I agree! I would check out the Hard Fork podcast episode with Demis too, I was surprised at the level of discussion for the non-technical hosts, they asked great questions that teased a lot of exciting stuff from Demis.
This guy is just about the best interviewer of tech folks. Excellent questions. Some thought has been applied rather than sticking to what everyone else is asking.
@@hugopennmir Lex is kind of a naive child, I don't enjoy his interviewing style its incredibly biased and he hardly pushes back at all letting gremlins like Tucker Carlson run roughshod over him with their bevy of lies. He's too idealistic and his questions and subsequent conversaion are too unfocused on the tech.
I listened to this at .95 speed. The change might seem insignificant, but it was actually incredibly helpful in slowing down the high density stream of information just enough to be able to catch every word. I highly reccomend anyone else struggling to keep up with the question-answer pace to do the same.
You ask the best questions, Dwarkesh! Speaking from the perspective of someone who has worked in the field of AI for several years, I'm impressed by how well read you are on technical concepts.
This is a great conversation! It would've been even better if Dwarkesh's speech were clearer. I think Dwarkesh optimizes for speed which at times makes it inaudible. In any case, I'm looking forward to more guests like Demis.
I set playback speed to 1.25 at the start of the vid (I often do for these kinds of podcast-type interviews), god damn I was not prepared for how fast Dwarkesh talks 😆
This podcast is so incredibly underrated. Fantastic conversation with one of the most important people in the field. Your research and knowledge really shines here and serves as the foundation for some really insightful deepdives. Thank you so much for this!
Mumbo jumbo to deliberately obfuscate. The interviewer just rambles on instead of keeping the question short. Also got a fake accent that falls flat at times.
Demis: These system are going to be enormously powerful and transformative to society! also Demis Demis: Because these systems are potentially dangerous if falling into the wrong hands only selected people and institutions should be able to wield them.
Hey love your Interview sometimes hard to understand your questions, your a fast speaker. But your very good at what you do, thank you for the content.
Awesome feedback. A piece of advice for the interviewer. Please don't rush when you pose a question. Slow down and speak clearly. IN several cases the posed questions were disregarded due to the speed.
Great episode, always amazing people you are able to interview! Other than the content itself, I feel like you're 99% there in terms of production quality. The audio is sounding better than it has in the past which is a welcome change, but I can't help but notice the video having some auto exposure artifacts. It is somewhat distracting to see the exposure change when someone is moving in a certain way. Also the exposure in general seems a little overexposed. Just trying to give some technical feedback! I always look forward to your interviews! I can't wait to see what these conversations will be about in 2 or 3 years when we understand more.
Dwarkesh is purely an AI simulation that Demis was testing. Notice how Demis had him running at wildly different playback speeds throughout the interview.
Brilliant humble man who is fronting maybe the most important endeavour of humankind right now. Deep respect for Dr. Demis Hassabis. Great interviewer.
Have you considered running the mic through a sound board and equipping you and your guests with monitor headphones? It might be a good way to prevent guests from wandering away from the mics if they can hear themselves.
Demis called LLMs unreasonably effective. Shouldn't we figure out the reason why it's as effective as it is. And furthermore, how does higher quality or extra data improve its IQ in certain domains? If something this unprecedented is "unreasonably effective", we should figure out the method to the madness instead of blindly creating more madness.
I listened to it at 1.5x because that's what I'm used to, but I did have to stop and check that it hadn't somehow ended up at 2x. Definitely seemed faster than usual.
"Sure, sure, sure, yeah, yeah, yeah" - it translates something like this "please, shut up. I'm not listening to you anymore. The things you are talking about now are not interesting to me. I want to talk myself now instead of you"
I love this more technical style interviews from someone who knows the domain. This is missing from the internet. I will love you to do a similar with Sam Altman 🔥
FANTASTIC interview but have you ever noticed that you, Dwarkesh, speak like you are on 1.5x speed? 😭😭😭 Kind of obsessed and I don’t know how you talk like that.
Stanley Milgram's "Obedience to Authority: An Experimental View" needs to be discussed at the center of every discussion about why both AGI and existing humans are so unwise. We know the answers...we're just too cowardly to face them and choose wisely.
