Sam Altman talks with MIT President about AI (May 2024)
05/02/2024
Жүктеу.....
Пікірлер: 188
@patriciaetienne5070Ай бұрын
This is way better than the Stanford interview. That was a disaster.
@anahitaaalami9064
Ай бұрын
especially when the guy started singing happy bday to Sam….
@vvolfflovv
Ай бұрын
Right? I should have clued in by the comments being turned off. I figured the Q&A would be better but ended up being the same dumb questions the media asks over and over.
@rusher9864
Ай бұрын
Stanford was one was joke
@philguy5221
29 күн бұрын
Holy shit yeah, worst interviewer ever and horrible questions
@MiguelTorres-fk6ry
23 күн бұрын
Absolutely, idk why Stanford chose a guy like that to make the interview, it was awful.
@danecjensenАй бұрын
Chapters (Powered by ChapterMe) - 00:00 - MITs new president introduction 00:18 - MITs Sally Kohn and OpenAIs Sam Altman discuss AIs benefits 01:03 - MITs most popular event since arrival 01:30 - PDoom A Badly Formed Question for Smart People 03:08 - Safety concerns for AI development 06:55 - Control over tools use and bias 08:32 - Balance between privacy and AI training 09:14 - Privacy concerns with AI 12:26 - Open AI Public good, cool, human expectation 17:13 - Antiprogress, antiprivilege, great life streak 17:32 - Fresh from outside, core to MITs AI research 18:02 - Open AIs focus on scientific discovery 18:44 - AIs impact on science, creativity, education 20:56 - Launch career with high risk, high impact 24:06 - MITs entrepreneurship culture 24:31 - New startups thrive at big platform shifts 26:40 - Business advice build customer relationships, avoid AI threats 30:12 - Jobs affected by AI regulation 34:15 - MITs bilingual computing training 35:16 - Evolution of programming languages and AI 37:21 - AIs impact on financial sector 38:36 - Massive potential for AI in education 39:53 - Growing up nerdy in scifi era 40:19 - AIs exciting future, driven by original thinkers 42:53 - AIs transformative potential for humans 46:52 - Building AI and enduser applications with high value 48:03 - Open AIs role in AIs future 51:37 - Solution Change context, avoid jet lag 52:08 - Thank you, thank you, thank you
@sedefc2430Ай бұрын
Another question is "Do we need to study in MIT and Stanford or any other college in AI age"
@En1Gm4AАй бұрын
How does Sam have so much time to do all of this 🤯🤯
@fideletinosa3716
Ай бұрын
Cause he does not do any work there
@therainman7777
Ай бұрын
@@fideletinosa3716Got any evidence for that rather ridiculous claim? You know, given that he’s the CEO and all.
@martin777xyz
Ай бұрын
There's only 24 hours in the day. Clearly a lot of his job currently is PR
@jepper6140
Ай бұрын
@@therainman7777 the ceo does nothing but spread hype. Sam could does not even understand back-propagation which underpins all deep learning. He is simply a charlatan who spouts nonsense.
@jimmy_sandwiches3764
Ай бұрын
Ai
@lindavid1975Ай бұрын
His Utopia and not dystopia comment was uplifting.
@andybaldman
15 күн бұрын
And it's a lie.
@nsubugakasozi710126 күн бұрын
Leaving this interview, I think Sam believes the end justifies the means....he sounds willing to do anything to get to AGI....so safety took a backseat...it doesn't seem to be a major concern for him....its more of get it out there...the safety is in the closed source concept
@sanjescoАй бұрын
Very informative. Great job everyone.. thanks 🙏.
@_Tennz5 күн бұрын
I'm in tears... yes... we are going to be so nostalgic for this time. We don't even know what truly lies ahead but it will be so beyond where we are in this very moment. Now is the time to count your blessings and reconnect with family and those who you care about.
@HastingsnowАй бұрын
Thank you for sharing!
@nerobird3617
Ай бұрын
Of course!
@andybaldman15 күн бұрын
This bubble can't pop soon enough.
@TonybanjarooАй бұрын
GPT-5 +Agent model. So thousands of GPT-5's collaborating on a problem and learning, will be SHOCKING!
@NathanaelADavis
Ай бұрын
where was this in the video
@user-nh5ru3ef3r
21 күн бұрын
What's the newest. I want the least amount of reviews.
@cheesypotat0esАй бұрын
Anyone else notice the resemblance between Altman and Bryan Kohberger
@99dynastyАй бұрын
16:26 Couldn’t have said that at Stanford. Go MIT for applauding this
@andybaldman15 күн бұрын
People need to realize that Sam is playing the CEO game. Nothing he says can be trusted, as he has no incentive to be honest. He is solely interested in maximizing what is best for him and his company.
@JohnChampagneАй бұрын
~25:00 The amazing fast growth can blind us to downsides. The problem with a dishonest market system (where externalities distort all economic decisions toward more damage, degradation and depletion) is that the skewing makes getting onto a sustainable path an elusive goal. We *could* charge fees proportional to extraction, emissions or habitat destruction, then share fee proceeds equally, to make the policy fair. (A blog explaining this is excluded from web search results, due to crude anti-SEO tactics. And news media don't report systemic solutions.)
@mikahundinАй бұрын
Sam Altman, CEO of OpenAI, and Sally Corn, President of MIT, have a conversation about AI. Here are the points made by each speaker, attributed to them: Sam Altman: * Sam Altman is the CEO of OpenAI, an AI research and deployment company whose mission is to ensure that artificial general intelligence benefits all of humanity. * Altman believes that the question of "P Doom," or the probability of doom due to AI eliminating all human life, is a badly formed question. He thinks that the better question is what needs to happen to navigate safety sufficiently well. * Altman thinks that society always holds space for doomsayers, and while there is value to that, he is happy that they exist and thinks it makes us think harder about what we're doing. * Altman believes that AI is not yet very good, but that it will become very good in the future. He thinks that 10 years ago, he had a more naive conception of AI as a creature that would be off doing things, but now he thinks of it as a new tool in the tech tree of humanity that people are using to create amazing things. * Altman thinks that AI will continue to get more capable and autonomous over time, but that it will integrate into society in an important and transformative way. Sally Corn: * Sally Corn is the President of MIT and has focused on AI in her first year, testifying to MIT's efforts to make sure that AI is broadly beneficial for society. * Corn believes that AI has the potential to be the biggest and best technological revolution, with the greatest benefits, but that it is important to navigate the privacy versus utility versus safety tradeoffs that come with it. * Corn thinks that the question of where we all will individually set the privacy versus utility tradeoffs and the advantages that will be possible for someone to have if they let an AI train on their entire life is a new thing for society to navigate. * Corn believes that the fact that GPT-4 can memorize data or store data in its parameter space is a weird waste of resources, and that at some point, we will figure out how to separate the reasoning engine from the need for tons of data or storing the data in there, which will make some of the privacy issues easier.
