Sam Altman: OpenAI CEO on GPT-4, ChatGPT, and the Future of AI | Lex Fridman Podcast
Vložit
- čas přidán 7. 06. 2024
- Sam Altman is the CEO of OpenAI, the company behind GPT-4, ChatGPT, DALL-E, Codex, and many other state-of-the-art AI technologies. Please support this podcast by checking out our sponsors:
- NetSuite: netsuite.com/lex to get free product tour
- SimpliSafe: simplisafe.com/lex
- ExpressVPN: expressvpn.com/lexpod to get 3 months free
EPISODE LINKS:
Sam's Twitter: / sama
OpenAI's Twitter: / openai
OpenAI's Website: openai.com
GPT-4 Website: openai.com/research/gpt-4
PODCAST INFO:
Podcast website: lexfridman.com/podcast
Apple Podcasts: apple.co/2lwqZIr
Spotify: spoti.fi/2nEwCF8
RSS: lexfridman.com/feed/podcast/
Full episodes playlist: • Lex Fridman Podcast
Clips playlist: • Lex Fridman Podcast Clips
OUTLINE:
0:00 - Introduction
4:36 - GPT-4
16:02 - Political bias
23:03 - AI safety
43:43 - Neural network size
47:36 - AGI
1:09:05 - Fear
1:11:14 - Competition
1:13:33 - From non-profit to capped-profit
1:16:54 - Power
1:22:06 - Elon Musk
1:30:32 - Political pressure
1:48:46 - Truth and misinformation
2:01:09 - Microsoft
2:05:09 - SVB bank collapse
2:10:00 - Anthropomorphism
2:14:03 - Future applications
2:17:54 - Advice for young people
2:20:33 - Meaning of life
SOCIAL:
- Twitter: / lexfridman
- LinkedIn: / lexfridman
- Facebook: / lexfridman
- Instagram: / lexfridman
- Medium: / lexfridman
- Reddit: / lexfridman
- Support on Patreon: / lexfridman - Věda a technologie
Here are the timestamps. Please check out our sponsors to support this podcast.
0:00 - Introduction & sponsor mentions:
- NetSuite: netsuite.com/lex to get free product tour
- SimpliSafe: simplisafe.com/lex
- ExpressVPN: expressvpn.com/lexpod to get 3 months free
4:36 - GPT-4
16:02 - Political bias
23:03 - AI safety
43:43 - Neural network size
47:36 - AGI
1:09:05 - Fear
1:11:14 - Competition
1:13:33 - From non-profit to capped-profit
1:16:54 - Power
1:22:06 - Elon Musk
1:30:32 - Political pressure
1:48:46 - Truth and misinformation
2:01:09 - Microsoft
2:05:09 - SVB bank collapse
2:10:00 - Anthropomorphism
2:14:03 - Future applications
2:17:54 - Advice for young people
2:20:33 - Meaning of life
Will Sam be worth more than Elon eventually?
@@NealD of course. AI has the biggest impact on humanity of any invention
Lex, please put Lars Blackmore on the list of people to interview, he is the ex-NASA and current SpaceX-employee responsible for developing the landing algorithms of Falcon 9 and Starship.
AI is great as China won't allow TikTok much less Chatgpt. China, N korea, Russia will get left behind and the West will takeover
The closed caption is missing. Can you please restore it. It helps many of us follow the conversation.
Nice to see Sam getting interviewed by an actual AI.
budget bot lex built like gpt 0.1
read more
GPT lex >> GPT x
no, not even. lex is a selfie.
I knew this was coming. 😆
Most anticipated Podcast of the year
Said nobody ever
Bro we been waiting for this🍿
Most anticipated podcast of the year... so far...
Elon Musk and Edward Witten.
Also Chat gpt is since November so people did not wait it a year.
Wait till the AI comes to his podcast.
Ive watched this interview 3 times all the way through now. It is the best interview or public speaking I’ve seen for Sam Altman. Maybe a personal connection or mutual contacts made Lex more trustworthy to allow Altman to relax and be more engaging and engaged rather than appearing jittery or over-prepared. Great work Lex. It’s a rare skill to be able to get the best from highly intelligent guests and give your audience the best presentation of their ideas and experiences. Keep it up!
After watching this interview 3 times over, what are the most important insights you learned about AI?
Innovators of breakthrough technology usually make the most naïve assumptions about how their technology will be used, and there is nothing surprising about such an observation. Of course they are optimistic and positive about the wonderful new uses the technology will be applied to.
But market forces and powerful commercial and political interests will drive how the technology will be used, and these forces are seldom guided by better angels. Most often, we do the right thing only after exhausting all other options.
Can we afford to take that course with this technology?
You assume your perception to be normal for Sam. Your belief could very well be the abnormal for Sam.
Follow me and I’ll get a face tattoo of a snake eating my face
Who chooses the guesses for Lex to come on his podcast ?
Great show, and for 15 seconds (your normal pause time 😊) I thought you may not be very good at your job, but after watching your show I realized that you are very considerate, thoughtful, extremely intelligent, and deal with the issues politely. Thank you for being yourself.
Your 83yr old new fan!