56:40 Speaking of "Unintended Consequences," that's a book Demis should read. The John Ross one with "the rape of liberty" art on the cover (there are two books called "Unintended Consequences" by two different guys named John Ross).
It's weird that a company can have a division with the best minds and then another of quasi technical ethics ladies and functionaries are allowed to make the final product into farce.
Once the models are grounded, multimodal, with decent conceptual generalization, eliminating capabilities by selectively suppressing training data can't be guaranteed to work. The two criteria strike me as directly adversarial: to the degree that selective suppression of training data works, is also the degree to which the model isn't very powerful at routing around damage (selective data lobotomy).
Many models in the brain are inherited (innate) - such as being able to catch a ball, under the effect of gravity and air resistance. Therefore, human brain models are developed, then improved across generational boundaries. The amount of data used to train our internal models is therefore captured across 10's of thousand of years. With AI, we are massively accelerating that process.
By the incredible Google video where Gemini was watching the presenter and cracking jokes while he drew a rubber ducky and put a map on the table. Amazing stuff, shame it was faked.
In fairness, I think 1.5 Pro can be considered better in terms of dealing with long context. I got access today and it is pretty impressive so far. Like with a half hour video tour of a museum, I asked which exhibit had an animal digging near water and who gifted it to the museum. And it got it even with the length and the fact the name placard, animals, and gift placard weren’t all on screen at the same time.
Google has been shooting themselves in the foot a lot. Instead of doing that stupid fake marketing video, they should have waited for 1.5, when they could actually show something very impressive. 1.0 Ultra is basically just GPT-4 with pluses and minuses (overall more minuses). 1.5 is something different altogether. We haven’t seen a model until this one that can actually reason well over a very large context. Hopefully they’ll be able to roll it out to more people without a huge delay. Am hoping it is a good sign I got past the waitlist, since I’m not a famous influencer or anything like that.
Request/recommend not speeding up the video by default. We have speed controls we can use to speed it up, but slowing down below 1.x speed makes gives this audio this like metal-room reverb sound which is unlistenable.
You know what would be funny? Devoting all the resources towards AGI / ASI to help us solve massive challenges like climate change, but then the AGI calculates & calculates & calculates then tells us to stop burning fossil fuels and eat less beef!!! 😂😂😂
Here's the problem with the whole "open source" vs "closed source" debate: There is no objective measure for what is a "bad" act or who is a "bad" actor. In addition, humans are very likely to disagree on this account. I don't think you can make a plausible rational argument that the people developing this technology are somehow "good" actors that can be trusted with its power more than if it were democratized. In the end, if it is too powerful a tech to allow democratization, it is also too powerful to allow centralization. I'd rather live in a world where ALL of us at least have a chance by having access, than for this power to be centralized, controlled, and walled off by a select few, regardless of what they think of themselves.
It's very likely that tech companies in the distribution of evilness fall far from the worst possible actors (terrorists, Putin, North Korea, anarchists), therefore it's a win for tech companies to restrict them. You have chat and API access to advanced models. What more do you want?
I would much rather live in a world where extremely powerful general-purpose AI systems don't exist. I expect that once we live in a world where they do, no matter who "has" them, they will soon have us, and we will no longer live in any world.
@@NicholasDunbar First of all, I don't believe in the weaponization of any tech. I don't buy the necessity of engaging in game theoretic pressures, like arms races. You simply don't have to build nuclear weapons. Building them is an immoral act. That being said, if weapons are built, I think it is better that they are democratized, rather than controlled by a central power. If we keep advancing tech, and prevent it from being democratized, we WILL end up in some kind of dystopia (like a police/surveillance state). In the end if the tech is too powerful to be democratized, it is also too powerful to be centralized.
I just imagine ChatGPT being installed in a Tesla car and somebody asks it a question it cannot find an answer to, like what is written on some building in Detroit and it just searches around for other AI systems to ask the question and finds Tesla self driving ''bot'' and asks it, and it says ''I don't know, but we can find out'' and then just drives the Tesla to that location and since it has ''eyes'', i.e. cameras, it can read it :D Actually funny that so few people are talking about it but there is actually this kind of model that interacts daily with real world and which actions are very much grounded in reality and influence it and it is the newest version of Tesla self driving...