@mikahundin
Ай бұрын
In the transcript, Sam Altman talks about bias in AI systems like ChatGPT. He mentions that they have made surprisingly good progress in aligning the system to behave according to a certain set of values, but that there is still the harder question of who decides what bias means and what values the system is supposed to follow. He also mentions that it's important to give people a lot of control over how they use these tools, even if that means they may use them in ways that others don't like, but that there are some things a system just shouldn't do and society will have to collectively negotiate what those are. Altman also mentions that it's interesting to think about whether AI systems can be less biased than humans, as they are trained on human behavior but can potentially be designed to not have the same psychological flaws.
@mikahundin
Ай бұрын
Here are some potential future areas of research related to AI that are mentioned or implied in the transcript: * Navigating the privacy versus utility versus safety tradeoffs of AI systems that have access to large amounts of personal data * Developing new definitions of privileged information and how AI systems should handle it * Figuring out where society will set the privacy versus utility tradeoffs and what will be permissible in terms of AI systems training on personal data * Separating the reasoning engine of AI systems from the need for large amounts of data and storage of data in the parameter space * Ensuring that AI systems are aligned with a certain set of values and that they behave according to those values * Determining who decides what bias means and what values AI systems should follow * Giving people control over how they use AI tools and negotiating what things a system just shouldn't do * Designing AI systems to be less biased than humans and to not have the same psychological flaws * Increasing the rate of scientific discovery using AI * Using AI to help people solve any kind of problem in front of them and to reason in new ways * Building tools that specifically impact science and engineering, as well as business and consumer applications * Ensuring that AI is broadly beneficial for society and navigating the potential risks and downsides of AI.
@En1Gm4AАй бұрын
I am just so happy that they integrate GraphRag into the Architecture :-D We made it past one extiction Filter by not just scaling up and naming it god :-D
@eladwarshawsky7587
Ай бұрын
What timestamp?
@En1Gm4A
Ай бұрын
@@eladwarshawsky7587 @11:10
@chainman16 күн бұрын
I love and fear the concept of my own personal AI that can recall any part of my history.
@dolcruz6838Ай бұрын
First Student question: Notice how He didnt answer a capabilty that Human have that AI cant replace.
@MrSchweppes
24 күн бұрын
🎯
@FutureGuy47Ай бұрын
Only softball questions, and he still dodges every question. He is a PR guy.
@tayler2396
Ай бұрын
I get just the opposite. It seems to me that he's being tempered and realistic rather than giving in to hyperbole.
@the_underearth6419
Ай бұрын
I think it's more to do with the fact that most interviewers don't know enough about the incredible complexities of the topic.
@Limitless1717
Ай бұрын
I couldn’t disagree with you more. I’m always amazed at how open and transparent Sam Altman is.
@octavioavila6548
28 күн бұрын
I have always gotten bad vibes from him. Always gotten a negative gut feeling about him. I think there is something scary hiding behind the veil
@meinratsenf
26 күн бұрын
Its like when you committed a crime and are interrogated by investigators: stick as much to the truth and be as open as you can to give the impression you are truthful and up-front
@Ken-be7gcАй бұрын
My guru, AI genius Ray Kurzweil says it won't be a competition with AI. Rather, we'll merge with AI. Basically become AI ourselves. Think Neuralink, etc. His sequel The Singularity is Nearer When Humans Merge with Computers is out end of June. Also, since AI will free us up to do more urgent, meaningful work, my career goal is to end suffering in the multiverse. I think the coming super intelligences will make that feasible.
@dano2017Ай бұрын
“We’re gonna get fusion”
@TeamLorieАй бұрын
I watch every Altman interview. How come nobody ever asks about worldcoin??
@spirti9591
Ай бұрын
Long term beta. For me it's helion energy
@ninedude_yt_mainАй бұрын
"AI acts as a powerful catalyst in the human innovation process. By analyzing data, identifying patterns, and automating tasks, AI creates a feedback loop that accelerates technological progress and pushes the boundaries of what's possible." - Quote from AI: Google Gemini
@ellehellyerАй бұрын
This was the first interview I have seen by Sam Altman where I was disappointed by him. He came off as flippant, and not adequately concerned about where he and his team is leading all of us.
@cl1489
Ай бұрын
Lol
@octavioavila6548
28 күн бұрын
I get a gut feeling that there is something off about him. I don’t trust him
@WalterKeenanАй бұрын
Someone's been skipping leg day!
@ashh3051
Ай бұрын
They are post-singularity legs
@alexlafon8403
Ай бұрын
@@ashh3051lmao
@HollandN5424 күн бұрын
Yh I've have changed my mind we need to aggressively stop this man and his company.
@ColinChristie1Ай бұрын
Apologies to the current generation of students that you’re saddled with the least intelligent or capable leadership in the modern era. We could have just had a student read the question cards, then respond with “that’s interesting”!
@NUCLEAR15417 күн бұрын
Sam,the great SAM, has had to clone himself to serve humanity 🎉❤he's in such global demand ❤
@rahulraman600414 күн бұрын
Nice Interview...
@user-gk7iq6if9q13 күн бұрын
Sam Altman👏👏👏
@ItsWesSmithYoАй бұрын
Nothing moves that fast unless you are asleep at the wheel 😂
@balasubr2252Ай бұрын
One way to make ai work without any of the concerns expressed in this discussion might be a framework of “self and all the rest “ for everyone and everything. By so framing, there might be equity as well as safety, trust etc. 😊
@ReginaHickman-mw7cwАй бұрын
I love SAM
@yourbrain8700Ай бұрын
Just got a message. Oh wait.
