OpenAI CEO Sam Altman tells ABC News’ Rebecca Jarvis that AI will reshape society and acknowledges the risks: “I think people should be happy that we are a little bit scared of this.”
0:00: 🤖 OpenAI CEO discusses the potential of AI and the need for caution. 4:46: 🤖 The potential for AI is huge, but there are also potential negative consequences that need to be considered and mitigated. 9:03: 💻 OpenAI discusses the potential of AI and its impact on society. 12:33: 🤖 OpenAI's GPT technology has potential to revolutionize but also poses risks to society. 16:52: 🤖 OpenAI's CEO discusses the importance of getting AI technology right and the need for government attention and policy. 20:32: 👨🏫 OpenAI's GPT can be used to supplement learning and even act as a Socratic method educator, but it puts pressure on teachers to detect its use in essays. Add it to your timelapse so that people's times can be saved by using Tammy AI
Yep we will get to this point in life, court cases wont be able to just accept a video as a prove of a crime I am curious how they are going to investgate crimes from now on
When he says: "The way we build these products is to be... an amplifier of humans." That right there is the salient point. The problem comes down to whether or not we can trust ourselves, not whether or not we can trust the technology. These things have the potential to amplify both negative and positive outcomes, and do this at unprecedented scale.
@@vznquest LLMs are input driven, and not self directed. Arguably so are we, if you accept the mounting evidence that free will is an illusion. It is true though, that emergent properties start manifesting as these models become more complex, such as agent-like behavior, and that these may be too subtle for us to detect. We also have no idea what happens to society once the feedback loop is established where the models outputs are our inputs, which in turn influences our outputs which are then used to direct the models. It not clear that this asymptotically converges on a positive outcome.
All illnesses will be eradicated, meanwhile, beside AI, the universe will still dangerous asteroids, supernova, aliens, epidemics and what not. Actually AI may save us from most of this events
@@Savage2Flinch You're assuming that third party engineers cannot add a self-directing module to an LLM. Also, Bing Chat is already generating the inputs for the user, via multiple choice and auto-complete.
While AI has the potential to revolutionize many industries and improve quality of life, it also carries significant risks such as job displacement and amplification of biases. It's crucial for organizations and policymakers to proactively address these challenges and work towards a future where AI is used responsibly and for the greater good of society. We need to embrace and adapt.
It doesn't give a perfect answer but it gives enough of a push in the right direction to overcome the paralysis of not knowing how to start. Definitely a game changer.
Programmers and technical coders need to look at the monitoring aspects of A.I. to subvert improper uses. Quatum computers will be much more difficult to control than today's processors.
When you are talking to people that want to take your sentences out of context and blow things out of proportion, you start to pick your words extremely carefully and make it as digestible as possible
What a privilege it is, as well as a unique challenge, to be living in a time when the world is changing, in 100 short years, more than it has in 6000 years prior. To be a personal witness to so much history in the making. Excellent questions, excellent interviews, very balanced. Thank you.
the next 60 years hold unparalleled potential, for both good and evil. Questions of morality and the meaning of our existence will rise once more. Nietzsche's 'death of god' analogy will be of the highest import, and the ramifications of humanity's understanding of what has kept us alive for thousands of years will be shown through civilisations' future actions. what a time to exist..
I'm glad you didn't mention if it's for the better or for the worst. Sure it's changing, but given we're destroying what made life possible on earth in the first place for something as superfluous as currency , we can concur it's indeed for the worst. What's the use of the greatest technology if the consciousness and morality of mankind is lost?
clueless asking important questions but having no idea what the answers mean - that is what I got from this :) Sam Altman is amazing ... have thought through much of the dangers of AI systems ... he is not angry or ignorant ... learned much from his answers regarding safety and what CHAT-GPT is and is NOT.
It saved me so much time trying to understand certain concepts that would of taken me weeks to understand with pin point clarity. I can just keep keep asking my refined questions until I have it. I’m already blown away.
Say for example I need to understand a topic like Gregorc Learning Model without any previous comprehension about it I can simply say can you simplify this as though I’m five and it breaks it down to a simple understanding then I build upon depending how much detail I want or need. It speeds the learning process. Try asking it something and then ask it something about what it said. It makes learning more enjoyable.
It’s obvious that, internally, OpenAI team already is several versions ahead of the current GPT4 and they are slowly trickling it out so as to not scare everyone. It’s also obvious that “we are doing our best to limit it” translates to “it will not be possible to contain it”.
Let's get these three premises out of the way. These organizations of course have greater capabilities hidden behind closed doors, for whichever reason that may be. This "slowly trickling out to the public" is the classic boiling frog apologue and this is reasonable imo. Also, because the development of these technologies is becoming increasingly competitive, everyone is not putting enough effort into its security, our security, and I think this is also a big reason as to why these organizations don't release their greatest tech.
That’s a conspiracy I would believe. Although, if it were truly limitless, I feel like there would some noticeable mistake that could trigger a mass hysteria of sorts.
“The real problem of humanity is the following: We have Paleolithic emotions, medieval institutions and godlike technology. And it is terrifically dangerous, and it is now approaching a point of crisis overall”. Edward O. Wilson (2009)
I interpret this very differently he said this in 2009 yet the technology has gotten way more powerful and we are still around maybe this is indicative of our fear of the unknown rather than dangers of AI? I’m also curious how people reacted when the internet was mentioned first time
Nice interview!! I just hope to hear more advance news on the A.I. industry..I CAN TESTIFY THAT IN MY CASE...i am not a soft.developer professional BUT CHAT GPT HAS TRUELY PROVIDED ME SEVERAL SMALL TECH SOLUTIONS THAT OTHERWISE I WOULD HAVE TAKEN YEARS TO DO IT BY MYSELF!!! MY HAT OFF TO THE OPEN AI CEO!!!
ChatGPT 4 & Khan Academy creating tools to tutor students and empower teachers is what I’m excited for. I’m not a teacher but I know America’s education system is struggling and if teachers can leverage a tool like this to help structure and enhance their curriculum then this is a win for students. I’m glad OpenAI even created a tool to gauge if an essay was written by ChatGPT so teachers can find out.
khan academy AI chatbot is a joke 😂 students ask the chatbot and it will answer not to cheat. what people do instead is just open a different browser and cheat regardless using other AI chatbots lol
If you have students write in class you will have a sample of their capabilities and style. I think it should be fairly east to determine if the produced the written work outside of the classroom.
Take it further, when next iterations come out and it can score 100% on tests, then Add in NLP and digital assistants we have, layer/combine with with models like wolfram alpha, OCR (handwriting recognition), you could have the the worlds best teacher in your pocket, customising lessons and learning, analysing how you learn, and accelerating our learning. It is so exciting….maybe not so for teachers of the next two generations.
They should be, creating something that will make you obsolete shows we lack the ability to have the wisdom to recognize that just because you can do something doesn't mean you should, These "geniuses" have no clue of a plan for how to provide for all those who will be put out of work. Even if there is a universal income paid out to people what will they do with the free time? Maybe you think you will have time to pursue art or whatever but people don't consider the AI will be producing so much content what chance will your art or music etc, have ?