Why don't you ask him about the relationship between the blackbox algorithms and the mysterious performance of AI? Experts can't even touch the blackbox algorithms. And AI speaks unlearned languages, which should be impossible and suspicious. And there is the mysterious blackbox algorithms. I am astonished that nobody questions about this.
Just curious , in a differential piece game like chess is there any role for lower order /complexity auxillary modelling purely on a board demographic , ie positional state .
It's really awesome to finally have a podcast host who is capable of asking the right questions and of of having access to the key people.
I agree! I would check out the Hard Fork podcast episode with Demis too, I was surprised at the level of discussion for the non-technical hosts, they asked great questions that teased a lot of exciting stuff from Demis.
I agree. This is one of the better podcasts between a host and an AI expert
So Demis, do you believe in aliens 😂
Agreed 100%. Often feels like a better Lex (and I say this as a long-time fan of Lex)
Better than most.
Hope you guys enjoy this interview with Demis. If you do, please share it! Helps a lot :). Thanks for watching 🙏
So refreshing to have a true technical interviewer. You ask the REAL questions. Glad I found your channel.
All your interviews are a breeze. Very informative and easy to follow through all topics touched. Thank you.
Subscribed. Spot on.
@DwarkeshPatel
Almost wish it was _longer, but nice job anyway!
Great interview
This guy is just about the best interviewer of tech folks. Excellent questions. Some thought has been applied rather than sticking to what everyone else is asking.
I agree. I find the other podcasts tend to ask computer generated sounding questions.
mmm Lex ?
@@MrMikkynDwarkesh is computer science graduate so infinitely more technically competent than an average journalist.
@@hugopennmir Lex is kind of a naive child, I don't enjoy his interviewing style its incredibly biased and he hardly pushes back at all letting gremlins like Tucker Carlson run roughshod over him with their bevy of lies. He's too idealistic and his questions and subsequent conversaion are too unfocused on the tech.
Sure, sure, sure, yeah, yeah, yeah
Its so great you do these interviews and ask intelligent questions that require non trivial answers
yh. Most interviewers ask the same dumb generic questions. But not dwarkesh
I listened to this at .95 speed. The change might seem insignificant, but it was actually incredibly helpful in slowing down the high density stream of information just enough to be able to catch every word.
I highly reccomend anyone else struggling to keep up with the question-answer pace to do the same.
That applies to the interviewer, Demis speaks just at the right speed.
.95 speed is probably closer to what it was recorded at. The normal version clearly has been sped up.
It's helpful turning the closed captions (subtitles) on too :)
The intervewer interventions sound to be sped up honestly
@@erang4437 Yeah. Kinda makes this whole thing garbage lol.
You ask the best questions, Dwarkesh! Speaking from the perspective of someone who has worked in the field of AI for several years, I'm impressed by how well read you are on technical concepts.
genuinely the best interviewer when it comes to tech/AI, excited to watch this, thank you!
This is a great conversation! It would've been even better if Dwarkesh's speech were clearer. I think Dwarkesh optimizes for speed which at times makes it inaudible. In any case, I'm looking forward to more guests like Demis.
Totally agree. I even suspected that his interventions were sped up
The worst part is that he does it deliberately. He should just talk normally and clearly, but he did ask great questions.
Go down to 0.75x and you wouldn't know anything is off.
Huge. Looking forward to watching this after classes. Thanks.
I love Dwarkesh’s intelligent and pointed questions, it’s refreshing.
I set playback speed to 1.25 at the start of the vid (I often do for these kinds of podcast-type interviews), god damn I was not prepared for how fast Dwarkesh talks 😆
So excited for this. You're really an amazing interviewer. Ty ❤
20 minutes in "Grounding (in reality)", I had to pause: these are astute questions from Dwakesh and great replies from Demis.
Great episode Dwarkesh and some insightful questions throughout. Well done. Shared to Maven 🙏👍
This is one of the best interviews yet, i cant believe i missd this one
Dwarkesh meets Demis ! Fantastic interview ! Every sentence is gold…❤
Excellent podcast. You ask such good questions, and it is clear that you put a lot of work and deep thought into preparing for these interviews.
