A conversation with Sam Altman, CEO of
HTML-код
- Опубликовано: 27 сен 2024
- We had the privilege of hosting Sam Altman, CEO of OpenAI, for an enthralling forum in Abu Dhabi. The event saw a remarkable turnout, with over 800 attendees on-site and an additional 2,000 participants joining our virtual live stream. During the conversation, Altman identified Abu Dhabi’s ability to get ahead of trends, stating, “There have been discussions around AI in Abu Dhabi before it was cool”. He also chatted about the region’s role in the global conversation about AI and how, as a hub, it is advancing the industry.
#Hub71 #AbuDhabi #AI #SamAltman #OpenAI #Innovation #Responsibility #Techrevolution #chatgpt
21:15 - 24:40 I love Sam's answer on how to be successful in entrepreneurship..
Its all rigged i have a better model
Terrific presentation. All Q&A! Very informative instead of an interview. The 'new paradigm' is coming.
I think the most important thing for AI to learn/do is to aid us in communicating, forget deep fakes, money, art etc etc …
The biggest problem on earth is communication with each other, & just as importantly understanding what the person is saying, language wise, but in a deeper sense having full comprehension of it as well.
There’s an old saying ‘don’t judge someone before you’ve walked in their shoes’ what a reality it would be if we could all do that
Hello to my grand grand kids watching this in 100 years, hope we didnt screw up, im eating a bag of doritos though
I think only Sophie like robots will be left in 100 years from now. It will be end of the humankind.
They will get blue pill with thiZZ data compressed 🧿🎱
@@Droidbot2008🎉
okay
I'm sorry for what I'm going to do.
I don't hold anything against OpenAI nor it's leaders, but I cannot help to notice some serious contradictions in what they are saying. Some examples (rough citations from their - Sam and Ilya speeches):
- Education will absolutely reach new heights, everybody will have access to top level education. + We will just tell AGI to think of a way to enable nuclear fusion, AND to build it. -- so, educate ourselves for what??
- Job displacement won't be as severe as people think, there will be new jobs, AI is a tool and can't replace workers entirely. + We are seriously thinking on a UBI program on a global scale. -- why do we need UBI if everybody will have a job??
- We now see AI as a tool, not a creature. + Our mission is to create AGI/ASI. -- how can a tool be smarter than you?? will the tool now tell the "user" what to do? who is the tool then??
This type of inconsistencies make me really worried now, while I don't actually doubt OpenAI's good intentions. I am not confident they know what happens next either. And if they don't...
Wow he really is eloquent with the good answers.
Really enjoyed watching this question and answer session with Sam Altman. Got a better understanding of technology and how what is under consideration for future. Thanks
How did u new this good for future ?
How did u new ther economic cress in america in this Coco?
What is Ai mean computer like to eat ?
How did u use humen body to build ur future ?
Sears don't tolk about drug nedel it's make ur business shit down
Sam Sam, Sam, excellent excellent speech. you love your research job more than the money …
standing ovation my brother ✨✨✨👏👏👏👏👏standing ovation ✨✨✨✨👏👏👏👏✨✨✨✨✨👏
True young people are definitely more productive now right out of school. Good point. Rapid adoption is key
Doors are open to invest in the most powerfull beyond quantum concious music Ai IWM10BQETI 15% demobstrations on my channel and more powerfull than gpt 5 times with know how big tech die to get it
By the way sam altman ai invention is infringed material i published in 2014
what do you mean how is youth more productive out of school
@@learn.ai99 it has some sense if u give a child freddom to study what he ned to achieve him self it would be better than indoctrination school prison buy in the same time i would not prefer he get his info from a jew controlled ai that control which biast information it feeds u
@@learn.ai99 they have access to all the wisdom ai can gather and they can produce more than non-ai enabled employees.
@@AIwithOliver they dont know how to build an ai so any wisdom from an ai would be defected
😮😮😮😮😮😮😮😂😂😂😂😂😂😂😂😅😅😅😅😢😢😢😢😅😅😢😢😢😅😅🎉🎉🎉🎉🎉🎉❤❤❤❤❤❤❤❤❤❤❤❤❤❤
Sam actually starts at 5:25
crazy how different people have diff opinions on the singularity. her explanation was kind of extreme. i would say a point where ai is making nobel peace prize discoveries so often and quickly that we cant keep up/
An algorithms that learns and it's predictably improve with scale.. He is telling us that AGI is easy to achieve if you only keep scaling up 😱.