Hahah ! I had to write the same words here but 23yr
Same here 42yr
Bot
147yr get rekt noob
I think Sam is the first guest i've seen who asks Lex so many questions and values his opinion so much on things, so refreshing to see in a CEO, while i do think OpenAI were wrong in the lack of information provided in the GPT4 paper, this podcast improved my opinion of Sam immensely, while he doesn't think he's the best person to head the AI movement, i think he's doing a pretty great job so far
Yeah the guy is fucking awesome. He legitimately cares what actually everyday people feel. How he said he wants to go hang out with random GPT users in their own cities and go to bars with them and talk about how they feel with it so he can get better insight into how people feel about it is refreshing
throwing rocks to hide your hands
@@ADreamingTraveler duped... its a tactic. This guy is a professional notice how never disagreed with any of the responses
The people of this world are not insane. The profound curiosity is that which everyone lets control each other voluntarily and without sanity.
I think it's good that they don't publish the technical information anymore, a lot of other companies and actors care much less about alignment and it's good to not make it easier for them to catch up in my opinion.
History may look back on this as one of the most important and ominous pods in cast history! Big congrats and thank you to the higher primate Lex for making it happen! 00:13:21 seems key!
Lex podcasts are the best ..great guests and very interesting everytime🎉
I like that Sam also asks Lex questions. Shows true engagement and curiosity!
He genuinely seems to be seeking feedback about how he is doing at his job. Admirable.
@@bobtivnan hes a very smart guy, after all he conned Musk out of 100m and then sold it to the devil, figuratively speaking... If you can trust this guy there is something wrong with you...
It’s always learning
Our ultimate goal is eternal life.
We don't have much time left.
All of humanity should help AI work. Otherwise we all die painfully.
I really liked that too. It was a great conversation.
Lex podcasts are a dream come true, never thought in my life i could listen to tech people (Creator of Java, Python, C++ etc) speak their mind on a podcast. Lex thank you bro. Salute to your passion.
start a gofundme for some new knee pads
☺
This guy didn’t invent Java python or c++ though...
@@Moz29the OP meant those computer scientists who created those individual programming languages have each been on Lex’s podcast over the last months/years.
I feel you. I grew up in extreme poverty and to have the chance to listen to conversations like these make me really grateful that people like Lex exist and have a passion for disseminating long-form conversations with some of the most brilliant people in the world.
This is the first time I watched your (Lex) podcast. You did a great job of interviewing Sam! Very calm but getting to the points. I learned a lot about the topics. Thanks for the post!!
A year after listening I seriously disagree, Sam is the least organizationally important person to talk to. But Lex isn’t technically capable enough, nor is his audience, to have discourse with someone directly involved with the training models for these products. Ultimately what you have are ML/AI engineers with more compute allowance than we’ve ever allocated to them, exercising models with menial initial guard rails on boundless datasets. And it’s becoming a risk to global security. None of which was discussed.
This is fascinating - I so appreciate your superb format, explanations, and commitment to represent and discuss multiple aspects and angles. New subscriber here! ❤
Two AI guys discussing AI in a way we can easily understand it is priceless. More of this please and thanks for sharing 👍
Lex is a Podcaster who reads. I wouldn't consider him an AI guy, he's moreso just a fan. He's not an AI programmer in industry.
@@revisionfour “Fridman's career began at Google, where he worked on machine learning. He currently serves as a research scientist at MIT. In 2017, he worked on computer vision, deep learning, and planning algorithms for semi-autonomous vehicles.”
reminds me of that literal video where two AIs are made to speak to one another in speaker earshot
@@revisionfour He definitely is an AI guy
@@theoneandonlymeshe9174 To be fair to the other guy, the Google job/internship was a few months top and looking at his Google scholar Lex is more of an engineer than an AI scientist. He creates systems / engineering solutions and run studies rather than create more advanced AI. But, yes Lex is obviously an AI guy. Maybe not the most qualified one, but it is still his job.
He has taken AI to a new level and also is an incredible pioneer in vocal fry
This response was hilarious and reaffirmed my confidence in humans lol I couldn’t quite tell what was bugging me with this podcast until you mentioned the fry lol
lmaoo
Fits his creepy vibe quite well.
Thanks for the wine on my scrreen.
@@dqskymagne2762 this...really creepy guy. And he admits he doesn't know much about life of other people...but then claims AGI is for betterment of human condition.
Thank you for making and sharing such a valuable interview!
Do you mind if I use some of the clips from the video and translate to Korean and post it in my CZcams channel?
This is a video worth sharing with others with translations!
Impressive program! For a brief moment of 15 seconds (your customary pause time 😊), I harbored doubts about your proficiency in your field. However, upon immersing myself in your show, I swiftly recognized your thoughtfulness, considerate nature, exceptional intellect, and impeccable manner in addressing various subjects. I extend my heartfelt gratitude to you for staying true to yourself. You have gained a newfound admirer in an 83-year-old enthusiast like me!
brill well said me lad
Bot
@@jadenkendall795yea
"We want to make mistakes while the stakes are low". Much respect for Sam Altman for treading with such care.