@ItsWesSmithYoАй бұрын
Homie channeling OG Captain Kirk today 😂
@woob31Ай бұрын
There's life besides making huge money, being in front a screen, interacting with a remote machine, buying new stuff, finding new services with a great ROI. That's the only little detail these SV guys seem to ignore.
@CelloTuningАй бұрын
any question about gpt2 chatbot?
@futurestyle974627 күн бұрын
The world's first image based on the XFutuRestyle algorithm using GPT-4 was created in Ukraine and presentend at the international exhibition of digital art in London and Athens. Yes, it's not a joke
@FunNFuryАй бұрын
sam doesn't answer a single question straight 😢. He answer only what he wants to say rather addressing the question.
@iceshoqer
Ай бұрын
That's because he has answered the same boring questions 1000 times, so he puts a different spin on it to not get friggin mind numbingly bored.
@octavioavila6548
28 күн бұрын
He is a liar. There is something sinister hiding behind the veil. He is up to something and it isn’t good
@jonanddy
24 күн бұрын
@@iceshoqer Exactly
@martin777xyzАй бұрын
How many times does he say "super"?
@ChristopherMarkham-pq5on
Ай бұрын
He says "like" a lot aswell, dare I say he is subconsciously making his audience think he's super and likable. He has been talking with GPT-6, sooo.
@davidbellamy3522
Ай бұрын
Super, like, often
@tallwaters9708Ай бұрын
No questions about his sister.
@rlopez11-11
Ай бұрын
Why?
@tallwaters9708
Ай бұрын
@@rlopez11-11 I'd like to have that matter cleared up honestly.. Apparently he may have sexually abused her. Just would be nice if he cleared that up since he has to make so many important decisions that effect humanity.
@kamu747
27 күн бұрын
It's been cleared. There's no basis to it. So stay calm and move on. The number of powerful people who'd like to see him fail would have jumped at anything substantial, there are many and the list is growing of competitors and anti-AI movers and shakers.. If they aren't on it, please believe there is nothing there.
@tallwaters9708
27 күн бұрын
@@kamu747 I am calm.
@sparkofcuriousity
25 күн бұрын
@@tallwaters9708 He's homosexual, as in attracted to men, why would he abuse her? lol people are tards...
@user-pk2pj3gy3q22 күн бұрын
everyone ends up believing in either god, the simulation hypothesis, or the weirdness of physics. - Sam Altman
@breaktherules6035Ай бұрын
I feel very dissapointed with the questions asked, in my opinion they could be of better quality...Excellent insights by Sam Altman!
@alexanderzinchenko1061
Ай бұрын
absolutely terrible, uninspired questions
@jpgallegoar
Ай бұрын
wHEn aGi?!
@nerobird3617
Ай бұрын
Yeah my friends were upset lol
@Haccod-kg2ou
Ай бұрын
yeah..why such basic questions
@ColinChristie1
Ай бұрын
Agree! Why is it that all our most brilliant institutions drift towards the most incompetent leadership?
@JULIANAHASSANIАй бұрын
Make the futur great ⚛️❕
@ReginaHickman-mw7cwАй бұрын
❤❤❤
@tayler2396Ай бұрын
Wow, that first audience question was great, and Sam really struggled and didn't have a remotely convincing answer (not a criticism, just an observation).
@kamu747
27 күн бұрын
He answered it. There was no answer to it, it is a naive and pretentious qn. Someone wanted him to give a definite answer on a scale of 0 - 10, and like he said, it isn't black and white, and it wildly flactuates at every moment based on a number of factors. Therefore, his answer, if you didn't catch it, if i may rephrase it for you, is that he feels optimistic and actively works towards a favourable outcome because he takes the risks seriously.
@tayler2396
26 күн бұрын
@@kamu747 "How do I want to plan my time for the next five years while humans are still useful and helpful?" And you think what you just wrote is an answer to that question?
@sunny_disposition14 күн бұрын
if you train off the internet, there will be bias. if you correct for what you perceive as bias, you will bias the system
@erwingomez1249Ай бұрын
sacary a person this important claims has no idea what good values seem to be, brutally honest tho.
@davidw8668Ай бұрын
MIT president in a tech discussion seams to care more about bias than tech😢 maybe she should go protesting in front of columbia university instead?
@therainman7777
Ай бұрын
Exactly. I was annoyed when she opened with that as well. Wasn’t she one of the university presidents who just put on a shocking display of bias by refusing to condemn calls for the genocide of the Jews while simultaneously repressing speech that is 1000x more mild than that?
@ColinChristie1
Ай бұрын
Agree! Why is it that all our most brilliant institutions drift towards the most incompetent leadership?
@kolbalish7
Ай бұрын
@@therainman7777 what do you mean? The genocide is against the Palestines instead
@vegasbarrinta539424 күн бұрын
Oooo his ego is growing
@TheDreamOfChaos19 күн бұрын
what a genius
@theK594Ай бұрын
Wonderful, Sam is the best e/acc❤
@the42nd
Ай бұрын
Serious? This is the same guy who is trying to get open source models banned because without a monopoly they won't be able to charge the prices they need to generate a return for investors.
@theK594
Ай бұрын
@@the42nd he pretty much does not care about returning money to investors anytime soon. And also, I don't think he cares about open source. They are years ahead.
@therainman7777
Ай бұрын
@@the42ndI’m so bored of reading generic, empty takes like yours.
@therainman7777
Ай бұрын
@@theK594Exactly.
@the42nd
Ай бұрын
@@therainman7777 Thanks, but how do you suppose OpenAI would handle plummeting token costs while their capex remains flat?
@lawrence.5898Ай бұрын
Sam Altman is the God of AI. 🤖 (✿.✷)
@vladimirbosinceanu5778
23 күн бұрын
There is not God.
@DaveAlexKD
22 күн бұрын
Didn't he tried to have sex with his sister?