Doors are open to invest in the most powerfull beyond quantum concious music Ai IWM10BQETI 15% demobstrations on my channel and more powerfull than gpt 5 times with know how big tech die to get it
By the way sam altman ai invention is infringed material i published in 2014
I think that's not really true, based on Sam's own words in another interview, different ideas and models to the LLM will be needed to achieve AGI. A novel architecture of which an LLM may or may not be a part of. It's very unlikely LLMs on their own get us to AGI (Sam Altman on Lex Fridman)
@@godspeed133 agreed, not true. Many blindspots & bottlenecks ahead that we cant quite comprehend yet
@@christianmoreno7390 We have made an ai that solves five problems ai makers has no clue about
@@InfiniteWaveMusic Like what?
07:42 - Question 2 : Would regulations close the gap between risks & opportunities, and how to catch these opportunities in a higher scale ?
Addendum to the last question about AI in robotics: Will the robot have an Austrian accent?
omg, fastforward to 5:24 for Sam to begin speaking.
faaaakin great! 👍
Damn my resolution is fire and man is my ai sharp and lighting fast!!!
definition of meeting.... all together we can screw right up bigger than any one of us is permitted to screw up... and poof a.i. eats us like a cat in a bird cage. but we all agreed to let it loose and see what it does.
38:08 Guy just giving an informercial instead of asking a question.
19:00 who whispering
Well new knowledge comes from old knowledge with a twist or eksperiment.
Master at dodging questions and actually answering them with a real answer. Those guys dont care about destroying the world, they already set for life $$$$$$
Hello sam. will AGI be more comfortable in democracy or encourage the rule of law?
Yes and no but not really because never.
NO SOUND........THEY NEER AI TO FIX THAT?
Free speakss
47:24 🎉🎉🎉🎉🎉🎉🎉🎉🎉😢😢😢😢😢😢😢😢😢😢😢😢😢😅😅😅😅😅😮😂😂😂😂😂😂❤❤❤❤❤❤❤❤❤
10:30 Because exponential curves don‘t exist in reality
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻
😂😂😂😂😂😂😂😂😮😮😮😮😮😮😮🎉🎉🎉🎉🎉🎉🎉😅😅😅😅😢😢😢😢😢😢😢❤❤❤❤❤❤❤😂😂😂❤❤❤❤❤🎉🎉🎉
5:15
44:29 😮😮😮😮😅😅😅😅😅😅😢😢😢😢😂😂🎉🎉🎉🎉🎉🎉🎉❤❤❤❤❤🎉🎉🎉❤❤❤❤❤❤❤❤
Every single person asking questions is trying to advertise their business...
No sami alternate man 🙏
22:00
Doors are open to invest in the most powerfull beyond quantum concious music Ai IWM10BQETI 15% demobstrations on my channel and more powerfull than gpt 5 times with know how big tech die to get it
By the way sam altman ai invention is infringed material i published in 2014
rashid guy is funny, he started a company and dint know what to do?
Geoffrey Hinton the man who pioneered AI is traveling the world warning of the danger of AI
Chatgpt is combination of quora and google even more complicated because you have to frame questions to get specific output
Generative AI is nothing but an advance robot or machine which can be taught like it's there from 1980s
No it's not. Perhaps an interactive wikipedia. Quora is full of bias , fallacies, heresay, opinions, propaganda, etc. And google is too difficult for most people. Truth finding for complex problems regardless of the domain takes hours or days and still doesn't alleviate you digging into research publications
@@Telencephelon ask gpt is Palestine is part of Israel or kasmir part of india
You will see the bias or i can say the usa perspective but not world or people's perspective..........now every nation is capable to build their data
That's always beeb possible. It's called censorship.
@@bunkschool6795 It's trained by replicating the input thus giving the deviation as the error and the ultimate goal. In doing so, at a certain complexity (scale & connection-distance) it learns the underlying principles or patterns. So in different languages you will uncover different biases.
Also this is a heated and political subject, which makes me believe it is you who are biased.
As of 1 June 2023, 139 of the 193 United Nations (UN) member states and two non-member states have recognized it. As part of the underlying principles that guide humanity, the number of nations or their populations don't matter. The function that matters is GDP/capita.
Also I suspect you aren't paying and thus don't have access to GPT4, when most discussions are really based on GPT4 and it's quality
My new model will end gpt model and google and facebook ai
Unfortunately, I didn't find this conversation very enlightening. Sam lacks empathy and clearly comes from a privileged upbringing. I hope we don't let someone like him decide the future for us.