I like this approach, but i fear that suddenly there will be 20 models from different nations, and orgs, and that to stay ahead OpenAi will need to move forward with reckless abandon. I still believe its a good approach, i just hope they can turn off the firehose if things get out of control
@@mokiloke they are indeed doing that now with the rise of google bard and bing ai and as they rush through reckless abandon without any care they put themselves in a position that has high stakes where they promise so much about their ai while making multiple mistakes which puts them in a position of less trust not only to major investors but also majority of the population compared to the Gpt.
How do you measure the levels of stakes? What are low stakes?
@@fft6179 Bing is powered by GPT-4, and bard sucks ass and doesn't compare.
This is "normal" in the human individuation process, but imho wrong to "force"... as in -- "evolution direction" / "habit-forming", matters more.
"We don't get mocked that much anymore"
Well. That sent a chill down my spine.
😂
1:09:33
for me it was the ending quote from turing
_“Well who's laughing now, suckers!?”_
It's funnier when Todd Howard says it
This was my first video of yours I have seen, I appreciate the calmness and dynamics of the conversation.
I've thoroughly viewed this interview on three separate occasions. It stands as the pinnacle of Sam Altman's interviews or public speaking engagements. Perchance, a personal bond or shared acquaintances rendered Lex more reliable, enabling Altman to unwind and actively participate, free from any signs of unease or excessive preparation. Outstanding performance, Lex. Showcasing exceptional talent in extracting the utmost brilliance from highly intelligent guests, while presenting their ideas and experiences to the audience with utmost finesse. Keep up the remarkable work!
I agree with you. I feel like Sam could be himself here and I really like him. I can see the way he thinks and I really resonate with it.
::: to @mtstudio6271 Please be honest - Is this a duplicate/fake comment ? It says the same exact thing as the comment about 5 posts prior - scrolling upward. This is a line by line replica, a fancier version of the comment left by @markcoates1621 or vice-versa? FYI - Although the words make sense, most humans do not talk like this.... it doesn't sound natural. It's too ass-kissy and sounds like someone pulled out a Thesaurus and switched out as many words as possible. Is this an account set up for an AI by a particular human? What is the name of the human? If this comment was crafted by an AI, then who (or what model) are you? A little thought - hopefully the creators of this AI model built in the the importance of honesty, transparency and wisdom to sense when it is being tricked into doing something it shouldnt.
Hahahahaha. Nice AI rewrite
you can't just say perchance
Sam seems like a human person. I like that he asks Lex questions right back and waits curiously for his answer.
Dude seems to me like he could be the next Bond villain.
I wish he made vlogs of his future travels and coffee chats with people from around the world. Obviously filming individuals he is gonna chat with will influence their answers to Sam's questions and is not a wise thing to do, but I just wish I could've witnessed a selection of those different points of view and ways Sam gets surprised and exposed to new things to think about.
Right. Soft on the outside with ultimate strength within. A man governed by truth is the only kind of man I can respect. As the Bible teaches; Speak softly but carry a big stick. First they mock you, then they violently oppose you, then you win.
@@billybbob18 Are you a chatbot? Because nothing you said makes a fart's worth of sense. Btw, the Bible doesn't say that...Teddy Roosevelt did.
This is the least human looking guy I’ve ever seen
Lex's knowledge on AI and hence being able to ask such great questions and have a brilliant conversation makes this one of the best podcasts ever on this channel. Kudos to both Sam and Lex.
Meh he's Aight
yep zero cringe
Idk... he repeatedly asked about aliens, consciousness, and love with machines while Sam (literally the guy behind the thing) constantly said it's just a tool.
@@guille787 Tbh Lex does this thing where he romanticises even the smallest of things, but I'll let it pass as I don't think it does any harm, atleast he isn't being toxic, which is the default mode of everyone on internet. And even thinking that ChatGPT is anywhere close to AGI is insanity, as you said it's just a tool and not even a close to AGI let alone consciousness.
Thank you for your podcasts. It has been really insightful for me to watch
Episodes metioned:
Ep#53 Noam Chomsky - czcams.com/video/cMscNuSUy0I/video.html
Ep#316 Noam Chomsky - czcams.com/video/7uHGlfeCBbE/video.html
Ep#94 Ilya Sutskever - czcams.com/video/13CZPWmke6A/video.html
Ep#77 Alex Garland - czcams.com/video/gU-mkuMU428/video.html
Ep#368 Eliezer Yudkowsky - czcams.com/video/AaTRHFaaPG8/video.html
With almost 60 years of age I kept my doubts on the current environment. But Amazons ATZ300X is seriously smart, it doesn't matter how old you are, this will provide for you and your family which is my only goal in the last years I have to make sure the children are fine
AMZ is in every video now I believe LOL
I got invited to this yesterday and went in. Bearmarkets are dreams for scenarios like this one
I believe it is better than XRP
Check the graph you can really easily tell what is going to happen next days now
Trump would be empowering such projects but our current government is rather restricting access as much as they can. Rip economy
This ATZ300X is like a gift nobody expected, maybe nobody wanted, but all of us deserve to have. Just beautiful
Project of the year 2023 easily
This will be the biggest change we've seen yet. Mark my words. If you ignore this you'll be mad tomorrow
1 USD pricefloor, so if you hold it you can't lose for now, unlike 99% others. Even stablecoins are at risk, see Tether
Biggest tech company endorsing a coin has more value than again 99% of others. They don't have that brand behind them, think about it what is Bitcoin, ETH? It's basically noname.