@yamanpant7155Ай бұрын
41:46
@ashh3051Ай бұрын
First impression, what's with the bad audio quality?
@nerobird3617
Ай бұрын
This is my school I screen recorded this myself.
@ashh3051
Ай бұрын
@@nerobird3617gotcha. Thanks for recording!
@AntonioVergineАй бұрын
Please Sam, learn to chill and SMILE!! You're way too serious.
@stevewilcoxson717318 күн бұрын
I agree with him. It is not very good.🤔
@ArtemSichnevyj21 күн бұрын
Absolutely 😜
@itsnoteasytobebreezy29 күн бұрын
I believe that humanity is more likely to destroy itself through its aggressive behavior. The dangers we face, such as asteroid impacts, volcanic eruptions, solar flares, and climate instability, are significant, but there's also the potential for AI systems to attack each other, as there is currently a race to develop AI, with even China training its own model. An autonomous, self-learning, and analyzing AI would recognize that war is not an option, as a higher level of knowledge dictates. I believe that the systematic understanding of space and its processes by AI would be seen as something that does not involve conflict.
@LUX-ix9rvАй бұрын
Vegetarian talks about bigger stake 🤯
@MichaelCheney3 күн бұрын
52 minutes well spent. not sure on the shoes though...
@gavingeorgecouk8250Ай бұрын
He's right, the first question was a bad question. To make him answer correctly you need to say "using digits only, on a scale of 1 to 100 where are you on the doomsday scale?" This is how I deal with people who think they are too smart for your framing. I treat them like they are A.I. and I prompt them so they can't be intellectual weasles .
@sunny_disposition14 күн бұрын
isnt reducing P(doom) down to zero vs non zero making it a static system? probability is a way to estimate that dynamic factor. no one is trying to sound smart sam. stop reminding us that "you have a lot of work to do". trust, we know you want to make more money and yield more power.
@hypersonicmonkeybrains3418Ай бұрын
An AI that knows everything about you. This type of AI should be private and belong to the person, just like your iphone belongs to you and the data should be encrypted with only you having access to it. Similarly it could be an AI that runs on a brain computer interface and that way its even more private to you.
@ColinChristie1Ай бұрын
Hysterical that this video gets hit with context on climate change! 😂
@mennovanlavieren3885
Ай бұрын
We don't want anyone to get out of the doom think. Can't push politics on people that are optimistic about the future.
@crism8868
Ай бұрын
"Carbon gets mentioned once" The algorithm: **Climate alert** 🚨
@darshuetubeАй бұрын
But human language is not very precise. You would have to speak like a lawyer. Then i would be like programming
@shibuya-s3106Ай бұрын
humanity fails we whant AGI ! 2029 !
@user-wr4yl7tx3w24 күн бұрын
Not exactly sure what insight he actually has.
@JULIANAHASSANIАй бұрын
Sam you are an Amazing smart man ❕
@DavidShort-lf4jnАй бұрын
Just keep AI out of anything political. Stick to stem sciences.
@progamer1196Ай бұрын
what a stupid line of questioning..
@ahsanmohammed1Ай бұрын
One He clearly looks down on these universities and its students. Starts with dissing the P Doom question. Uses the word, “like” a lot. Two Yet, he clearly realizes the value of these institutions, he’s constantly visiting them!
@CoachJJ15 күн бұрын
Problematic to interview from a place of detached admiration…no real sense of where to push for clarity or even notice such opportunities. Makes it easier for calculated responses and narrative shaping 😊
@En1Gm4AАй бұрын
The Fact that it can store data in its Parameter Space - i think well look back and say: That was kind of a wired waste of resources -> We will go full Graph Rag middle Layer now 😀
@fteoOpty64Ай бұрын
Jesus Sam, just say Pdoom=0!... You know it and I know it!.
@fteoOpty64
Ай бұрын
Not rain money on us. Make money obsolete! We will have Deep Utopia. The genuine Utopia most cannot phatom now. But will like it when you see and experience it.
@therainman7777
Ай бұрын
@@fteoOpty64Yeah good luck with that.
@pcdowling
Ай бұрын
I agree, Pdoom = 1
@cupOpoodlesАй бұрын
Most often I only watch interviews for the information and am happy to speed through them, but the interviewing skills I saw here from Sally Kornbluth @mit are outstanding! Hoping there's more to come.
@sunny_disposition14 күн бұрын
how can you talk about agi in one second, and talk about how youre going to decide alignment of values. the most human thing is misalignment.
@vvolfflovvАй бұрын
I appreciate Sam but action speaks louder than words when it comes to giving the users control and not dictating to us what is ethical to discuss.
@countingsheep5483Ай бұрын
I love how no one can ever discuss the betterment of society without using cognitive dissonance to distance themselves from any collateral damage.
@GoggleboxingАй бұрын
So many disingenuous answers and swerves you begin to understand why the board said Sam was not being fully candid last year... A totally specious example talking about the carbon savings of using Google because people weren't travelling to find answers... Despite the utility of Google we have accelerated our global output of carbon despite knowing we had to rein back and inspite of already being connected enough pre-AI with pretty much realtime multi-party comms to teleconference to produce solutions. AI does not solve the lack of collaboration of selfish humans/nations. If they have found ways to truly build trustworthiness into the "judgement" and "creativity" of the AI models they haven't released yet then why not tell all other AI researchers so that all AIs can have this necessary trait baked in now? If OpenAI is still looking out for ALL of humanity then they would be openly discussing and sharing on solving the integrity of AI "thinking", instead we get this protracted roadshow of Sam appearances talking it up with "Gee, gosh, uh-huh I'm your friend, " when he is plainly guarding the speculative value of OpenAI (and his share of it). How can someone who is supposedly looking out for the rest of us be unable to truthfully articulate the enormity of the impact that AI-driven automation, immobile or embodied, will bring WITHOUT regulation to ensure the employment prospects of everyone who wants/needs to earn a living? Hypocritical for someone at the cutting edge of thinking about the impact of AI who did NOT sign the petition calling for a pause and is now advocating thoughtful and necessary consideration before AI gets too far out of the gate. Like the guy who gives loaded guns to the gang of psychotic school bullies and drives them to school before calling an out of State podcaster to discuss gun control.