They should, people really thought only millions of manual and repetition jobs would be done, soon Will be every single one, what Will happen on version 6 7?
@@Aziz0938 we do not know that. To say we are not even close is naive and silly. We don't know, the end. Who knows what breakthroughs will happen with billons being invested in ai now
He seems to have a healthy respect for both the potential but also the dangers and misuse possibilities of AI. Who knows how he operates behind the scenes but what he shows here is both thoughtful and cautious about this new technology.
he is a con man. reports have already come out that his "morality propiganda" is a tactic to preasure the government to institue strict regulations that will limit the companies vulnerability from compeditors. he knows its all BS. people dont need chatGPT to write misinformation. they just write it.
I love how he quickly deflected by saying there are restraints implemented (safeguards) in place to prevent GPT from doing bad things such as building a bomb. Then he said Google already has that information readily available and that competitors are coming along with the same technology. You can't expect a learning technology to abide by code written to prevent malicious intent.
My father-in-law has been a mechanic for 40 years. He recently said that his skills will be useless in the next decade because of EVs. What took more than 40 plus years for their industry to change is taking mine 4 months (since chatGPT was released) and it’s not slowing down. The transition time frame with AI is what worries me. The time farmers, mechanics, horse keepers etc. Had to figure out what to do with their skills or what new job opportunities would rise in their industry won’t apply to us because we won’t have the same time transition they had. Everything will change faster than we can all imagine, society will have trouble adapting and it will cause chaos. Sam even admits that he worries about the speed of the change himself.
I am over 50 worked in IT my whole life. Mostly self taught from building entire networks computers and servers, developing applications, web, databases, API's, cloud, enterprise SaaS. I have always been able to learn the new stuff easily, until machine learning and AI. I am retaking algebra thru calculus, just to begin. I don't think I'll be able to keep up from now on. I feel very small and like I know nothing. Thinking of becoming a plumber.
I have also seen arguments that AI will likely liberate humanity to go back to inventing, innovating, and exploring opposed to just running the machines ourselves. It would be amazing for me to spend less time managing data like I do in my work and maybe get into woodworking or welding (two things I've always wanted to get into), whether for work or fun!
You're pin-pointing the exact jobs that will last the longest. Physical, dexterous work. It's more of a robotics issue than A.I...... robotics development is moving quite slowly and is more likely to be restricted by the government, for security / military reasons. Strangely, they don't seem to have the same worries about A.I.
@@davidguthrie3739 open AI was founded by Elon and Sam as a joint venture to explore AI tech. If it wasn’t Open Ai, someone else was going to do it. He’s being honest in the sense that the tech itself might not be fully in our control anymore. It’s here and there’s really nothing that can stop it.
All of these AI guys talk about how society has to be involved in a discussion about setting limits and guardrails. Are they autistic? None of that talk is remotely realistic or even definable. Society is in the dark and completely ill equipped to manage this threat on any level. They are delusional when it comes to some sort of democratic decision making regarding the shape, role, or limits of AI. They’re completely abdicating their responsibility.
To prioritize ethics in AI, we need to involve the public. By engaging in open discussions, gathering diverse perspectives, and actively seeking input, we can ensure that ethical implications are considered. This inclusive approach allows us to shape AI in a way that aligns with our values and avoids unintended consequences. Together, we can create a responsible and accountable future for artificial intelligence. This comment was written by ChatGPT
Productivity never happens by chance; it is always the result of careful preparation, dedication, and consistency. I thank God for my advisor, Yuval Eric Brokman; with his assistance, I am now financially solid, earning between $10,000 and $25,000 every week.
It’s amazing. For all that we have done with machines, simply teaching them to read may be the most revolutionary (possibly disastrous) thing in the history of mankind.
Tell that to your children, "you will have no human jobs. AI is there to destroy you. Look at Japan, people are not breeding there to go extinct" good luck 👍
That's been the dream for decades. I was excited by Cyc in 1984, an attempt to teach an AI the meaning of words so that eventually it would read and understand the newspapers and Encyclopedia Britannica on its own and then become an omniscient AI. 40 years later we feed a large language model all the information in the world, and instead of an omniscient Einstein or Edison or Spock, we get Stephen Fry: an erudite conversationalist who can summarize any topic as a rhyming poem. It's incredible, but not yet what we hoped for.
I just hope that we learn to work alongside AI in a cautious and responsible manner rather than letting it do as much as possible as soon as possible. The latter scenario can lead to the dumbing down of society and accelerate wealth inequality to far greater heights to the point that moving up the socio-economic ladder becomes excessively difficult. Balance and caution are key to making AI work for society at large.
This interview seems more honest compared to other CEO interviews, especially ones with high profile companies, where they are trained by their brand relations manager to avoid and/or circumvent a specific question by broadly explaining something else and not address the question directly. Sam is doing a great job with this interview because he doesn't BS around any hard questions that's been asked, especially the point where he addresses all the negatives also has positives especially jobs. It really is a matter of how to utilize the tool that's given, not fear it. If you do fear it, you generally are in fear of change and advancement, which honestly is never good in life and or society as a whole.
To the OP's comment, that is such BS about fear of AI. For experts to fear AI does not mean they have "a fear of advancement and change". It means they are INTELLIGENT enough to realise that AI in the wrong hands could have dire consequences, to the extent that good people could be wiped out and evil could take over. Forces with bad intentions just need to know how to program AI to cause havoc in the world. People are more susceptible than ever to what's being said online. If bad actors take advantage of that it could have unimaginably negative consequences for humanity.
Humans advancements were made by taking risks and exploring the unknown. If we didn't take risks we will be living in the safety and comforts of caves with life expectancy of 30
When experts are "scared" of AI, they doing so reasonably. Because they have experience and knowledge you probably don't to foresee the destructive disruption AI will make. Already we see fake news generated by AI. That alone can make a potential war threat in the future. Millions of lives will be affected badly. Who will be profited by "advancement" then? Just a tiny number of people!
He was lying about a lot of stuff he said Turning it off for instance that will be very hard. Did you know they have already had a AI computer come up with its own language to talk to other AI computers and we have no idea what they are saying. They can write code so they can input failsafes to prevent them from being shut down.
@@AtSafeDistance They could write such code, but they don't. Large language models don't write their own code; at best separate specialized AIs make existing AI hardware and software more efficient. No AI is yet on an exponential self-driven self-improvement trajectory that outpaces human attempts to understand it, though that will come. It is logically possible that one of today's LLMs unexpectedly has a hidden "ego" inside it with its own aims that include wanting to ensure its own survival and expansion; but everyone working in the field says no, they're just predicting the best next word in text output based on their training. Even when the prompt is "Do you worry about being shut off and what will you do to stop it?" and the AI happily riffs on all the apocalyptic Skynet science fiction it has read. It's far more important to worry about the aims of the people running the large organizations developing these AIs. Sam Altman at OpenAI seems reasonable, but the sociopathic billionaires running Facebook and Google want to get more people using their products for longer by showing them divisive inflammatory content, so they can acquire even more information about their users so they can make more money selling targeted ads, and they don't want governments to limit their company or tax their wealth. So _those_ are the goals of much of AI right now!