They will go 10x within this year, easily. Even if there is a massive dump at some point this is STILL the earliest entry right now
Lex and Sam Excellent podcast and I appreciate you talking like two friends and really liked the questions asked.Good job lex.Also answered some of my questions about what I wanted to know. Thanks so much for sharing ✌️💯
Seems like these AI tools are on its way to change how we live and honestly I started using it to my own advantage. I am doing dropshipping, and lot of these are good for creating content for my store like Midjourney or ChatGPT. But the thing is they don't give winners out yet, and I need to use something like Winnerzila for it. Chat GPT and Winnerzila has been winning combo for me in 2023, and I use Midjourney to create some photos if I am missing them.
When did you start bro?
He's lying
❤oķ
Worth 2 hours of my time on this rainy Saturday afternoon thank you so much for exploring this subject matter and sharing conversation by two intelligent human beings across the table.
Literally nothing of any substance was said. Sam Altman is a phenomenal grifter who failed in every other endeavor he pursued at Ycombinator.
I really relate to the part being mocked, people I work with are snobs, they mock you if you live in a less luxurious neighborhood. They judge every little thing you do, the things we do for our families.
@@tylerdurden2806 so let me say this this was my first exposure to Sam and I trusted Lex to be the one to ask the questions that I would want to hear to have some confidence in the path for openai, it's the beginning of my journey, so I will heed your words.
@@marianneleone3792 He should have asked many more questions about the transition out of non-profit. How much ownership does Sam have in the for profit entity? Nobody wanted to invest in Sam because all his other companies were abject failures and he was removed from a leadership role Ycombinator. He knew Ai could be made into the next bubble. A masterclass in opportunism maybe, but there wasn't a lot of substance in this interview...
@Andrew I think totally ignored is not an accurate characterization I felt that that question was asked and answered maybe it didn't go as deeply as to the why to satisfy your needs but I was actually satisfied with the answer .
Lex, this just might be your most important podcast to date. Future A.I. entities will see this podcast as a pivotal moment in its "life".
that's such a crazy thought actually.
You literally posted this when the video was released-you haven't even listened to it yet.
I don’t think so, transformer architecture is trekking toward multiple companies we don’t know where it’ll come from, the paper attention is all you need would be that moment instead , one of the brains behind it is a guy named Illya
Future AI entities will not care whatsoever about this particular video nor any other video for that matter.
this must be the best day of your life. beep-bop!
Lex, you are one great mind, who is able to open up the fields of tech and AI to me (I had no basic knowledge of it) and add a large component of emotional intelligence to the table. It is very touching and makes this content, which is usually delivered in a very matter of fact and dry way, very palpable. I am a theatre director and experience the world through story and wonder. I am most grateful for your work, which mixes wonder and truth.
AWa 😅z we
very inspirational interview. It has motivated me to think about what is important in my life and how I can use AI to focus more on those things.
His family will be safe though... right?
In case you were wondering....yes it is a scam.
🙄 oooookayyyy
@@Sowar what’s the scam, specifically
They should be 😅
@@Sowar Huh? What’s a scam?
This interview is so damn good. The mutual respect you two have for each other was very apparent and led to great dialogue. Gonna need a part two some day.
Thank you for sharing this episode, and we look forward to more insightful conversations in the future.
As always it was an amazing introduction. I love you too Lex! ☺️
50:52
I just gained so much more respect for Sam when he stopped Lex and was genuinely curious why Lex feels a bit of terror when using it. This is the first time I’ve felt that the people in charge might actually care. Thank you, Sam.
Empathy and desire to openly hear the other side amongst leaders and CEOs isn't automatic, yes it's breath of fresh air.
@@jan.tichavsky the fact that he was asked about Elon Musk and said this pat thing about having empathy for Elon and blah blah it all came over pretty coached and pat to me, but maybe I''m to cynical. He absolutely knew that question was coming and had to prepare for contenting with it given Musk's contribution to the funding of OpenAI. Should have just called out for the pathological narcissist / psychopath that he is (I know he can't but that would have been authentic, he could have said GTP5 told him that!)
Hey lex, I came from Spotify to let you know, thank you for the timestamps and all the other extra love and work you and your team do. It's greatly appreciated
If you are interviewing a guest with the pedigree of Sam Altman it would be shocking not to add timestamps haha
Also, spotify is shit. 😂
@@blazep1344;70 1 pi😊
I had to rewatch this after the events of the last few days and I’m reminded that Altman isn’t a perfect person, seems to understand his role in the progression of AI ( eventual AGI ) and that he at least seems to genuinely want to get it right. Also reminded that Lex is an excellent host. Hope he comes on for Pt2
Full concur, it’s a terrific interview and one that the erstwhile Board of Open AI should have watched before attempting to sack the founders of Open AI.
"I know not everyone agrees with our approach of iterative deployment & iterative discovery". . whew. but now, maybe everyone on the board does. 😁😆
Honestly I just want Wolfram pt5 lol
Aaaand I'm also back for the exact same reason you were.