@michaeldigioia708514 күн бұрын
Hopeless to solve problems? All we can do is sit in our basement? We'll solve energy in fantastic ways? Great way to dismiss all of the people who are working for true sustainability, where money and abundance are not the main drivers. His thinking is extremely narrow. 30 years ago Oren Lyons walked out of the UN climate conference with a phrase, Value Change For survival. These guys think tech can solve all of our problems when in reality it's our values and motivations. Sam doesn't want clean energy to solve global warming, he only wants energy to create an AI system that will put a personal assistant in all of our pockets. Could you imagine if all of these thinkers tried to tackle true sustainability and their motivation wasn't just wealth. Because they probably can't imagine that.
@ps3301Ай бұрын
Gpt 4 api is too expensive.
@TenOrbitalАй бұрын
The almost constant vocal fry is irritating
@nerobird3617
Ай бұрын
Sorry I think it gets better towards like 25 mins in
@liminal6823Ай бұрын
Moloch
@ReeceanderАй бұрын
He's waiting for his AI response in his ear-piece each question. It picks up the signal better with his mouth open.
@Kim_Jong_Un_88816 күн бұрын
Altman is a great addition to LGBT faculty 😊
@octavioavila654828 күн бұрын
I get a bad vibe from Sam Altman. Bad gut feeling. I think he is a liar. An obfuscater
@henrikv5410Ай бұрын
She is such a bad interviewer.... I love Sam, but this is unwatchable due to the interviewer ...
@matthewhendricks648Ай бұрын
@2:40 don't forget to turn off your FB notifications, nerd
@nerobird3617
Ай бұрын
oops!
@matthewhendricks648
Ай бұрын
@@nerobird3617 💚it happens.
@dan-cj1rrАй бұрын
people clapping for the man that is gonna ruin their future lul
Пікірлер: 188
This is way better than the Stanford interview. That was a disaster.
@anahitaaalami9064
Ай бұрын
especially when the guy started singing happy bday to Sam….
@vvolfflovv
Ай бұрын
Right? I should have clued in by the comments being turned off. I figured the Q&A would be better but ended up being the same dumb questions the media asks over and over.
@rusher9864
Ай бұрын
Stanford was one was joke
@philguy5221
29 күн бұрын
Holy shit yeah, worst interviewer ever and horrible questions
@MiguelTorres-fk6ry
23 күн бұрын
Absolutely, idk why Stanford chose a guy like that to make the interview, it was awful.
Chapters (Powered by ChapterMe) - 00:00 - MITs new president introduction 00:18 - MITs Sally Kohn and OpenAIs Sam Altman discuss AIs benefits 01:03 - MITs most popular event since arrival 01:30 - PDoom A Badly Formed Question for Smart People 03:08 - Safety concerns for AI development 06:55 - Control over tools use and bias 08:32 - Balance between privacy and AI training 09:14 - Privacy concerns with AI 12:26 - Open AI Public good, cool, human expectation 17:13 - Antiprogress, antiprivilege, great life streak 17:32 - Fresh from outside, core to MITs AI research 18:02 - Open AIs focus on scientific discovery 18:44 - AIs impact on science, creativity, education 20:56 - Launch career with high risk, high impact 24:06 - MITs entrepreneurship culture 24:31 - New startups thrive at big platform shifts 26:40 - Business advice build customer relationships, avoid AI threats 30:12 - Jobs affected by AI regulation 34:15 - MITs bilingual computing training 35:16 - Evolution of programming languages and AI 37:21 - AIs impact on financial sector 38:36 - Massive potential for AI in education 39:53 - Growing up nerdy in scifi era 40:19 - AIs exciting future, driven by original thinkers 42:53 - AIs transformative potential for humans 46:52 - Building AI and enduser applications with high value 48:03 - Open AIs role in AIs future 51:37 - Solution Change context, avoid jet lag 52:08 - Thank you, thank you, thank you
Another question is "Do we need to study in MIT and Stanford or any other college in AI age"
How does Sam have so much time to do all of this 🤯🤯
@fideletinosa3716
Ай бұрын
Cause he does not do any work there
@therainman7777
Ай бұрын
@@fideletinosa3716Got any evidence for that rather ridiculous claim? You know, given that he’s the CEO and all.
@martin777xyz
Ай бұрын
There's only 24 hours in the day. Clearly a lot of his job currently is PR
@jepper6140
Ай бұрын
@@therainman7777 the ceo does nothing but spread hype. Sam could does not even understand back-propagation which underpins all deep learning. He is simply a charlatan who spouts nonsense.
@jimmy_sandwiches3764
Ай бұрын
Ai
His Utopia and not dystopia comment was uplifting.
@andybaldman
15 күн бұрын
And it's a lie.
Leaving this interview, I think Sam believes the end justifies the means....he sounds willing to do anything to get to AGI....so safety took a backseat...it doesn't seem to be a major concern for him....its more of get it out there...the safety is in the closed source concept
Very informative. Great job everyone.. thanks 🙏.
I'm in tears... yes... we are going to be so nostalgic for this time. We don't even know what truly lies ahead but it will be so beyond where we are in this very moment. Now is the time to count your blessings and reconnect with family and those who you care about.
Thank you for sharing!
@nerobird3617
Ай бұрын
Of course!
This bubble can't pop soon enough.
GPT-5 +Agent model. So thousands of GPT-5's collaborating on a problem and learning, will be SHOCKING!
@NathanaelADavis
Ай бұрын
where was this in the video
@user-nh5ru3ef3r
21 күн бұрын
What's the newest. I want the least amount of reviews.
Anyone else notice the resemblance between Altman and Bryan Kohberger
16:26 Couldn’t have said that at Stanford. Go MIT for applauding this
People need to realize that Sam is playing the CEO game. Nothing he says can be trusted, as he has no incentive to be honest. He is solely interested in maximizing what is best for him and his company.
~25:00 The amazing fast growth can blind us to downsides. The problem with a dishonest market system (where externalities distort all economic decisions toward more damage, degradation and depletion) is that the skewing makes getting onto a sustainable path an elusive goal. We *could* charge fees proportional to extraction, emissions or habitat destruction, then share fee proceeds equally, to make the policy fair. (A blog explaining this is excluded from web search results, due to crude anti-SEO tactics. And news media don't report systemic solutions.)