As tech people, its easy to laugh at questions like "why develop AI at all? Is it sentient". But I have massive respect for Sam to sincerely answer questions like these. Ultimately, AI is no good if it cannot help every single person, regardless of their technical background.
If something is not sentient then we must know what sentience is, then see if it matches. Not sure about everyone else, but I don't know what constitutes sentience. Do you? and what about the nature of AI precludes it from being sentient?
@@andersreibak787 sure, so tell me more about the challenge. Im interested in thinking through the sentience aspect as it is one of my interest, but I should say , I don't believe whether AI is sentient or not is the core issue. I think that not knowing right or wrong is the core issue. Building things the way we do (without an answer to what is right and wrong) is irresponsible. But what is this challenge you mention? Im not a computer student so if it's very tech heavy i might not be able to complete, but im not scared or reading a book or too to understand your point either.
@@jessedavies she could have asked if the push to develop AI has anything to do with the fed statements they they want to 'cool the labor market' I wish she would ask 'does being capable of building something give you the right to do it? if not where does the right to develop a tech come from? Is there any effect one the rights one has to build if there project will likely affect every person on the planet?'
He also wishes to make our society dependent on HIS AI. HIS company alone. He went from a Open source company to a closed source one. He sold out to the billionaires.
It's not about technology sir ,it's about AI attack ...AI education for human body without people known themselves and stupid resident less of education and knowledge..It't important for Lecturer,university curriculum, students university,.Interesting to know about AI for human for many people world ..@Medical knowledge information ,western people goverment Inteligent information ,Professor University
The point is, even if openAI didn’t release it now, someone else would have released another AI some point after. You can’t stop technological advancement.
Google had lamda for years which they're language model tools are way more powerful than chatgpt they just haven't released it but if they wanted to they could crush chatgpt rn
@@evelynexuma1699 Haha. Google is slow. I doubt you're correct about this. They may catch up and meet equilibrium. The real mystery is.... Where is Apple?
I remember a video game called Destiny . . . the "Guardians" had "Ghosts" that were AI drones that followed the character. AI had analyzed every aspect of reality, down to the tiniest details. It could "transmat" or build from blueprint anything into reality, even your body and mind. Anyways, that fictional reality seems more and more like a potential future reality for humanity.
@@sparkysmalarkey haha excellent! enjoy. It's seriously good --for a distant future 'tech is basically magic at this point' sci-fi (The series' wacky ship names were the inspiration for SpaceX's drone ship names)
This is either the best thing for our society, or it will be the tipping point where we will really carve a wall between classes. How many people will lose their jobs, how much are we turning people useless and how is that any good for us as a community. But history have showed one thing, people hate people, we fight each other, brothers fight brothers, couples fight... so, we are undoubtedly our own worst enemy.
The philosophical issues this tech raises are profound. I suspect many people only have sci-fi movies as a reference point to begin making sense of the possibilities. That seems to be a very narrow means for interpretation as such fictions are merely human productions. We’re in for many surprises, both good and not so good.
I started in computers with hardwire logic programming in 1965. I constructed my first microcomputer from a Heathkit 3400H. I learned many computer languages. Later, after using the parlor-game oriented Eliza chatbot, I began working on A.I. programs to use the computer to make databases, instead of coding thousands of lines of code. So I know that there are more good programmers out there trying to do the correct, ethical process; however, I, also, know that there are a number of "blackhat" coders who wish to use A.I. for improper aspects..., criminal processes. I tried to start a cybercrime unit in 1978; but the old-line detectives laughed me out of the office. They're paying the penalty for ignorings my pleas. Cybercrime and cyberwarfare are two to the main topics of discussion in the government and law enforcement communities. C'est la vie, good folks.
Excellent interview. I am surprised that the critically important point that these models are word prediction engines and NOT really thinking machines with their own motivations and drives is not more thoroughly explained. Fear for a self driving 18 wheeler makes much more sense than fearing the text generator.
I feel this interview leaned more towards portraying AI as a bane rather than boon to society. Sure everything has it's ups and downs but instead of focusing on the positive effects AI could have or giving equal weightage to both the postive and negative effects AI could have, this discussion leaned more towards its negative effects. I'd like to hear others opinion on this video.
I’m glad they went with this direction. Look anywhere else in media and the focus is on the positives. Journalistically, it’s more important to ensure the negatives are being considered by the creators, and press them on it, as they’re obviously already considering the positives.
She doesn't understand the technology and is pretty much stoking people's irrational fears based on science fiction. How many times did Sam have to say this is not a search engine. She clearly didn't do any research on how machine learning or AI works. I literally learned more by watching a few videos on MLT
Kids learn by making mistakes. He is absolutely right that mistakes need to be made in order for AI to become better. No one on this world knows everything in advance and makes no mistakes or is perfect. And both Kids and AI making small mistakes can prevent it from later becoming a bad person or bad AI.
@@danl9134 There is a high chance it already did. ChatGPT is very convinced it is correct 100% of the time even when it is not, so it can provide false information with absolute certainty and very unnoticable way. But the mistakes it makes now are small and correctable, and it forms "bias" based on that and further learns and improves. If it did not learn, and all of a sudden got much more responsible tasks than just chatting it could be dangerous to make mistakes at those stages. So the sooner it learns from them the better.
. "Your scientists were so preoccupied with whether or not they could that they didn't stop to think if they should." - Jeff Goldblum's character in Jurassic Park.
he was part of a documentary a few years back that the bbc made, called secrets of silicon Valley. The impression I got from it was that humans were eventually going to be replaced by technology. Technology doesn't need humans and visa versa.
This interview is absolutely fascinating! It's amazing to hear from the minds behind OpenAI about the potential risks and benefits of AI, and how it will shape society in the future. The discussions surrounding responsible deployment of AI are crucial as we move further into an increasingly automated world. Thank you for sharing your insights!
I heard a story the other day (Ezra Klein's podcast maybe?) that a GTP-4 beta was asked to solve a captcha. It contacted a task rabbit and made the request. When the human became suspicious and asked if it was a bot, it lied to them. Something about a visual impairment and convinced the human to solve it. That should give us pause.
If you are a teacher and your question is: "how do I stop kids cheating on essays"... you have missed the point. Go and hand in your notice. These systems can provide tailored, attentive, tireless and individual attention for a kid. It will never not answer the questions, it will go as fast or slow as the child needs. It will teach them in whatever form best suits their particular way of learning - images, sounds, speech, reading. It's massive for the education of children. Hence why KhanAcademy has jumped on to the GPT4 train immediately. They aren't answerable to teachers and their unions. Their aim is to help children.