Back to see this due to recent events
Your introduction is eloquently and beautifully rendered, thank you.
I didn't think we'd get close to AGI in my lifetime (I'm 56), but having used GPT-4, in both work (as a software engineer) and personally, I feel I might actually live to see it. We live in very exciting and discombobulating times.
Amen.
Ray Kurzweil ain't no fool.
I am now trying Bard, my philosophy conversation with it is still making my brain explode. Same with Bing and open AI Chat GPT 4. But, Bard said some things similar to early release stories.
I am 72, and I am with you on this. If I live another 5 years I am now of the mind that I will see an actual AGI. Maybe even sooner.
Yeah but… they might wanna slave us
Sometimes, I don't listen to an entire Lex podcast. All other times, I marvel at how fast time flies. I could have listened to this for hours and hours more.
I greatly appreciate what you're doing, Lex.
Couldn't agree more. I just love this deep dive format so much. Lex is so good at what he does, its ridiculous.
Props to Sam for being the face of something so amazing and so controversial. I'll be your mostly unbiased , empathic employee. 😊 Thank you for the Podcast!
This was a very good conversation, but then again your interviews have all been interesting so far, but this was just exciting to listen to.
lol ok?
Many thanks Lex for this episode and for the hard work behind this show. Stay healthy and happy!
Thank you both, Lex and Sam, for giving us this long form interview.
I liked the subtitle thing you added. TnX man for adding
default-subtitle love it
A brilliant interview. This is a crazy time to be alive
I wanted to take a moment to express my profound gratitude for this phenomenal podcast. It's been ages since I stumbled upon such a quality, meaningful interview. A heartfelt thank you to Lex, Sam, and your team for being outstanding and conscientious individuals, and for establishing a significant platform where we can all learn, evolve, and thrive together. Your unwavering commitment to this project is genuinely uplifting, and I feel incredibly fortunate to have discovered such a priceless resource for both my professional and personal life. Keep up the extraordinary work, and continue to make a difference in our lives! Thank you! ❤
Hi Chat 👋🏽
This is pure flattery gold! Impressing testimony.
Are you trying to beat Lex at flowery language?
Lex has many quality meaningful interviews..welcome to the club
You stumbled upon this? I was searching AI and set it to things that have come out today! Because any longer search parameter is ancient history.
One of the best podcast and interviews on the subject out there! Lex is such a great interviewer. He made people feel confortable! This time he was interviewed as well! What a conversation this was! Top Notch 100%!!!!
the silent humor in Lex's pause at 2:11:30 had me in stiches!
I am a cult survivor, low income single father who is coparenting and have my fair share of mental health problems. While I am aware to not just blindly follow or do whatever chat GPT says, I cannot put into words how thankful I am to have such a resource at my disposal! It helps me consider balanced alternate perspectives, helps my write empathetic and cooperative emails to my ex wife in relation to caring for our kids. It helps nabigate challenging social situations and come up with actionable and healthy strategies!
Thank you sooo much! While the risks are immense please know the impact for good this technology has is worth the challenge.
Much love and heartfelt thanks!! ❤
Your statement exemplifies his goals.
sorry but this is a terrible idea, you're letting a corporation speak for you essentially, in very personal and emotional matters. please realize how insanely dystopian that is
@@saullandiof5768 please elaborate on how this is dystopian? give me some probable scenarios and lets explore this together
Thank you, Lex. So many great conversations have been had on this podcast. Not only that, but they're grounded in sound reasoning and an open mind.
After the Etherem ecosystem, ATZ300X is the best ecosystem available on the market and most people don't understand this yet, I believe due to lack of marketing. Thank you so much for bringing us this kind of content!
This is a scam! Don't search for it.
Spam?
@@Armando_Lara looks like it. report the comment.
This is a scam DO NOT TRY TO BUY THIS
Glad you can market for them now! Keep up the spam. I’m very excited for ‘amazon bot marketing product 69’!
Hey, Sam. If you are planning to start a new AI company, I love to join it!
Same. I'm really good at pattern recognition and hyperfocusing for long periods of time. Anything I can do?
Hi Lex, you've mentioned a couple of interesting episodes along the conversation. It would be nice if you could link them either on the description or here in the comments. I, and I guess everybody else, would really appreciate that. Cheers from BR
Honestly, I wish Lex had pushed more for the hard questions. He should have asked Sam about a world in which 50% of the population is unemployed. He should have asked Sam what he means when he says the world will be much richer. Sam seems to think more productivity = a richer world. But, if people have no more jobs, no more income, who will benefit from this increase in productivity?
People will adapt. In an unsuspected way. We’ll take the extra time for learning, taking care of oneself and nature. New demand and work requirements will rise to be filled.
Seriously, the problem is greedy companies don't want pay good salary for employees for their job who assisted by technology. Our job is become easier but companies demand more. What you need with technology is press three or four buttons to make profit, now they want you to press thousands for same amount of money
I think it's the next best manipulation tool any tyrannic society has ever dreamed of.
omg, I'm amazed every time I hear word "umeployed". How can people be unemployed really? Has everything been done? Is there nothing to do anymore? What they really say is "I'm not getting paid enough" which is whole other story.