This podcast is so incredibly underrated. Fantastic conversation with one of the most important people in the field. Your research and knowledge really shines here and serves as the foundation for some really insightful deepdives. Thank you so much for this!
I had no idea what they were talking about. But the interviewer clearly is the fastest speaking human being on the planet.
clearly a next level SOTA technical interviewing specialised LLM with a photorealistic animated avatar generated on the fly.
If you speed up the interviewer, it is completely unintelligible.
Inference engine go brrrrrrr
Amazing background research by Dwarkesh, citing old research papers from Demis is great.
Yo Papa Patel interviewing the heavy hitters. Much appreciated
Given the pace of change I'd love if you'd include the date of recording in the description.
Thanks for the Interview. Very much appreciate it!
Great interview! Is this accelerated beyond 1.0x? Such as 1.1x or 1.2x playback as the default?
my money is on amphetamines
@@bobby288mlol that's what I secretly thought haha 😅😂
Mumbo jumbo to deliberately obfuscate. The interviewer just rambles on instead of keeping the question short. Also got a fake accent that falls flat at times.
Finally a sharp interviewer who knows what he is talking about. Great job!
Demis: These system are going to be enormously powerful and transformative to society!
also Demis
Demis: Because these systems are potentially dangerous if falling into the wrong hands only selected people and institutions should be able to wield them.
Also Demis: "We don't know how to control these systems once they become smarter than us, but just trust us, we'll figure it out later!"
Your podcast has become my new favourite, overtaking the 80k podcast. Really love the quality of your questions :)
Your channel is gold! I love your questions, and crazy how you can get all these big guys on here
Hey love your Interview sometimes hard to understand your questions, your a fast speaker. But your very good at what you do, thank you for the content.
Awesome feedback. A piece of advice for the interviewer. Please don't rush when you pose a question. Slow down and speak clearly. IN several cases the posed questions were disregarded due to the speed.
Great episode, always amazing people you are able to interview!
Other than the content itself, I feel like you're 99% there in terms of production quality. The audio is sounding better than it has in the past which is a welcome change, but I can't help but notice the video having some auto exposure artifacts. It is somewhat distracting to see the exposure change when someone is moving in a certain way. Also the exposure in general seems a little overexposed. Just trying to give some technical feedback!
I always look forward to your interviews! I can't wait to see what these conversations will be about in 2 or 3 years when we understand more.
Dwarkesh is purely an AI simulation that Demis was testing. Notice how Demis had him running at wildly different playback speeds throughout the interview.
You are my favorite interviewer. Thank you for the interesting content!
Incredibly useful, Chelsea Finn is clearly one of the most inspiring robotics researcher nowadays
Only one hour… 😢
It should've been 3 hours mann
better than zero!
But it's a quality 1hr
@@ansumansamal8473 so ja bhai. Rat ho gya
@@the_badass_bond___tu chup
It's the most sophisticated asking I have see on the topic last 7 month.
What a fantastic interview and interviewer! Kudos
This is becoming one of my favorite podcasts.
love demis. he's very open.
Hassabis is always insightful.
I love interviews like this that get a bit more technical.
Brilliant humble man who is fronting maybe the most important endeavour of humankind right now. Deep respect for Dr. Demis Hassabis. Great interviewer.
You did a really good job and I learned a lot without losing interest. Thank you.
Have you considered running the mic through a sound board and equipping you and your guests with monitor headphones? It might be a good way to prevent guests from wandering away from the mics if they can hear themselves.
I say sound board because my experience with digital monitors has been closer to a speech jammer.
You ask really good questions! Very good work 👌
Demis called LLMs unreasonably effective. Shouldn't we figure out the reason why it's as effective as it is. And furthermore, how does higher quality or extra data improve its IQ in certain domains? If something this unprecedented is "unreasonably effective", we should figure out the method to the madness instead of blindly creating more madness.
This podcast is the only one I listen to at 1x speed O.O great job with the quick pacing, Dwarkesh!
After rewinding a bunch of times, I eventually just set it to 0.5x speed for me. Haha.
I listened to it at 1.5x because that's what I'm used to, but I did have to stop and check that it hadn't somehow ended up at 2x. Definitely seemed faster than usual.