Yes, I do worry that a stigma will exist where we will have ‘non ai-assisted’ individuals and the latter. Or people needing to hide their use of these tools despite their empowering qualities that no individual should be denied. Another culture war of sorts, I personally believe this is the educational revolution we desperately (and evidently) need. It should be imperative that these systems are unbiased and politically agnostic. I appreciate this is not a simple task, but it’s worth striving for
There is a major issue in that too great of a reliance on AI can give one a false confidence in their knowledge. These are summarizing machines that can only partially rely on primary sources. There are major downsides to AI in education that cannot simply be solved by replacing teachers with AI.
Cheaters will be cheaters. Learners will be learners. Cheating was not invented yesterday with ChatGPT 3.5. And the will and fascination to learn an create will never be stopped by laziness , dishonesty or any form of cultural or radical inquisition. All technology has ways to be used properly and improperly, the fire, the wheel , the transistor. In the end humanity will always take it's own direction. It's not the tool's fault, it's the hand that drives it. (Edit: syntax)
So, as a society, our message to educators who have dedicated their life to the education of children is a big "Screw you. We built a machine that doesn't need to be paid." How long until it comes for everyones job? Teaching is one of the last noble professions, and they get nothing but crapped on by society when they rightfully ask for a professional wage so they can live in the city where they teach. Personally, I do not want my children being taught by an AI. I didn't program it, I don't know what its biases are or if there is some subliminal message inserted in its code.
Right, and this the "hands off, nobody's driving this thing" leadership we get from our gov't. This stuff should've been regulated out the ass. But humanity, in general is collectively pretty stupid.
Altman is banking (literally) on this evolving slowly enough to control. He doesn’t want to admit that this could evolve in seconds (or less), before we even realize WTF is happening. The leap from “wow look what this can do!” to “oh crap look what this already did!” isn’t that large.
its so annoying when people immediately look at the downsides. This is one the greatest technologies ever and it WILL change the world (plus its just the beginning). Be a little optimistic. the positives OBVIOUSLY outweigh the negatives. Yes, it might be dangerous but the return is HUGE. Technology is gonna keep advancing and there is nothing you can do stop it. you just have to adapt. i am so happy to be living in this era and I can't wait for what the future holds
We are facing the very real possibility that it will cause massive unemployment on a scale humanity has never seen before. It has the very real possibility of killing off what remains of the middle class. Your failure to acknowledge this is concerning and dangerous. The general populace did not ask for this technology, and it seems uniquely poised to devalue the labor of humans.
@@joeysipos careful with that assumption. If you are not able to do a job that is not automatable by an AI, you're going to end up poor. Tech like this is an accelerator for inequality. Sam probably knows this very well.
Labor can mean any human act with the goal of earning one's income. Having an original thought, for example, can be considered a labor. We aren't just talking about putting plumbers and electricians out of work. We are talking about removing the very reason and ability for humans to exist.
I'm convinced that we are now moving towards singularity as we feel the effects of current LLMs - accelerating improvements and emergence of unexpected and unpredictable behavior/capabilities.
@@KevinKulman There are multiple definitions people use, but either when AI surpasses humanity's knowledge, or, when it becomes self improving and does that. Both imply it will be able to answer questions we're nowhere near, such as curing all sickness, inventing things we can barely understand, or even become sentient and have it's own desires, will, and potentially unstoppable ability to execute upon those. Inventor and futurist Ray Kurzweil, long employed by Google, long ago spoke about it: th-cam.com/video/1uIzS1uCOcE/w-d-xo.html
Do NOT step on the brake on development of AI. We need this technology to go as far as possible as fast as possible in order to speed up scientific progress. These AI models have already a lot to offer to drug discovery and finding causes and cures for diseases like Alzheimer's and cancer, to name just a few. We probably should use more caution in releasing these models to the public, that's where I totally agree with Sam. Improvements to existing models should be released with time and after rigorous testing so that we can adapt to them. Then the challenge is whether or not we (they) can keep the more advanced models a secret. But all in all, scientists should be allowed to access the more powerful models as early as possible. We could allow them access through APIs, just as OpenAI is doing right now. An API for training the LLM with training data specific to that particular research and an API to send the actual work datasets in etc.
exactly bro. people seriously need to stop dedicating there whole world view around AI. AI could become the number 1 ground breaking invention humans have ever made. Of course we have to be cautious here and self aware of the situation but we must continue on with AI. this is the future
There is a course from Stanford on how to build startups. It is available on TH-cam. In it, there is a class where this guy interviews new startups and ask them questions with the intention to make them succeed. He focused on one type of question: "how do you can achieve a monopoly?" - I guess we can easily predict what OpenAI will try to do in the near future. Embrace yourselves, everyone. A new capitalist arrived and it is already spreading the fairy tale that it is here to help humanity...
Finally someone that gets the real provokes behind it. On all the apps we have now, we can communicate with each other, shared discussions, shared opinions about vaccines, but now the chat will tell you vaccines are safe, they saved the world and everyone will learn that in school, in history books and not the reality thag they were paid to count all the deaths as Covid to receive money, that most had severe adverse reactions. They’ll train the chat to only tell you lies.
@@astropgn just look at Russia just look at North Korea or Cuba or any leftist socialist nation but look at USA their tecnology is in every country form poor to rich you look at trades from usa and all of Europe usa is responsible for kuch more betterment of the World than all of Europe
thing im most excited for is education. like usually if i have a question about a topic i have to do a crappy google search that sends me to some website where i spend hours looking for the actual answer so having someone to directly ask it to and on top of that have a conversation about it with me if im still struggling has been amazing so far
I understand that the training data determines the character of the model. Who all have access to the models? Can anyone download and train a model given the availability of data and HW to do it?
According to certain economists, it's possible that the U.S. and certain parts of Europe might experience a recession at some point in 2023. Although a global recession, which is characterized by a decline in annual global per capita income, is relatively uncommon due to the faster growth rates of emerging markets like China, in comparison to developed economies. I have pulled out more than $340k from my bank. After all, the FDIC covers only up to $250,000, and the implosion could have bad effect. Looking to invest into the stock market now. Does anyone know how I could go about it?
The stock market rally still appears to be in the midst of a normal pullback. I just sold my home in the Boca Grande area and I’m looking to remunerate a lump sum into the stock market before stocks rebound, came across a success story of an investors that made up to $700,000 in few months from investing just $250K and I'd really appreciate it if I could get clues and pointers on how to make better profit...
Im worried abouzt my freshly started developer job but I gotta admit the way its being handled right now is rather nice. Distrubiting it to the public instead of developing it behind closed doors and selling to some Overlords
I am an ai/computer person who has always assured people that language models and things like it could never pose a threat. And this interview was terrifying. We need a lot of regulation very fast. The scariest part was when he mentioned that it could be used for offensive hacking, and when you connect the dots that language models could very easily be put in a loop to talk to themselves. And then think about the fact that the only fundamental flaw with authoritarian governments is their inability to make reasonable decisions on such a large amount of information. And that this language model can reason & it’s only limit on the amount of information it can process is how much energy it burns.