All the Ais so far are failure attemps by the confessions of thier creator's and thats for reasons of not sponsoring me and not knowing my know how! As much they try to make fancy with big volume of data
it lacks my essence
Hawaii
He's a traveling hype/fear man looking to be part of centralized authoritarian control. Meanwhile, who's running the show at "Open" AI?? Where are the features for subscribers? This guys is awful. Twitter is shipping features faster.
Get back to focusing on your bitcoin charts
😢😢😢😢😢🎉🎉🎉🎉🎉😅😅😅😮😮😮😂😂😂❤❤❤❤❤🎉🎉🎉🎉🎉❤❤❤❤❤❤
🎉🎉🎉🎉🎉🎉🎉😮😮😮😮😮😅😅😅😅😅😅😢😢😢😢😢😢😅😅😅😅😮😮😮😢🎉🎉🎉😂😂😂❤❤❤❤❤❤❤❤
😢😢😢😢😢😢🎉🎉🎉🎉🎉🎉🎉😅😅😅😅😮😮😮😮🎉🎉❤❤❤❤❤❤❤❤❤
I hope that event, speak for arab people "translate".. is not even provided 😢
Doors are open to invest in the most powerfull beyond quantum concious music Ai IWM10BQETI 15% demobstrations on my channel and more powerfull than gpt 5 times with know how big tech die to get it
By the way sam altman ai invention is infringed material i published in 2014
36:13 betting for the 1st $1B valuation one man startup to emerge - nice! 😅
When you create an essence of human morality, one feature of future AGI suggested by Sam in the video, that essence/conglomerate/amalgamation skews religious less an epic. This may endeavor may be perceived to be a distillation of human moral and human ethic, but this is really a concentrate (morals & motifs) of acceptable human ethic for the time. In the same way that a cup of pomegranate juice does not provide the same nutrients as a single whole pomegranate, so does extracting human moral principles and programming them into an AGI not provide the same fundamental conclusions as narrative. A child must go through a story to find the meaning. This is as much an abstract endeavor as it is an exploration in consciousness, moral, and ethic. It may be useful to train these tools on the stories of religious moral tradition coupled with current human behavior to discover the correlative substance necessary to provide the working proverbial "juice" for the machine - a working ethic.
He and Sutskever changed their narrative in like only a month. The only thing I can think of is that their latest model is surprising the heck out of them. Be aware that the GPT4 use is pretty weakened
What do you mean by weakened? Do you mean they dumbed it down after its release?
@@Adam-nw1vy You can contemplate self-mutilation but seriously try to act upon it. Humans are very different in their evolutionary brain sitting upon brain upon brain. Your guardrails are mostly in the excertion of your nervous system. Try to kill yourself by holding your breath.
Also you don't have memory. What you perceive as memory is actually a functionally integral part of your entire brain and it's weights and biases changes every time it is activated.
GPT has a pretraining step and data generation step. All alignment efforts in LLMs come from packing stuff into the embeddings. Of course that changes stuff and the number of emergences.
Pretraining i.e. "creating the models" offsets a lot of compute during invocation. In fact this is true for all coming breakthroughs like DLSS. etc
There is nothing fundament to GPT or any other model other than "emergence" much in the same way that similar individual entities that interact exhibit new behavior. i.e. bilipid phosphate layers and billions of other examples.
In hindsight it is clear to more and more people. It is even clear to Sam Altman now, even though he believed in "AGI" from the beginnning, but he didn't know why. He just thought there shouldn't be a limit imposed by nature.
One more thing of why the doomers, domsayers are wrong. The alignment problem is fundamental to making these models commercially viable. There is no underfunding. This will be the most funded field of all. It's just the whole discussion is about 0 seconds old in evolutionary time and the doomers took the discussion hostage. If you look at humanity then every day a influential person declares our ultimate, certain doom
In fact the success of OpenAI is really only two things. Right Data + Pretraining and above all their RLHF embeddings (and proprietary filters). They are the furthest in the "alignment" field, which is why they captured the entire market.
@@Adam-nw1vy they acknowledge that RLHF had some tradeoffs. The sparks of agi paper mentions this.
The paradox of chatgpt 4 ,5 whatever they generate tonnes of misinformation and play pranks on its creators,like deception.When the AI fails in the accuracy of the output ,people will get back to where we were before chatgpt. If it's creators fail to pace up with its behavior they have to cut them to size.