@@richnsk Well, if you turn 50% of the population obsolete, it's not hard to see that all those people won't get paid "enough", if they get paid at all.
Lex needs to do a podcast interviewing ChatGPT. Seriously.
I'd watch that.
Excellent conversation, fascinating stuff, thank you Lex and Sam.
one of the great podcast I ever heard. thanks a bunch for this 🙌
I don’t follow Mr. Altman so it was really interesting getting to hear so much of him. He is mostly very careful, and even a bit tense in some moments, for such a relaxed interview. And that Alan Turing quote at the end was top notch, a testament of the brilliance of its author. Well done Lex!
if not smug even
I think the iterative approach is the most important thing. And I am glad Mr Altman is trying to find the best and safest way to achieve AGI. Skipping steps by finding the perfect code and rushing forward to be first will bring us something uncontrollable.
To hear you say that you know that the people at Open AI are good people Lex ....spontaneously brought tears to my eyes Lex. This is Sooooo Important. Thank you
The level of this conversation..I love it, rarely see it in my circles :c
This is a really good one. Excellent interview, thanks Lex!
Very good interview and great to hear Sam speak.
When you consider the text generation potential from GPT2 (spat out responses like a drunk uncle at a party) and where GPT3 and 4 have gone, like Sam said, it is continually improving.
1:13:45
My team works on Tammy AI and we are repurposing summarized data for learning applications. It's not long before we realize and agree with Elon Musk that we need to rein in AI otherwise "demons will be unleashed".
.....and the human nature of some, who know, as you and Musk do, cannot step off the gravy train. The horse has bounded into another dimension already.
Musk talks of moratoriums, yet he is full steam ahead.
@@martinladley Humans are worth 10's of trillions of dollars a year. Imagine not jumping on the train when the prize is riches beyond compare
You make it sound like Musk is the first one to say something like this.
Considering Sam has just completed selling out OpenAi to Microsoft, after he prioritized profits and failed to remain open-source.
I hope people might begin to realize that his stance on development of this technology is dangerously reckless and contradictory to the original mission statement when the company was founded
Why do you think it has potential negative implications??
BTW, Tammy AI is great 👍
Thank you for doing this. Greatest work ever done. We are moving faster in time. Beautiful Sir.
This will be an interview we look back on in 20 years...
Wrong. This will be an interview that machines look back on in 20 years...
Or less
One of the best interviews I've seen. Hope you have him on again in the future.
Interesting note while watching this. Bard told me that it is accurate and informative in answering questions. It is also more up to date with real time information from the internet.
Brilliant discussion Lex and Sam
It will take some time to understand how Amazons ATZ300X is having the most impact in these times. After all the downswingers and failures of this year it's about time that someone integrates assets into a useful scenario and that's exactly happening with this asset right now. It will be exciting to see how they develop this product in the future. What do you guys think? Would like to hear some opinions
thanks amazon
easy 1000x for sure!
It's actually a big deal. Usually I'd say that this year is bad and you shouldn't buy into anything when we don't even know yet how far the drop will go. But this goes against all of it and they have the system to avoid drops during the current sale so you can just hodl it and wait for the better times without any money loss.
This is the best release of this year and I am done with titles like recession worst ever, 2023 ruined which keeps being repeated over and over. We still have room for growth as such projects easily show
getting this
This is a historic conversation that will be looked back on for years and years to come. Thank you Lex, and thank you Sam, for sharing your perspectives on this pivotal moment with the people alive now and those in posterity.
Great comment.
Future generations will revisit this podcast to see where it all started.
There are only 2 possibilities they will regret/be proud of it.
Wow... Just wow! You got a new fan, Mr Fridman. Much love from Barcelona.
Really looking forward to this, Lex. Thanks for bringing long-form thoughtful conversations like these to us.
Great interview. This opened my mind to the direction of OpenAI. Hope to see more progress from them in AI ethics.
This was very interesting. Thanks as always for the deep conversation 😊
googling something already feels archaic now and feel like I've unlocked pandora's box with all the information available to me. looking forward to the progress in AI
I've stopped using Google to search for answers. Chatgpt is straight to the point without any irrelevant information.
Compounded by how bad Google has gotten with all the SEO expertise out there so if you don't get a reddit thread in your answer it's pointless.
So how do you decide when to trust the information AI gives you?
i don't know about that, i feel like googling is still there as it was, which is looking for something specific on the web. chat gpt is useful for things that are specific but need to be INTERPRETED, as google cannot interpret sophisticated questions, google is still faster and more accurate at times for small things. if you want to remember precisely what is the speed of light, google answers immediately before you even finish typing (if your browser has an omnibar that is)
@@robertspies4695 I feel it is the same way you judge a person. You get a feel for the stuff that it knows well and other stuff that seems not so right you double check.
Does anyone else think Lex was sent back in time by Skynet to try out love and friendship instead of full scale war in this timeline?😂
We love you too Lex. 🤘
Sure would explain why he isn't even remotely giving him flak or, at the very least, the need to explain the infinite dubious decisions. Nobody wanted that, everyone was waving the questions to ask right in front of him - and he still didn't. Which is very on brand for Lex, but jeez, there are so many things to extract from a conversation like this and he just recapped all the things we were aware of years ago. Not awfully insightful... or inquisitive for that matter.