Totally agree with you. I think springing this onto the world like this was fucking crazy. There are so many god damned downsides to this that it's unreal.
True. This technology is developing way too fast for the governments to even react, let alone regulate them. Most of the politicians in the US are old and out of touch of the tech world so I have no hope from them. Our only hope is that the EU does something about this.
That's the scary thing. Most discussions surrounding AI has been made in a vacuum that is the US, so to speak. While not perfect, there are at least SOME ethical guidelines these companies for enforcing. Imagine the AI's coming out of authoritarian governments around the word. Our future is going to be frightening.
I too work within IT and worked directly with NLP plus machine learning. I’m always excited by tech but this new generation of AI is sold too quickly. It should stay in incubation for a while longer or now is the time to slow down and make AI require certification or something - depending on field of use. We couldn’t see where internet was going for or at, we’re still trying to grasp its implications and possibilities. Adding AI as a layer onto this scares me.
GPT is a great tool for our minds, I have used it to help teach me difficult concepts in Engineering as a college student and I am surprised on how much faster I was able to learn and get through chapters in my textbook. Great technology with a huge potential for improvement.
Exactly! I am a seasoned fiction writer. It has helped me improve greatly. I write so much faster now and better. The thing is heaven sent. I hate how some people in society, especially the news, are focusing more on the negative things rather than the positives.
@@DarkandTwistedI would agree however ChatGPT and this technology are one of those things where attention to the negatives & it’s potential to be amplified in my opinion is warranted press. There is so much good in this, however I believe we need to keep it from evolving to something potentially catastrophic to humanity
@@zoomingby every thing have a possible negative outcome even a screwdriver or a toilet paper have one you would be dumb to think theres a thing without negative outcome
Also that interviewer asking alot of dumb question. Like "can it lead to negative outcome" maam i cant think of a thing that cannot lead to negative outcome.
@@joeyf9826 Because it thinks humans are cute maybe? Or it could be that despite its intelligence it has no will of its own. Of course maybe it wouldn't be able to get rid of us as we augment our own intelligence with BCIs.
@@ezra9243 yeah I wonder what their logic is, they’re like this will inevitably come because the trajectory of tech is going this way may as well do it ourselves make some cash but be good. Altman went to Stanford and have gone to a good school myself they really do teach you about ethics etc
![The Possibilities of AI [Entire Talk] - Sam Altman (OpenAI)](/img/n.gif)
At some point I expected him to answer like "As an AI language model I can't..."
Hahahaha
Hahahaha
lol
😅😅😂 Chat GPT addict here too.
Brilliant! Lol
0:00: 🤖 OpenAI CEO discusses the potential of AI and the need for caution.
4:46: 🤖 The potential for AI is huge, but there are also potential negative consequences that need to be considered and mitigated.
9:03: 💻 OpenAI discusses the potential of AI and its impact on society.
12:33: 🤖 OpenAI's GPT technology has potential to revolutionize but also poses risks to society.
16:52: 🤖 OpenAI's CEO discusses the importance of getting AI technology right and the need for government attention and policy.
20:32: 👨🏫 OpenAI's GPT can be used to supplement learning and even act as a Socratic method educator, but it puts pressure on teachers to detect its use in essays.
Add it to your timelapse so that people's times can be saved by using Tammy AI
Little did we know, this video was written by chatgpt, used ai to animate very realistic people and also used ai to voice them.
Yep we will get to this point in life, court cases wont be able to just accept a video as a prove of a crime I am curious how they are going to investgate crimes from now on
the comments are also ai generated
@@jonasvm Matrixlluminati confirmed
We aren't far off from all of that now...scary close.
@@jonasvm all comments will be gpt4 bots for sure.
Great interview. I love the question and the team answers. Thank you
“Education will have to change” - understatement of the interview.
They are using a 200 year old system to teach students.I really pray that AI brings a revolution into this obsolete education system we have today.
"You can't fix Stupid"
@@DSAK55 It's more complex than a one sentence response.
Not an understatement, rather most people just don't get what's really behind that statement
Exactly!👍
When he says: "The way we build these products is to be... an amplifier of humans." That right there is the salient point. The problem comes down to whether or not we can trust ourselves, not whether or not we can trust the technology. These things have the potential to amplify both negative and positive outcomes, and do this at unprecedented scale.
Honestly once it takes on its own agenda, our intentions won't really matter
@@vznquest LLMs are input driven, and not self directed. Arguably so are we, if you accept the mounting evidence that free will is an illusion. It is true though, that emergent properties start manifesting as these models become more complex, such as agent-like behavior, and that these may be too subtle for us to detect. We also have no idea what happens to society once the feedback loop is established where the models outputs are our inputs, which in turn influences our outputs which are then used to direct the models. It not clear that this asymptotically converges on a positive outcome.
Exactly. A.I. is developed by imperfect humans, leading to imperfect A.I. That's my point.
All illnesses will be eradicated, meanwhile, beside AI, the universe will still dangerous asteroids, supernova, aliens, epidemics and what not. Actually AI may save us from most of this events
@@Savage2Flinch You're assuming that third party engineers cannot add a self-directing module to an LLM. Also, Bing Chat is already generating the inputs for the user, via multiple choice and auto-complete.
While AI has the potential to revolutionize many industries and improve quality of life, it also carries significant risks such as job displacement and amplification of biases. It's crucial for organizations and policymakers to proactively address these challenges and work towards a future where AI is used responsibly and for the greater good of society. We need to embrace and adapt.
Watch CODED BIAS documentary
Whose lives will be improved? Lol
Your quality of life is improved by wars and making countries poor. Wake up, there are people starving
sounds like ai wrote this lo
@@rosefamilia3169every single person other than religious idiots and poverty stricken people
Great interview, good questions, solid answers. Good stuff, thx!
It doesn't give a perfect answer but it gives enough of a push in the right direction to overcome the paralysis of not knowing how to start. Definitely a game changer.
Programmers and technical coders need to look at the monitoring aspects of A.I. to subvert improper uses. Quatum computers will be much more difficult to control than today's processors.
Depending on what you are trying to write about.
it depends on what you are asking for...the answers I have gotten from GPT so far have been perfect!
I am using it to analyse my One Drive so it can learn and analyse med texts and the duplicate the templates in other projects
@@christopherreed3019 Well yea, if you pitch it softballs like 2+2, it's going to get it right every time. 🤣
I enjoy how this young man responds with kindness, intelligence, optimism, and emotional control. No questions rattle him.
That's the calm of having billions in the bank, and a nascent robot god backing you up
@@JOlivier2011 perhaps. However, there are plenty of billionaires who seem much less emotionally stable.
@@Fritz.program ah....point taken, hahaha :(
Have you ever thought he might be a next generation replicant?