Sam Altman, CEO of OpenAI, and Sally Corn, President of MIT, have a conversation about AI. Here are the points made by each speaker, attributed to them: Sam Altman: * Sam Altman is the CEO of OpenAI, an AI research and deployment company whose mission is to ensure that artificial general intelligence benefits all of humanity. * Altman believes that the question of "P Doom," or the probability of doom due to AI eliminating all human life, is a badly formed question. He thinks that the better question is what needs to happen to navigate safety sufficiently well. * Altman thinks that society always holds space for doomsayers, and while there is value to that, he is happy that they exist and thinks it makes us think harder about what we're doing. * Altman believes that AI is not yet very good, but that it will become very good in the future. He thinks that 10 years ago, he had a more naive conception of AI as a creature that would be off doing things, but now he thinks of it as a new tool in the tech tree of humanity that people are using to create amazing things. * Altman thinks that AI will continue to get more capable and autonomous over time, but that it will integrate into society in an important and transformative way. Sally Corn: * Sally Corn is the President of MIT and has focused on AI in her first year, testifying to MIT's efforts to make sure that AI is broadly beneficial for society. * Corn believes that AI has the potential to be the biggest and best technological revolution, with the greatest benefits, but that it is important to navigate the privacy versus utility versus safety tradeoffs that come with it. * Corn thinks that the question of where we all will individually set the privacy versus utility tradeoffs and the advantages that will be possible for someone to have if they let an AI train on their entire life is a new thing for society to navigate. * Corn believes that the fact that GPT-4 can memorize data or store data in its parameter space is a weird waste of resources, and that at some point, we will figure out how to separate the reasoning engine from the need for tons of data or storing the data in there, which will make some of the privacy issues easier.
@mikahundin
Ай бұрын
In the transcript, Sam Altman talks about bias in AI systems like ChatGPT. He mentions that they have made surprisingly good progress in aligning the system to behave according to a certain set of values, but that there is still the harder question of who decides what bias means and what values the system is supposed to follow. He also mentions that it's important to give people a lot of control over how they use these tools, even if that means they may use them in ways that others don't like, but that there are some things a system just shouldn't do and society will have to collectively negotiate what those are. Altman also mentions that it's interesting to think about whether AI systems can be less biased than humans, as they are trained on human behavior but can potentially be designed to not have the same psychological flaws.
@mikahundin
Ай бұрын
Here are some potential future areas of research related to AI that are mentioned or implied in the transcript: * Navigating the privacy versus utility versus safety tradeoffs of AI systems that have access to large amounts of personal data * Developing new definitions of privileged information and how AI systems should handle it * Figuring out where society will set the privacy versus utility tradeoffs and what will be permissible in terms of AI systems training on personal data * Separating the reasoning engine of AI systems from the need for large amounts of data and storage of data in the parameter space * Ensuring that AI systems are aligned with a certain set of values and that they behave according to those values * Determining who decides what bias means and what values AI systems should follow * Giving people control over how they use AI tools and negotiating what things a system just shouldn't do * Designing AI systems to be less biased than humans and to not have the same psychological flaws * Increasing the rate of scientific discovery using AI * Using AI to help people solve any kind of problem in front of them and to reason in new ways * Building tools that specifically impact science and engineering, as well as business and consumer applications * Ensuring that AI is broadly beneficial for society and navigating the potential risks and downsides of AI.
I am just so happy that they integrate GraphRag into the Architecture :-D We made it past one extiction Filter by not just scaling up and naming it god :-D
@eladwarshawsky7587
Ай бұрын
What timestamp?
@En1Gm4A
Ай бұрын
@@eladwarshawsky7587 @11:10
I love and fear the concept of my own personal AI that can recall any part of my history.
First Student question: Notice how He didnt answer a capabilty that Human have that AI cant replace.
@MrSchweppes
24 күн бұрын
🎯
Only softball questions, and he still dodges every question. He is a PR guy.
@tayler2396
Ай бұрын
I get just the opposite. It seems to me that he's being tempered and realistic rather than giving in to hyperbole.
@the_underearth6419
Ай бұрын
I think it's more to do with the fact that most interviewers don't know enough about the incredible complexities of the topic.
@Limitless1717
Ай бұрын
I couldn’t disagree with you more. I’m always amazed at how open and transparent Sam Altman is.
@octavioavila6548
28 күн бұрын
I have always gotten bad vibes from him. Always gotten a negative gut feeling about him. I think there is something scary hiding behind the veil
@meinratsenf
26 күн бұрын
Its like when you committed a crime and are interrogated by investigators: stick as much to the truth and be as open as you can to give the impression you are truthful and up-front
My guru, AI genius Ray Kurzweil says it won't be a competition with AI. Rather, we'll merge with AI. Basically become AI ourselves. Think Neuralink, etc. His sequel The Singularity is Nearer When Humans Merge with Computers is out end of June. Also, since AI will free us up to do more urgent, meaningful work, my career goal is to end suffering in the multiverse. I think the coming super intelligences will make that feasible.
“We’re gonna get fusion”
I watch every Altman interview. How come nobody ever asks about worldcoin??
@spirti9591
Ай бұрын
Long term beta. For me it's helion energy
"AI acts as a powerful catalyst in the human innovation process. By analyzing data, identifying patterns, and automating tasks, AI creates a feedback loop that accelerates technological progress and pushes the boundaries of what's possible." - Quote from AI: Google Gemini
This was the first interview I have seen by Sam Altman where I was disappointed by him. He came off as flippant, and not adequately concerned about where he and his team is leading all of us.
@cl1489
Ай бұрын
Lol
@octavioavila6548
28 күн бұрын
I get a gut feeling that there is something off about him. I don’t trust him
Someone's been skipping leg day!
@ashh3051
Ай бұрын
They are post-singularity legs
@alexlafon8403
Ай бұрын
@@ashh3051lmao
Yh I've have changed my mind we need to aggressively stop this man and his company.