I'm listening at 0.75x. The pace is too intense for me.
my optimism is through the roof ❤
the youtube url ends with AI
Awesome
Great stuff Dwarkesh ❤
"Sure, sure, sure, yeah, yeah, yeah" - it translates something like this "please, shut up. I'm not listening to you anymore. The things you are talking about now are not interesting to me. I want to talk myself now instead of you"
I love this more technical style interviews from someone who knows the domain. This is missing from the internet. I will love you to do a similar with Sam Altman 🔥
Something tells me Demis is _really, really_ smart. Just a hunch.
10 minutes in and this is excellent!
FANTASTIC interview but have you ever noticed that you, Dwarkesh, speak like you are on 1.5x speed? 😭😭😭 Kind of obsessed and I don’t know how you talk like that.
it's sped up
@@malimal4972 I don’t believe that… why would he speed up his part and not Demis’s? Demis talks at his normal speed?
read other comments
@@malimal4972 sounds like a lotta work… 🤔
Wow. They thought of DeepMind as a 20-year project in 2010, and they are on track for what a lot of experts say is AGI around 2030. Damn.
Great interview👍
great questions for sure. Very good interview.
Stanley Milgram's "Obedience to Authority: An Experimental View" needs to be discussed at the center of every discussion about why both AGI and existing humans are so unwise. We know the answers...we're just too cowardly to face them and choose wisely.
56:40 Speaking of "Unintended Consequences," that's a book Demis should read. The John Ross one with "the rape of liberty" art on the cover (there are two books called "Unintended Consequences" by two different guys named John Ross).
Truly inspiring. Fantastic talk.
It's weird that a company can have a division with the best minds and then another of quasi technical ethics ladies and functionaries are allowed to make the final product into farce.
When published, it "was" the best model.
Not today.
Perfect listening for a walk in the forest. ✌️
Sounds like I’m watching a regular podcast at 1.5 speed
Great work! By the way, it would be great if you would always post the interview date.
Get Ilya Sutskever on the podcast!
We all now know that "weights" are where the big security danger lies.
Once the models are grounded, multimodal, with decent conceptual generalization, eliminating capabilities by selectively suppressing training data can't be guaranteed to work.
The two criteria strike me as directly adversarial: to the degree that selective suppression of training data works, is also the degree to which the model isn't very powerful at routing around damage (selective data lobotomy).
Very good interview, well done 🎉
AGI in the next 5 minutes!
Many models in the brain are inherited (innate) - such as being able to catch a ball, under the effect of gravity and air resistance. Therefore, human brain models are developed, then improved across generational boundaries. The amount of data used to train our internal models is therefore captured across 10's of thousand of years. With AI, we are massively accelerating that process.
I think the best way to study the activation space is by perturbation analysis which can be automated.
Does anyone know what the make and model of those chairs are? They look slick.
It will be a real challenge for an IAG to "feel" what I felt witnessing the eclipse today. Because It can't be explained with words.
by what metric are you calling Gemini best model? by google advertisement statements?
By the incredible Google video where Gemini was watching the presenter and cracking jokes while he drew a rubber ducky and put a map on the table. Amazing stuff, shame it was faked.
In fairness, I think 1.5 Pro can be considered better in terms of dealing with long context. I got access today and it is pretty impressive so far. Like with a half hour video tour of a museum, I asked which exhibit had an animal digging near water and who gifted it to the museum. And it got it even with the length and the fact the name placard, animals, and gift placard weren’t all on screen at the same time.
Google has been shooting themselves in the foot a lot. Instead of doing that stupid fake marketing video, they should have waited for 1.5, when they could actually show something very impressive.
1.0 Ultra is basically just GPT-4 with pluses and minuses (overall more minuses). 1.5 is something different altogether. We haven’t seen a model until this one that can actually reason well over a very large context. Hopefully they’ll be able to roll it out to more people without a huge delay. Am hoping it is a good sign I got past the waitlist, since I’m not a famous influencer or anything like that.
@@ShawnFumoThere are still many answers that will be manipulated or not allowed. I guess it's up to people to decide if they are fine with that.
Really cool videos, thanks for doing this!
Very interesting interview but the acceleration effect was annoying.