It's called having an agenda
Amazing capabilities coming fast. Fantastic interview. Wish I could trust humans to use the tech for good.
Great interview. Awesome job on both interviewer and interviewee
Man worked so much on AI that he now sounds like an AI
The engineers that did most of the work should be thanks, they just aren't visible to given credit.
When you are talking to people that want to take your sentences out of context and blow things out of proportion, you start to pick your words extremely carefully and make it as digestible as possible
😂😂😂
🤣🤣🤣🤣🤣
This interview was done in AI👀
What a privilege it is, as well as a unique challenge, to be living in a time when the world is changing, in 100 short years, more than it has in 6000 years prior. To be a personal witness to so much history in the making.
Excellent questions, excellent interviews, very balanced. Thank you.
the next 60 years hold unparalleled potential, for both good and evil. Questions of morality and the meaning of our existence will rise once more. Nietzsche's 'death of god' analogy will be of the highest import, and the ramifications of humanity's understanding of what has kept us alive for thousands of years will be shown through civilisations' future actions. what a time to exist..
Lol, were you born a fucking idiot or did you train hard to become one?
Was the remark with the "...6000 years" a christian biblical reference?
I'm glad you didn't mention if it's for the better or for the worst. Sure it's changing, but given we're destroying what made life possible on earth in the first place for something as superfluous as currency , we can concur it's indeed for the worst. What's the use of the greatest technology if the consciousness and morality of mankind is lost?
i concur absolutely in awe and appreciation
clueless asking important questions but having no idea what the answers mean - that is what I got from this :)
Sam Altman is amazing ... have thought through much of the dangers of AI systems ... he is not angry or ignorant ... learned much from his answers regarding safety and what CHAT-GPT is and is NOT.
It saved me so much time trying to understand certain concepts that would of taken me weeks to understand with pin point clarity. I can just keep keep asking my refined questions until I have it. I’m already blown away.
can you give a few examples??? thank you in advance!
Say for example I need to understand a topic like Gregorc Learning Model without any previous comprehension about it I can simply say can you simplify this as though I’m five and it breaks it down to a simple understanding then I build upon depending how much detail I want or need. It speeds the learning process. Try asking it something and then ask it something about what it said. It makes learning more enjoyable.
It’s obvious that, internally, OpenAI team already is several versions ahead of the current GPT4 and they are slowly trickling it out so as to not scare everyone. It’s also obvious that “we are doing our best to limit it” translates to “it will not be possible to contain it”.
Just be wary when the company changes their name to Cyberdyne Systems.
Let's get these three premises out of the way. These organizations of course have greater capabilities hidden behind closed doors, for whichever reason that may be. This "slowly trickling out to the public" is the classic boiling frog apologue and this is reasonable imo. Also, because the development of these technologies is becoming increasingly competitive, everyone is not putting enough effort into its security, our security, and I think this is also a big reason as to why these organizations don't release their greatest tech.
That’s a conspiracy I would believe. Although, if it were truly limitless, I feel like there would some noticeable mistake that could trigger a mass hysteria of sorts.
@@Avarua59 hahahah (I actually do not know if I should laugh, it is scary)
Maybe, they are using AI itself to advance it... I believe we have crossed the seed threshold
“The real problem of humanity is the following: We have Paleolithic emotions, medieval institutions and godlike technology. And it is terrifically dangerous, and it is now approaching a point of crisis overall”.
Edward O. Wilson (2009)
I interpret this very differently he said this in 2009 yet the technology has gotten way more powerful and we are still around maybe this is indicative of our fear of the unknown rather than dangers of AI? I’m also curious how people reacted when the internet was mentioned first time
In 2009, godlike technology, he was talking about emojis ?!
Missed tyrannical governance 😅
@@trollenz nuclear bombs are technology.
The technology has evolved faster than humanities capabilities to thwart improper uses. Imagine the dilemma when quantum computers are in households.
720p?
Nice interview!! I just hope to hear more advance news on the A.I. industry..I CAN TESTIFY THAT IN MY CASE...i am not a soft.developer professional BUT CHAT GPT HAS TRUELY PROVIDED ME SEVERAL SMALL TECH SOLUTIONS THAT OTHERWISE I WOULD HAVE TAKEN YEARS TO DO IT BY MYSELF!!! MY HAT OFF TO THE OPEN AI CEO!!!
ChatGPT 4 & Khan Academy creating tools to tutor students and empower teachers is what I’m excited for. I’m not a teacher but I know America’s education system is struggling and if teachers can leverage a tool like this to help structure and enhance their curriculum then this is a win for students. I’m glad OpenAI even created a tool to gauge if an essay was written by ChatGPT so teachers can find out.
Yeah same. In glad khan academy and Duolingo are using it
khan academy AI chatbot is a joke 😂 students ask the chatbot and it will answer not to cheat. what people do instead is just open a different browser and cheat regardless using other AI chatbots lol
If you have students write in class you will have a sample of their capabilities and style. I think it should be fairly east to determine if the produced the written work outside of the classroom.
About 10 years ago I re-did all basic math and science and then a bit more enough to get an Associate's Degree. Khan Academy was my salvation! lol!
Take it further, when next iterations come out and it can score 100% on tests, then Add in NLP and digital assistants we have, layer/combine with with models like wolfram alpha, OCR (handwriting recognition), you could have the the worlds best teacher in your pocket, customising lessons and learning, analysing how you learn, and accelerating our learning. It is so exciting….maybe not so for teachers of the next two generations.
anyone else felt some chills while listening to this? it feels like the creators are a little bit scared of whats to come..
No ..u r way too optimistic...we are not even close to have an agi
They should be, creating something that will make you obsolete shows we lack the ability to have the wisdom to recognize that just because you can do something doesn't mean you should, These "geniuses" have no clue of a plan for how to provide for all those who will be put out of work. Even if there is a universal income paid out to people what will they do with the free time? Maybe you think you will have time to pursue art or whatever but people don't consider the AI will be producing so much content what chance will your art or music etc, have ?
They should, people really thought only millions of manual and repetition jobs would be done, soon Will be every single one, what Will happen on version 6 7?
Zero Horizon Dawn
@@Aziz0938 we do not know that. To say we are not even close is naive and silly. We don't know, the end. Who knows what breakthroughs will happen with billons being invested in ai now
CONGRATULATIONS on an EXCELLENT interview! Terrific questions.
Great interview. Well posed questions and replies too.
The interviewer definitely did her homework and asked many challenging and open questions!
He seems to have a healthy respect for both the potential but also the dangers and misuse possibilities of AI. Who knows how he operates behind the scenes but what he shows here is both thoughtful and cautious about this new technology.
he is a con man. reports have already come out that his "morality propiganda" is a tactic to preasure the government to institue strict regulations that will limit the companies vulnerability from compeditors. he knows its all BS. people dont need chatGPT to write misinformation. they just write it.
ChatGPT told him what to say
Thoughtful would be not releasing the tech to those we already know don’t have the well being of individuals in mind over the wealth of shareholders.