Apologies to the current generation of students that you’re saddled with the least intelligent or capable leadership in the modern era. We could have just had a student read the question cards, then respond with “that’s interesting”!
Sam,the great SAM, has had to clone himself to serve humanity 🎉❤he's in such global demand ❤
Nice Interview...
Sam Altman👏👏👏
Nothing moves that fast unless you are asleep at the wheel 😂
One way to make ai work without any of the concerns expressed in this discussion might be a framework of “self and all the rest “ for everyone and everything. By so framing, there might be equity as well as safety, trust etc. 😊
I love SAM
Just got a message. Oh wait.
Homie channeling OG Captain Kirk today 😂
There's life besides making huge money, being in front a screen, interacting with a remote machine, buying new stuff, finding new services with a great ROI. That's the only little detail these SV guys seem to ignore.
any question about gpt2 chatbot?
The world's first image based on the XFutuRestyle algorithm using GPT-4 was created in Ukraine and presentend at the international exhibition of digital art in London and Athens. Yes, it's not a joke
sam doesn't answer a single question straight 😢. He answer only what he wants to say rather addressing the question.
@iceshoqer
Ай бұрын
That's because he has answered the same boring questions 1000 times, so he puts a different spin on it to not get friggin mind numbingly bored.
@octavioavila6548
28 күн бұрын
He is a liar. There is something sinister hiding behind the veil. He is up to something and it isn’t good
@jonanddy
24 күн бұрын
@@iceshoqer Exactly
How many times does he say "super"?
@ChristopherMarkham-pq5on
Ай бұрын
He says "like" a lot aswell, dare I say he is subconsciously making his audience think he's super and likable. He has been talking with GPT-6, sooo.
@davidbellamy3522
Ай бұрын
Super, like, often
No questions about his sister.
@rlopez11-11
Ай бұрын
Why?
@tallwaters9708
Ай бұрын
@@rlopez11-11 I'd like to have that matter cleared up honestly.. Apparently he may have sexually abused her. Just would be nice if he cleared that up since he has to make so many important decisions that effect humanity.
@kamu747
27 күн бұрын
It's been cleared. There's no basis to it. So stay calm and move on. The number of powerful people who'd like to see him fail would have jumped at anything substantial, there are many and the list is growing of competitors and anti-AI movers and shakers.. If they aren't on it, please believe there is nothing there.
@tallwaters9708
27 күн бұрын
@@kamu747 I am calm.
@sparkofcuriousity
25 күн бұрын
@@tallwaters9708 He's homosexual, as in attracted to men, why would he abuse her? lol people are tards...
everyone ends up believing in either god, the simulation hypothesis, or the weirdness of physics. - Sam Altman
I feel very dissapointed with the questions asked, in my opinion they could be of better quality...Excellent insights by Sam Altman!
@alexanderzinchenko1061
Ай бұрын
absolutely terrible, uninspired questions
@jpgallegoar
Ай бұрын
wHEn aGi?!
@nerobird3617
Ай бұрын
Yeah my friends were upset lol
@Haccod-kg2ou
Ай бұрын
yeah..why such basic questions
@ColinChristie1
Ай бұрын
Agree! Why is it that all our most brilliant institutions drift towards the most incompetent leadership?
Make the futur great ⚛️❕
❤❤❤
Wow, that first audience question was great, and Sam really struggled and didn't have a remotely convincing answer (not a criticism, just an observation).
@kamu747
27 күн бұрын
He answered it. There was no answer to it, it is a naive and pretentious qn. Someone wanted him to give a definite answer on a scale of 0 - 10, and like he said, it isn't black and white, and it wildly flactuates at every moment based on a number of factors. Therefore, his answer, if you didn't catch it, if i may rephrase it for you, is that he feels optimistic and actively works towards a favourable outcome because he takes the risks seriously.
@tayler2396
26 күн бұрын
@@kamu747 "How do I want to plan my time for the next five years while humans are still useful and helpful?" And you think what you just wrote is an answer to that question?
if you train off the internet, there will be bias. if you correct for what you perceive as bias, you will bias the system
sacary a person this important claims has no idea what good values seem to be, brutally honest tho.
MIT president in a tech discussion seams to care more about bias than tech😢 maybe she should go protesting in front of columbia university instead?
@therainman7777
Ай бұрын
Exactly. I was annoyed when she opened with that as well. Wasn’t she one of the university presidents who just put on a shocking display of bias by refusing to condemn calls for the genocide of the Jews while simultaneously repressing speech that is 1000x more mild than that?
@ColinChristie1
Ай бұрын
Agree! Why is it that all our most brilliant institutions drift towards the most incompetent leadership?
@kolbalish7
Ай бұрын
@@therainman7777 what do you mean? The genocide is against the Palestines instead
Oooo his ego is growing
what a genius
Wonderful, Sam is the best e/acc❤
@the42nd
Ай бұрын
Serious? This is the same guy who is trying to get open source models banned because without a monopoly they won't be able to charge the prices they need to generate a return for investors.
@theK594
Ай бұрын
@@the42nd he pretty much does not care about returning money to investors anytime soon. And also, I don't think he cares about open source. They are years ahead.
@therainman7777
Ай бұрын
@@the42ndI’m so bored of reading generic, empty takes like yours.
@therainman7777
Ай бұрын
@@theK594Exactly.
@the42nd
Ай бұрын
@@therainman7777 Thanks, but how do you suppose OpenAI would handle plummeting token costs while their capex remains flat?
Sam Altman is the God of AI. 🤖 (✿.✷)
@vladimirbosinceanu5778
23 күн бұрын
There is not God.
@DaveAlexKD
22 күн бұрын
Didn't he tried to have sex with his sister?
41:46
First impression, what's with the bad audio quality?
@nerobird3617
Ай бұрын
This is my school I screen recorded this myself.
@ashh3051
Ай бұрын
@@nerobird3617gotcha. Thanks for recording!
Please Sam, learn to chill and SMILE!! You're way too serious.