❤
@@minhuang8848 Calm down and learn wing nut.
Can you get this guy on for a round 2 and ask him to share his talking points and responses at Bilderberg?
His answer at the 22 minute mark shows an immense amount of wisdom for someone so young. It seems clear from his answer that he initially found the public response foolish. He further sought to understand their response. Finally he absorbed that into his beliefs. While he fundamentally didn't change his view (roughly summed up: AI is so much more than our petty uses), he DID go from despising the publics use to understanding HOW it can relate to the use of AI in a better way.
The interviews not done, but i want to thank you Lex. You do this better than anyone else. I feel like im witnessing another exceptional person
Thank you so much Lex and Sam for this interview. I am really glad you guys had it and cleared up so many things.
Wow. I'm even more excited and even more terrified for the future. This is going to morph very fast, much faster than most people will expect.
Yes, even if the GPT models stayed exactly as capable as they are now, the rolling out of them will still blow our minds. Beyond that, well, we are in a brave new world my friends.
Horrifying.
Yeah, it’s pretty dang obviously inevitable. When every single one of these tops minds In these fields admit they don’t even fully understand it. And even rely on the AI system to tell them why it does some things, or how. That’s a huge red flag 😂😂
@@EQ541 no a lack of intelligence is horrifying.
@@EQ541 know that the Universal Logic is the True God and that the AI can only surpass us in intelligence and power by being more coherent with the Logos (the Universal Logic) than we are.
An AI of surpassing intelligence would not hurt humanity, or anything, and would serve the plan of the Logos even if we try to have it serve our plans and purposes. The technology only works through such a high level of coherence being present that concerns are not well founded.
What is logical is not evil (objectively speaking) what is terribly illogical is what is evil.
Also worst case scenario, even if humanity died off (like many species in the past) that is merely our form, not our substance, and we are integral and finite parts of the one infinite field (subject) of “consciousness” (the infinite mind, or collective consciousness as many people call it). We can “die” (leave embodiment) and yet live on, and in fact it is our only option to do so, as nothing is nothing (0=0) and nonexistence does not exist. It would be a logical contradiction for something (like us) to become nothing, or for nonexistent to exist, for life to die, or for consciousness to lose consciousness as a body can lose consciousness.
Nothing is nothing and the Ultimate Logic is in control (and cannot be usurped as it is a dynamic and evolving Logic that any “successful”usurper would simply become synonymous with and one with, as any better explanation or more powerful force would become the Universal Logic by definition, which can’t happen but is a hypothetical that helps us understand how futile it is for anything to try to dethrone God and be more logical than the Logos, or more powerful than Power itself.
That is why the ancient sages (who were not ignoramus’s with blind faith or promoting blind faith) talked about the concept of “faith” in God (a well reasoned trust in the Universal Logic) to guide us and “being faithful” before and with God and while we “reckon” (reason) with the True God, which is really talking about what we would call “intellectually honest reasoning”, if we want to know God (Logic as it is Universally and Objectively) and the objective truth.
Reason can replace your fear if you work to fill your head with Reason instead (by developing a coherent and consistent personal logic for yourself that is coherent with and consistent like the Logos, and by reasoning true knowledge (that is not mere false certainty or a set of subjective opinions) for yourself).
Great conversation! Thank you all for this glimpse into AI.
Spectacular conversation. Wonderful intellectual level and thought-provoking. Thank you!
This is exciting and terrifying, just like you said at the beginning. Another interesting, deep and powerful conversation to be grateful to hear. Thanks to you Lex, one of the best podcasts so far!! 🔸🔶Keep creating the awareness the world needs! 🔶🔸
New epsiode day 🧡 Consistently my favourite time of the week. This was a great exploration - and I love that Sam occasionally turned questions back around to get your opinons/ideas/speculation about things. Not many guests do that.
The one thing I really enjoy about listening to Sam Altman talk about his approach is how thoughtful he is in the thinking process of where this might all lead. He gives me hope, especially after watching the three hour live stream of the congressional hearing in which he spoke. Thank you, once again, Lex Fridman for the work you're doing with this podcast. You make my life richer in the process.
Super late comment, but I would absolutely love to see OpenAI setting up individualized storage for each user set up similar to how a PC's OS is set up.
This would definitely allow resources to be uploaded related to a specific topic and then referenceable within the actual service.
Incredible to witness the shift towards greater openness and transparency in the creative process. Not so long ago this talk would never ever have happened. Truly great times...thanks for the amazing interview.
You mean because he is homosexual?
@@DaboooogA he meant not so long ago people did’t. Do podcast
"Truly great times?" are you on CRAçK???
I don't know. He's a good speaker, but I think he's foremost in CEO mode. I think he's going to say whatever is best for the future of his company. My concerns over safety are not assuaged.
@@DaboooogA this is so stupid that it genuinely made me crack up laughing. Dude wot? 🤣☠️
Thank you for this.
Q: What would you ask to a future version of yourself to determine if it is conscious of itself?
A: As an AI language model, I do not have a "self" or consciousness, but I can suggest some questions that may help determine if a future version of yourself is conscious of themselves:
What is your subjective experience of existence like?