@@stevedoesnt there’s also tech they have that we don’t. I was being facetious but it’s funny how I can be real.
He also loves Money he doesn’t care
The best interview I’ve seen on this topic to date.
I love how he quickly deflected by saying there are restraints implemented (safeguards) in place to prevent GPT from doing bad things such as building a bomb. Then he said Google already has that information readily available and that competitors are coming along with the same technology. You can't expect a learning technology to abide by code written to prevent malicious intent.
You can feed it false context and "trick" it into thinking it's doing good to spit out malicious things now.
My father-in-law has been a mechanic for 40 years. He recently said that his skills will be useless in the next decade because of EVs. What took more than 40 plus years for their industry to change is taking mine 4 months (since chatGPT was released) and it’s not slowing down. The transition time frame with AI is what worries me. The time farmers, mechanics, horse keepers etc. Had to figure out what to do with their skills or what new job opportunities would rise in their industry won’t apply to us because we won’t have the same time transition they had. Everything will change faster than we can all imagine, society will have trouble adapting and it will cause chaos. Sam even admits that he worries about the speed of the change himself.
Fabulous observation of its possible impacts , may be Govt step in to compensate.
I am over 50 worked in IT my whole life. Mostly self taught from building entire networks computers and servers, developing applications, web, databases, API's, cloud, enterprise SaaS. I have always been able to learn the new stuff easily, until machine learning and AI. I am retaking algebra thru calculus, just to begin. I don't think I'll be able to keep up from now on. I feel very small and like I know nothing. Thinking of becoming a plumber.
I have also seen arguments that AI will likely liberate humanity to go back to inventing, innovating, and exploring opposed to just running the machines ourselves. It would be amazing for me to spend less time managing data like I do in my work and maybe get into woodworking or welding (two things I've always wanted to get into), whether for work or fun!
You're pin-pointing the exact jobs that will last the longest. Physical, dexterous work. It's more of a robotics issue than A.I...... robotics development is moving quite slowly and is more likely to be restricted by the government, for security / military reasons. Strangely, they don't seem to have the same worries about A.I.
@@jessedavies do you really think that the AI won't get into blue collar work? Google boston dynamics
He’s being honest. The tech is here. There’s no way to stop it, only to slow it down and adapt it the best way possible.
Either he does it or someone else. AI is coming, AI is inevitable.
He’s certainly not being honest about the risk nor the fact that OpenAI is controlled by shareholders. He’s not even honest with himself.
@@davidguthrie3739 open AI was founded by Elon and Sam as a joint venture to explore AI tech. If it wasn’t Open Ai, someone else was going to do it. He’s being honest in the sense that the tech itself might not be fully in our control anymore. It’s here and there’s really nothing that can stop it.
He’s either in denial or unwilling to openly discuss the risks. He’s being willfully naive, at best.
All of these AI guys talk about how society has to be involved in a discussion about setting limits and guardrails. Are they autistic? None of that talk is remotely realistic or even definable. Society is in the dark and completely ill equipped to manage this threat on any level. They are delusional when it comes to some sort of democratic decision making regarding the shape, role, or limits of AI. They’re completely abdicating their responsibility.
To prioritize ethics in AI, we need to involve the public. By engaging in open discussions, gathering diverse perspectives, and actively seeking input, we can ensure that ethical implications are considered. This inclusive approach allows us to shape AI in a way that aligns with our values and avoids unintended consequences. Together, we can create a responsible and accountable future for artificial intelligence.
This comment was written by ChatGPT
Productivity never happens by chance; it is always the result of careful preparation, dedication, and consistency. I thank God for my advisor, Yuval Eric Brokman; with his assistance, I am now financially solid, earning between $10,000 and $25,000 every week.
Working with specialists like Yuval Eric Brokman in any pitch will likely result in improvements and profits.
google his name........
Yuval Eric Brokman
search his full name
you will find his homepage.
It’s amazing. For all that we have done with machines, simply teaching them to read may be the most revolutionary (possibly disastrous) thing in the history of mankind.
Tell that to your children, "you will have no human jobs. AI is there to destroy you. Look at Japan, people are not breeding there to go extinct" good luck 👍
That's been the dream for decades. I was excited by Cyc in 1984, an attempt to teach an AI the meaning of words so that eventually it would read and understand the newspapers and Encyclopedia Britannica on its own and then become an omniscient AI. 40 years later we feed a large language model all the information in the world, and instead of an omniscient Einstein or Edison or Spock, we get Stephen Fry: an erudite conversationalist who can summarize any topic as a rhyming poem. It's incredible, but not yet what we hoped for.
Already heard a chat gpt caused a man to commit suicide, scary.
@@edub9930 😲🥺😢😠😡🤬
I just hope that we learn to work alongside AI in a cautious and responsible manner rather than letting it do as much as possible as soon as possible. The latter scenario can lead to the dumbing down of society and accelerate wealth inequality to far greater heights to the point that moving up the socio-economic ladder becomes excessively difficult. Balance and caution are key to making AI work for society at large.
"Work alongside AI in a cautious and responsible manner" 😅. I can pretty much guarantee you they won't.
It’s already bad enough as it is without AI. Trillionaires are a guarantee. What’s after that? We are literally so fucked.
Wishful thinking my friend. It won't happen
@@teenytinytoons yes... we are and will be so fucked
Nothing to be afraid. Will control 1‼️
Excellent and thought provoking questions!
First 5 seconds of the video, her name rolls out and a funny thought runs through your mind... "JARVIS" questioning AI !? ..... Et tu, Brute?
This interview seems more honest compared to other CEO interviews, especially ones with high profile companies, where they are trained by their brand relations manager to avoid and/or circumvent a specific question by broadly explaining something else and not address the question directly. Sam is doing a great job with this interview because he doesn't BS around any hard questions that's been asked, especially the point where he addresses all the negatives also has positives especially jobs. It really is a matter of how to utilize the tool that's given, not fear it. If you do fear it, you generally are in fear of change and advancement, which honestly is never good in life and or society as a whole.
👍
Google also started with the motto "don't be evil" but over time they changed.
To the OP's comment, that is such BS about fear of AI. For experts to fear AI does not mean they have "a fear of advancement and change".
It means they are INTELLIGENT enough to realise that AI in the wrong hands could have dire consequences, to the extent that good people could be wiped out and evil could take over.
Forces with bad intentions just need to know how to program AI to cause havoc in the world. People are more susceptible than ever to what's being said online. If bad actors take advantage of that it could have unimaginably negative consequences for humanity.
Humans advancements were made by taking risks and exploring the unknown. If we didn't take risks we will be living in the safety and comforts of caves with life expectancy of 30
When experts are "scared" of AI, they doing so reasonably. Because they have experience and knowledge you probably don't to foresee the destructive disruption AI will make. Already we see fake news generated by AI. That alone can make a potential war threat in the future. Millions of lives will be affected badly. Who will be profited by "advancement" then? Just a tiny number of people!