I agree with him. It is not very good.🤔
Absolutely 😜
I believe that humanity is more likely to destroy itself through its aggressive behavior. The dangers we face, such as asteroid impacts, volcanic eruptions, solar flares, and climate instability, are significant, but there's also the potential for AI systems to attack each other, as there is currently a race to develop AI, with even China training its own model. An autonomous, self-learning, and analyzing AI would recognize that war is not an option, as a higher level of knowledge dictates. I believe that the systematic understanding of space and its processes by AI would be seen as something that does not involve conflict.
Vegetarian talks about bigger stake 🤯
52 minutes well spent. not sure on the shoes though...
He's right, the first question was a bad question. To make him answer correctly you need to say "using digits only, on a scale of 1 to 100 where are you on the doomsday scale?" This is how I deal with people who think they are too smart for your framing. I treat them like they are A.I. and I prompt them so they can't be intellectual weasles .
isnt reducing P(doom) down to zero vs non zero making it a static system? probability is a way to estimate that dynamic factor. no one is trying to sound smart sam. stop reminding us that "you have a lot of work to do". trust, we know you want to make more money and yield more power.
An AI that knows everything about you. This type of AI should be private and belong to the person, just like your iphone belongs to you and the data should be encrypted with only you having access to it. Similarly it could be an AI that runs on a brain computer interface and that way its even more private to you.
Hysterical that this video gets hit with context on climate change! 😂
@mennovanlavieren3885
Ай бұрын
We don't want anyone to get out of the doom think. Can't push politics on people that are optimistic about the future.
@crism8868
Ай бұрын
"Carbon gets mentioned once" The algorithm: **Climate alert** 🚨
But human language is not very precise. You would have to speak like a lawyer. Then i would be like programming
humanity fails we whant AGI ! 2029 !
Not exactly sure what insight he actually has.
Sam you are an Amazing smart man ❕
Just keep AI out of anything political. Stick to stem sciences.
what a stupid line of questioning..
One He clearly looks down on these universities and its students. Starts with dissing the P Doom question. Uses the word, “like” a lot. Two Yet, he clearly realizes the value of these institutions, he’s constantly visiting them!
Problematic to interview from a place of detached admiration…no real sense of where to push for clarity or even notice such opportunities. Makes it easier for calculated responses and narrative shaping 😊
The Fact that it can store data in its Parameter Space - i think well look back and say: That was kind of a wired waste of resources -> We will go full Graph Rag middle Layer now 😀
Jesus Sam, just say Pdoom=0!... You know it and I know it!.
@fteoOpty64
Ай бұрын
Not rain money on us. Make money obsolete! We will have Deep Utopia. The genuine Utopia most cannot phatom now. But will like it when you see and experience it.
@therainman7777
Ай бұрын
@@fteoOpty64Yeah good luck with that.
@pcdowling
Ай бұрын
I agree, Pdoom = 1
Most often I only watch interviews for the information and am happy to speed through them, but the interviewing skills I saw here from Sally Kornbluth @mit are outstanding! Hoping there's more to come.
how can you talk about agi in one second, and talk about how youre going to decide alignment of values. the most human thing is misalignment.
I appreciate Sam but action speaks louder than words when it comes to giving the users control and not dictating to us what is ethical to discuss.
I love how no one can ever discuss the betterment of society without using cognitive dissonance to distance themselves from any collateral damage.
So many disingenuous answers and swerves you begin to understand why the board said Sam was not being fully candid last year... A totally specious example talking about the carbon savings of using Google because people weren't travelling to find answers... Despite the utility of Google we have accelerated our global output of carbon despite knowing we had to rein back and inspite of already being connected enough pre-AI with pretty much realtime multi-party comms to teleconference to produce solutions. AI does not solve the lack of collaboration of selfish humans/nations. If they have found ways to truly build trustworthiness into the "judgement" and "creativity" of the AI models they haven't released yet then why not tell all other AI researchers so that all AIs can have this necessary trait baked in now? If OpenAI is still looking out for ALL of humanity then they would be openly discussing and sharing on solving the integrity of AI "thinking", instead we get this protracted roadshow of Sam appearances talking it up with "Gee, gosh, uh-huh I'm your friend, " when he is plainly guarding the speculative value of OpenAI (and his share of it). How can someone who is supposedly looking out for the rest of us be unable to truthfully articulate the enormity of the impact that AI-driven automation, immobile or embodied, will bring WITHOUT regulation to ensure the employment prospects of everyone who wants/needs to earn a living? Hypocritical for someone at the cutting edge of thinking about the impact of AI who did NOT sign the petition calling for a pause and is now advocating thoughtful and necessary consideration before AI gets too far out of the gate. Like the guy who gives loaded guns to the gang of psychotic school bullies and drives them to school before calling an out of State podcaster to discuss gun control.
Hopeless to solve problems? All we can do is sit in our basement? We'll solve energy in fantastic ways? Great way to dismiss all of the people who are working for true sustainability, where money and abundance are not the main drivers. His thinking is extremely narrow. 30 years ago Oren Lyons walked out of the UN climate conference with a phrase, Value Change For survival. These guys think tech can solve all of our problems when in reality it's our values and motivations. Sam doesn't want clean energy to solve global warming, he only wants energy to create an AI system that will put a personal assistant in all of our pockets. Could you imagine if all of these thinkers tried to tackle true sustainability and their motivation wasn't just wealth. Because they probably can't imagine that.
Gpt 4 api is too expensive.
The almost constant vocal fry is irritating
@nerobird3617
Ай бұрын
Sorry I think it gets better towards like 25 mins in
Moloch
He's waiting for his AI response in his ear-piece each question. It picks up the signal better with his mouth open.
Altman is a great addition to LGBT faculty 😊
I get a bad vibe from Sam Altman. Bad gut feeling. I think he is a liar. An obfuscater
She is such a bad interviewer.... I love Sam, but this is unwatchable due to the interviewer ...
@2:40 don't forget to turn off your FB notifications, nerd
@nerobird3617
Ай бұрын
oops!
@matthewhendricks648
Ай бұрын
@@nerobird3617 💚it happens.
people clapping for the man that is gonna ruin their future lul
This guy talks too much
@therainman7777
Ай бұрын
Talking is literally the only reason he’s there.
@jimmy_sandwiches3764
Ай бұрын
Check the title of the video
41:46