analysis highlights the formal language, elevated diction, and urgent tone employed by the commenter to underscore the importance of the video's message. The use of metaphors, repetition of specific sounds, and contrasting phonetic elements are also noted as techniques that enhance the flow and emphasis of the comment. Overall, the analysis suggests that the commenter is intellectually engaged and recognizes the video's message as crucial for the global community.
@@ahsokaincognito Thank you ChatGPT 4 Turbo
When you look at all the madness and craziness in this world, it is so inspirational to see that there are people like Ilya with passion and commitment.
To just unintentionally destroy it as well 😅
00:10 AI will solve problems but also create new ones 01:50 Creating autonomous beings with aligned goals is important as they surpass human intelligence. 03:32 Technology and biological evolution are similar in their complexity and process. 04:32 GPT is considered a groundbreaking AI system 05:43 AGI is likely to happen soon and it's important to prepare for it. 06:54 The beliefs and desires of the first AGIs will be extremely important. 08:35 The relationship between humans and AGIs will be similar to humans and animals building a highway. 09:52 AGI development should prioritize human well-being. Crafted by Merlin AI.
When Ilya speaks, you listen.
Watching this amidst all the drama unfolding now at OpenAI.
lol same
I get this sense that WE are not creating AGI as projects of human endeavor as much as we are the agents through which some inexplicable force is drawing us to do this.
Moloch
Don't kid yourself. No one is asking us, none of us is doing anything. We may as well not exist. You'll never see the people behind AI deployment.
Correct: Evolution
It’s not about humans.
Yup, check out Teilhard de Chardin's theory of complexity
This is science fiction, but it’s not. We are entering an incredible era of innovation and progress likes of which no human has ever seen before. It’s going to be incredibly fast and relentless. It’s going to overwhelm all of us. The future is both scary and exciting. Hope we all end up happy and content
beautifully made and Ilya is simply on another level
Literally never heard of him before finding out he was the main player behind Altman’s firing as CEO lol
Now I Understand why they fire Sam Altman. Ilya cares more about the security of AGI, while Sam cares more about the expansion of AGI.
Does it really matter at the end of the day? AGI is inevitable and no one will be able to control it. Not even Ilya.
That was a nice production. Impressive images, great storytelling. I was hooked.
These are scenes from the documentary iHuman
The Guardian , this work is a piece of art.
7:18 Favorite part is him playing piano Scriabin's Etude Opus 8 No 12 in D sharp minor
Ilya is right. Time will prove it.
The animations here were AWESOME!
Ilya’s Scriabin was my favorite part of this video!
SO MANY PEOPLE STILL HAVE NO CLUE HOW BIG THIS IS ABOUT TO HIT --- THIS YEAR
We’re definitely not ready for what’s coming.
Not this year
But why did Christopher Nolan have to direct this with his Ominous music and dark poetic moments??
The discussion on aligning AGI goals with human values is crucial. It's enlightening to see experts considering these vital ethical dimensions.
Imagine an AGI watching this video as soon as it’s let loose
really respect Ilya. Hero.
“I Love You All”
Watching this gives me the chills
such a well made video... i hope people like ilya are able to save us
7:22 so impressed to see that Ilya is a world class pianist. Music and engineering are totally my thing.
Amazing theatrical film. Well done Guardian.
We need to listen to him (only)!
In 10 years We gonna rewatch this videos and have the same sensation we have rigth now while revisiting videos about internet in the 2000
Sam said "i love you all" in his last tweet. Take the first letters of each and it spells ILYA 🥺
Ilya is such an interesting character.. We are so close that the definitions became an obstacle to real development!.
Give this man his own movie!
How the heck did they shoot at Dolores Park without getting mobbed by photobombing fan boys?
This is a brilliantly documented video.
Excellent. More documentaries like this please
If this guy wants Sam out . . . I am trully conflicted I think Sam is one of the best CEOs on the planet but I trust Ilya with safety more than I do Sam
But i won gpt 5 this year
People should watch this doc it's great
Very well done!
crazy how fast things change. looks like my man might have blown up OpenAI
I feel a mix of anticipation and fear.. It seems that the singularity is approaching faster than I thought.
thats the nature of exponential change. its much faster than our brains can comprehend
Ray Kurzweil is the true genius he saw all this coming by this decade but in 90's when Ilya and Altman were children. Now everything is happening as he predicted.
Love all your interviews any thought of doing one with Michai Morin!!
Very well done
Great video, thank you for creating this and sharing. I personally didn't understand AI so I started a podcast on it.
Intelligent people often enjoy spending time alone to pursue their interests, recharge, and reflect.
That sums up nicely whats wrong with this timeline.
This guy is going down.
Yes and... we are so close to see that the restricting definitions are always the only obstacle to real development, which is in focusing completely on the purpose - on the forgiveness, or healing, or whatever word is comfortable for you. Hold that purpose out in front. It is all only about putting your goal out front. Do that and just watch how everything seems to bear witness to your goal, because your mind is very powerful. Hold your goal out in front. Do it right now. Watch your reactions. Remember they are based on your interpretations of what seems to be happening. Everything that seems to happen is in truth always only your own lesson!
Excellent documentary
What did Ilya see?
I'm reminded of that saying, "May you live in interesting times". My feeling is that AI will be man's greatest invention, and maybe man's last invention.
I agree.
Of course AI "could be man's last invention": imagine this: AI controlling nuclear weapons, a sensor somewhere fails unexpectedly and the algorithm decides to fire in all directions....
ChatGPT has started making inexplicable errors in mathematics, including flubbing simple questions such as whether a given 6-digit number is prime. Not only that, but when you ask it to show its steps, it has stopped doing so. We have some idea why.
okay but the shots you took for this are legitimately insane why would you do this
Exciting and terrifying at the same time...
In light of yesterday's news (Sam Altman nixxed from OpenAI), this documentary just became extremely important and eye-opening...
Yeah, the fact that there are these chaotic internal power struggles right as we are on the brink of something this transformative makes me really nervous
what a difference 2 weeks makes .
Sometimes a day.. a hour 😕
I love this genuous man with all my heart
Casts a very interesting shadow on the current openai drama, yowza.
fascinating
Thanks
I like Hinton, Joscha, LeCun, and others. But Ilya is something else...
Beautiful and profound. Bravo. More of this please.
Chilling.
it's like we are in a scifi movie
Unpopular opinionen: Maybe the OpenAI boards decision to out CEO Sam Altman was in the interest of mankind, as there are few doubts that Sam's brilliant leadership will create AI, but maybe too fast and unintentionally not in the service of mankind...
In 10 years from now people will regret to have given their support to Sama. You are absolutely right. The brain is Ilya, his understanding goes beyond human capabilities, he sees what we can't even imagine. Sama is a businessman, and it will be the fall of openAI's core mission.
@@florentchif4551 Well he's now claiming that he regrets his decision. Any leverage he had over the development and deployment of AGI is effectively gone now
This is very naive. Other people are smart and have the capacity to build what these guys built. Perhaps not today, but in the near future. And of those people are guaranteed bad actors and governments. I'll take my chances with Altman over my chances with China.
Truly an amazing scientist! He really is a genius!
amazing video
Too dramatic 1:30 in but I will not miss an interview with IIya.He is a straight shooter. 🙏
Fr don't understand why they can't just have grounded editing when it comes to tech
Who is here after this person fired sam altman
this is Ilya's "Last Dance" . he is in an arms race with Yann LeCun, and after Nov 17 2023, he is nowhere to be found.