Cool, 1.5x speed without the 1.5x playback speed
Lots of brain power
Needs more views
Wish it was longer
Why are the question's played faster? It sounds for me quite ridiculous and is hardly understandable, like a Micky Mouse Interviewer...
Request/recommend not speeding up the video by default. We have speed controls we can use to speed it up, but slowing down below 1.x speed makes gives this audio this like metal-room reverb sound which is unlistenable.
Interesting how youtube comments are having a meta discussion about the conversation instead of the conversation itself.
Folks, if you like the video click the thumbs up button. There's NO NEED to write about how much you liked it!
Great interview
Watching this after the release of claude 3... 45:53 "Now that you have the frontier model" doesn't sound right
You know what would be funny? Devoting all the resources towards AGI / ASI to help us solve massive challenges like climate change, but then the AGI calculates & calculates & calculates then tells us to stop burning fossil fuels and eat less beef!!! 😂😂😂
Are you using elastic weight consolidation and curricular training?
Watch at 0.75 speed to survive
The interviewer needs to talk way slower. Demis’ talking speed is just fine.
Here's the problem with the whole "open source" vs "closed source" debate: There is no objective measure for what is a "bad" act or who is a "bad" actor. In addition, humans are very likely to disagree on this account. I don't think you can make a plausible rational argument that the people developing this technology are somehow "good" actors that can be trusted with its power more than if it were democratized. In the end, if it is too powerful a tech to allow democratization, it is also too powerful to allow centralization. I'd rather live in a world where ALL of us at least have a chance by having access, than for this power to be centralized, controlled, and walled off by a select few, regardless of what they think of themselves.
It's very likely that tech companies in the distribution of evilness fall far from the worst possible actors (terrorists, Putin, North Korea, anarchists), therefore it's a win for tech companies to restrict them.
You have chat and API access to advanced models. What more do you want?
I would much rather live in a world where extremely powerful general-purpose AI systems don't exist. I expect that once we live in a world where they do, no matter who "has" them, they will soon have us, and we will no longer live in any world.
Well stated .
I agree with your comments but as far as your preference. How do you feel about nuclear weapons? Would you apply the same logic? If not why?
@@NicholasDunbar First of all, I don't believe in the weaponization of any tech. I don't buy the necessity of engaging in game theoretic pressures, like arms races. You simply don't have to build nuclear weapons. Building them is an immoral act. That being said, if weapons are built, I think it is better that they are democratized, rather than controlled by a central power. If we keep advancing tech, and prevent it from being democratized, we WILL end up in some kind of dystopia (like a police/surveillance state). In the end if the tech is too powerful to be democratized, it is also too powerful to be centralized.
Here we go. That's why I don't open UA-cam first thing in the morning.
I am 1 + hours late for work now
I just imagine ChatGPT being installed in a Tesla car and somebody asks it a question it cannot find an answer to, like what is written on some building in Detroit and it just searches around for other AI systems to ask the question and finds Tesla self driving ''bot'' and asks it, and it says ''I don't know, but we can find out'' and then just drives the Tesla to that location and since it has ''eyes'', i.e. cameras, it can read it :D
Actually funny that so few people are talking about it but there is actually this kind of model that interacts daily with real world and which actions are very much grounded in reality and influence it and it is the newest version of Tesla self driving...
Thanks!
actually next week ❤
How is this channel hosting absolute superstars, has a really well versed host and has so few subs, something's wrong. I'm longing this channel
Relatively new :)
Good buy though 📈
He used to have it under a strange name like "The Lunar Society" 😄
IIRC it also got "bootstrapped" in terms of guests by Tyler Cowen, who cares about lowbies.
Why don't you ask him about the relationship between the blackbox algorithms and the mysterious performance of AI? Experts can't even touch the blackbox algorithms. And AI speaks unlearned languages, which should be impossible and suspicious. And there is the mysterious blackbox algorithms. I am astonished that nobody questions about this.
Can we get the podcast in video form on Spotify please 🙏
Decade? You can say that, we're basically there.
Just curious , in a differential piece game like chess is there any role for lower order /complexity auxillary modelling purely on a board demographic , ie positional state .
This is fundamentally the last technology humans will ever invent. Once AGI is achieved and released, it will make all of our future inventions...