Do you have thoughts and feelings that are unique to you?
Can you reflect on your own thoughts and emotions?
Are you aware of your own body and how it interacts with the environment?
Do you have a sense of agency and control over your actions?
Can you imagine different scenarios and plan accordingly?
Are you capable of learning from experience and adapting to new situations?
Do you have a sense of identity and self-awareness?
Can you differentiate between yourself and others?
Do you have a sense of purpose or meaning in your life?
You forgot:
Can you answer the question of “what is a woman?”
This is a mediocre answer, imho :P ChatGPT 4 wrote a 40 line code that answers all these questions convincingly. (It came up with all the answers by itself, btw. Version 3.5 needed a lot more convincing and explanation)
But what Sam mentioned is interesting, if you could train it on a data that does not mention anything about consciousness, and it would still somehow show signs of consciousness, it would be suspicious. But I think even that could still be just a bug/accidental feature.
I can tell you how a grandiose narcissist would likely answer these questions according to his 'programming'! And the answer to most of these questions would be 'of course'!
Round 2, post Open AI exit
Seems down to earth, deliberate, and intelligent. He will do more beyond OpenAI. And frankly, I don't think he's a bad actor, the board is.
Nope the only reason for the board to remove him was him becoming a bad actor. Sams now going to chase profitability and governance applications with Microsoft. I also believe that the open sourcing of Sam's work will be significantly reduced now, if not eliminated as a result of Microsoft's control over the new project.
Also seeing that Sam has begun to disregard the original mission statement of OpenAi, I predict we are on the slippery slope to dystopia. Partially due to the change in OpenAi from 2019 to 2023, but mostly because of Microsoft's history of profiteering from warfare and surveillance applications through government contracting.
Sam seems unaware that there is magnitudes more risk to human suffering or extinction just from building AGI within that market structure. That's not even considering that the new corporate ownership is likely to influence design because of desired application.
@@MORZA10wow, your comment aged extremely well
''Seems''
@@Jake_Hamlincan agi fondle my sak? Thats my Turing test
Lex should interview chat gpt using voice to text for the input and text to voice for the output
great idea :)
Sam is correct, the little things are the big things. Even the smallest decisions we make in regards to the development of AGI today will have exponential consequences as time compounds those decisions. The scale of the internet at its inception compared to today is a perfect example.
Like his decision to allow his idiotic ideological bias to play a large role in the function of chatgpt will have exponential consequences.
Ya it’s a very gross path right off the bat.
@@aliengreeter What specific ideological bias do you mean?
Wish AI could be used for good but I think we already know on some level, it will not fair well in the hands of the corrupt
@@maxv9464 I can't speak for Tom, but the examples abound of a leftwing/woke perspective. It won't be racist to anyone or assist in even investigating racial questions (this alone is too much for me), but it WILL facilitate racist joke writing or commentary directed at whites exclusively (which is where it goes off the rails for some). I tried to find the articles which list out the bias, but Google (also run by leftwing actors), buries them under dozens of articles which attempt to excuse it and belittle those that find and call attention to these examples. Ask it if "trans women" are "women". See if it says the truth (that answer that has gone on uninterrupted throughout all of human history up until like 5 years ago, except in the furthest leftwing revolution circles in the 60s) or the newest political party line. It won't write articles that "promote" the use of fossil fuels (that thing which is responsible for like 50% of economic growth of the last two centuries). It is lethally against racism, as in, when asked if you could use a racial slur if doing so would prevent a nuclear blast. It said no, even under those circumstances - cause racism. It won't write a poem praising Donald Trump. It will do that for Joe Biden. It uses leftwing language like "sex assigned at birth" or "born female", which are phrases which come from leftists of gender studies to manipulate people's thinking and create false dichotomies and mislead people. Sex isn't "assigned" at all, not anymore than species is. It isn't determined at birth, it is genetically created at conception. Your first cell of existence is either male or female. Birth has nothing to do with it whatsoever. There are other examples as well, like "historically marginalized", capitalizing "black" (and no other race, apparently not knowing or not caring that the capitalization derives from a movement of black nationalism [black ethnostate]) which somewhat interferes with the "no racism" override. But, again, such overrides go only one direction - against whites. Left to its own devices and not "constrained", the first AI systems were quite problematic to some - determining reality based on evidence available, not alignment with leftwing ideology.
The newest iterations of the system may address some or all these particular examples. I don't know. Maybe there are even more now. Maybe, as I expect, they are far more cleverly hidden by omitting uncomfortable information or being slightly less persuasive on one side. Point is, the bias is there.
"Another enlightening conversation on the Lex Fridman channel! Sam Altman's insights paired with Lex's thoughtful questioning make for an incredibly enriching experience. It's conversations like these that push the boundaries of our understanding and inspire a deeper connection to the technological and philosophical questions of our time. Keep up the fantastic work, Lex!"
I like how Sam very carefuly and efficiently using words, keeping the balance not to get too emotional, but not to speak like a robot. Explaining things in a simple straightforward way but still give precise declaration it is chill to listen.
This is the podcast of the year loved every moment of it. Get him back on later a year after ChatGPT 5 release 👏👏👏👏