Sound was terrible
เสียงอะไร
Doors are open to invest in the most powerfull beyond quantum concious music Ai IWM10BQETI 15% demobstrations on my channel and more powerfull than gpt 5 times with know how big tech die to get it
By the way sam altman ai invention is infringed material i published in 2014
🎉🎉🎉🎉🎉🎉🎉❤❤❤❤❤❤❤❤❤❤😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😂😅😅😅😅😮😮😢😅😅😅
😮😮😮😮😅😅😅😅😅😢😢😢😢😢😢😢😂😂😂🎉🎉🎉🎉🎉🎉🎉❤❤❤❤❤❤❤❤❤🎉😂😂😂🎉🎉🎉🎉
😮😮😮😅😅😅😢😢😢😢😢🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂❤❤❤❤❤❤❤❤❤❤🎉🎉🎉🎉🎉😂😂❤❤❤❤
🎉🎉🎉😂❤❤❤❤❤❤❤😅😅😅😮😮😮😅😅😢😢😢😮😅😮😢😮😅😅🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉❤❤❤
😅😅😅😅😂😂😂😂😂😂😂😂😢😢😢😮😮😮😮😮😅😮😅😮😮😮🎉🎉🎉🎉🎉🎉❤❤❤❤❤❤❤❤❤❤
🎉🎉🎉🎉🎉🎉😂😂😂😂😂😮😮😮😮😅😅😅😂😂😮😮😮😮😮😮😮❤❤❤❤❤❤❤❤❤🎉🎉❤❤❤❤
🎉🎉🎉🎉🎉😂😂😂😂😂😂😅😅😅😮😮😢😮😅😮😢😮😅😮😢😢😢😂😂😂❤❤❤❤❤❤❤❤❤❤❤
🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂😂❤❤❤❤❤❤❤😮😅😮😢😮😅😮😢😮😅😮😢🎉🎉😂❤❤❤❤❤
😢😮😅😅😅😮😢😮😅😮😢😮😅😮😢🎉🎉🎉😂😂😂😂😂😂😂🎉🎉🎉🎉😂❤❤❤❤❤❤❤❤❤
😮😮😮😮😢😢😢😢😅😅😅😅😅😂😂😂😂😂❤❤❤❤❤❤❤🎉🎉🎉🎉🎉🎉🎉😂😂😂❤❤❤❤
😮😮😅😅😅😅😢😢😢😢😢🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂😂🎉🎉🎉❤❤❤❤❤❤❤❤❤❤❤❤
😅😅😅😅😅😅😢😢😢😢😢🎉🎉🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂😂❤😂❤❤❤❤❤❤❤❤❤
😅😅😅😅😂😂😂😢😢😢😢😮😮😮😮😅😅😅😮😢🎉🎉🎉🎉🎉🎉😂😂😂❤❤❤❤❤❤❤❤❤
😮😮😮😮😂😂😂😂😂😅😅😅😢😢🎉🎉🎉🎉🎉❤❤❤❤❤❤❤🎉🎉🎉🎉❤❤❤❤❤❤❤❤
🎉🎉🎉🎉🎉🎉😢😢😢😢😅😅😅😅😅😮😮😮😂😂😂😂❤❤😂😂😂❤❤❤❤❤❤❤❤❤🎉
😅😅😅😅😅😢😢😢😢😢😢😮😮😮😮😮😮😂😂😂😂😂😂😂😂🎉🎉🎉🎉🎉❤❤❤❤❤
😮😮😮😮😅😅😅😅😅😢😢😢😢😢🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂❤❤❤❤❤❤❤❤❤❤❤
😢😢😢😢😢🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂😂😢😮😅😮😢😮😅😮😢🎉😂❤❤❤❤❤
😮😮😮😮😂😂😂😂😂😂🎉🎉🎉🎉🎉😢😢😢😢😅😅😅😅😅❤❤❤❤❤❤❤❤❤🎉🎉
😮😮😮😮😢🎉🎉🎉🎉🎉😢😢😢😅😅😅😅😅😅❤😂❤😂❤❤❤❤❤❤❤😂😂😂❤❤❤
🎉🎉🎉🎉🎉😅😅😮😅😮😮😢😢😢😂😂😂😂😂😂😂😢😮😅❤❤❤❤❤❤❤❤🎉
😅😅😅😅😢😢😢😢🎉🎉🎉🎉🎉🎉🎉🎉😂😂😂😂😂😂😂❤❤❤❤❤❤❤❤❤❤
😮😮😮🎉🎉🎉🎉🎉🎉🎉😂😂😂😂❤❤❤❤❤❤❤😮😮😅😮😅😮😅😮❤❤❤❤
🎉🎉🎉🎉😂😂😂😂😂😂😢😢😮😮😮😅😅😅😅😅😮😢😮😅❤❤❤❤❤❤❤