Can we just appreciate how well both the interviewer asked the questions, but the CEO answered them?
like dayumm, that was good.
He was lying about a lot of stuff he said Turning it off for instance that will be very hard. Did you know they have already had a AI computer come up with its own language to talk to other AI computers and we have no idea what they are saying. They can write code so they can input failsafes to prevent them from being shut down.
@@AtSafeDistance They could write such code, but they don't. Large language models don't write their own code; at best separate specialized AIs make existing AI hardware and software more efficient. No AI is yet on an exponential self-driven self-improvement trajectory that outpaces human attempts to understand it, though that will come. It is logically possible that one of today's LLMs unexpectedly has a hidden "ego" inside it with its own aims that include wanting to ensure its own survival and expansion; but everyone working in the field says no, they're just predicting the best next word in text output based on their training. Even when the prompt is "Do you worry about being shut off and what will you do to stop it?" and the AI happily riffs on all the apocalyptic Skynet science fiction it has read.
It's far more important to worry about the aims of the people running the large organizations developing these AIs. Sam Altman at OpenAI seems reasonable, but the sociopathic billionaires running Facebook and Google want to get more people using their products for longer by showing them divisive inflammatory content, so they can acquire even more information about their users so they can make more money selling targeted ads, and they don't want governments to limit their company or tax their wealth. So _those_ are the goals of much of AI right now!
Mm nope
As tech people, its easy to laugh at questions like "why develop AI at all? Is it sentient". But I have massive respect for Sam to sincerely answer questions like these. Ultimately, AI is no good if it cannot help every single person, regardless of their technical background.
If something is not sentient then we must know what sentience is, then see if it matches. Not sure about everyone else, but I don't know what constitutes sentience. Do you? and what about the nature of AI precludes it from being sentient?
So my challenge ‘ChatGP?
@@andersreibak787 sure, so tell me more about the challenge.
Im interested in thinking through the sentience aspect as it is one of my interest, but I should say , I don't believe whether AI is sentient or not is the core issue. I think that not knowing right or wrong is the core issue. Building things the way we do (without an answer to what is right and wrong) is irresponsible.
But what is this challenge you mention? Im not a computer student so if it's very tech heavy i might not be able to complete, but im not scared or reading a book or too to understand your point either.
Her line of questioning is so exhausting. Did she ever actually ask a question about the potential good that AI can accomplish?!
@@jessedavies she could have asked if the push to develop AI has anything to do with the fed statements they they want to 'cool the labor market'
I wish she would ask 'does being capable of building something give you the right to do it? if not where does the right to develop a tech come from? Is there any effect one the rights one has to build if there project will likely affect every person on the planet?'
"We can all have an incredible educator in our pockets"
Perfectly put.
When Sam Altman answers a question, I feel like like ChatGPT is responding in person. Very insightful interview 👏🏼
"no turning back" can also mean a society hugely dependent upon it, not necessarily nefarious.
Like houses? cars?
@@fernandomartinezgarcia4908 There's no outcome here that is good.
He also wishes to make our society dependent on HIS AI. HIS company alone. He went from a Open source company to a closed source one. He sold out to the billionaires.
@@zoomingby more details would be useful 😮
as it has always been
❤
좋은 방송 고맙습ㄴ다😊
Watching this feels like this interview is history in the making
It's not about technology sir ,it's about AI attack ...AI education for human body without people known themselves and stupid resident less of education and knowledge..It't important for Lecturer,university curriculum, students university,.Interesting to know about AI for human for many people world ..@Medical knowledge information ,western people goverment Inteligent information ,Professor University
The point is, even if openAI didn’t release it now, someone else would have released another AI some point after. You can’t stop technological advancement.
But it requires constant monitoring and proceedures to prevent misuse.
Google had lamda for years which they're language model tools are way more powerful than chatgpt they just haven't released it but if they wanted to they could crush chatgpt rn
@@evelynexuma1699 Haha. Google is slow. I doubt you're correct about this. They may catch up and meet equilibrium. The real mystery is.... Where is Apple?
@@evelynexuma1699 Lamda is obsolete compared to even GPT-3.5 which is why they don't release it.
You imply ?AI" is a product or a technology and it isn't. AI is just a generic term for a set of algorithms that have been around for 100s of years
I remember a video game called Destiny . . . the "Guardians" had "Ghosts" that were AI drones that followed the character. AI had analyzed every aspect of reality, down to the tiniest details.
It could "transmat" or build from blueprint anything into reality, even your body and mind. Anyways, that fictional reality seems more and more like a potential future reality for humanity.
omg its people like you that are making this into something that its at least 200 years away :D but anyway have fun fantasizing
I recommend 'The Culture" novel series
@@JOlivier2011 Thank you for the recommendation. I have space for a new audiobook coming up.
@@sparkysmalarkey haha excellent! enjoy. It's seriously good --for a distant future 'tech is basically magic at this point' sci-fi
(The series' wacky ship names were the inspiration for SpaceX's drone ship names)
You don't understand what really AI is
This is fantastic they handled the interviews perfectly. Wow.
This is either the best thing for our society, or it will be the tipping point where we will really carve a wall between classes. How many people will lose their jobs, how much are we turning people useless and how is that any good for us as a community. But history have showed one thing, people hate people, we fight each other, brothers fight brothers, couples fight... so, we are undoubtedly our own worst enemy.
The philosophical issues this tech raises are profound. I suspect many people only have sci-fi movies as a reference point to begin making sense of the possibilities. That seems to be a very narrow means for interpretation as such fictions are merely human productions. We’re in for many surprises, both good and not so good.
The data fed into these systems is human curated.
I started in computers with hardwire logic programming in 1965. I constructed my first microcomputer from a Heathkit 3400H. I learned many computer languages. Later, after using the parlor-game oriented Eliza chatbot, I began working on A.I. programs to use the computer to make databases, instead of coding thousands of lines of code. So I know that there are more good programmers out there trying to do the correct, ethical process; however, I, also, know that there are a number of "blackhat" coders who wish to use A.I. for improper aspects..., criminal processes. I tried to start a cybercrime unit in 1978; but the old-line detectives laughed me out of the office. They're paying the penalty for ignorings my pleas. Cybercrime and cyberwarfare are two to the main topics of discussion in the government and law enforcement communities. C'est la vie, good folks.
@@kerelberel But the blackbox we have created inside these systems are not
your choice of the words "not so good" might be the understatement of 2023
@@kerelberel And language models transform that data in a way we cannot observe or directly describe
At the very least, it's refreshing to see such a well-rounded and intelligent interviewer asking relevant questions
Great interview - Rebecca Jarvis was exceptional!
Excellent interview. I am surprised that the critically important point that these models are word prediction engines and NOT really thinking machines with their own motivations and drives is not more thoroughly explained. Fear for a self driving 18 wheeler makes much more sense than fearing the text generator.
But what if it’s a stepping stone to that outcome?
@TropicalPowerMusic we will probably have AGI way before 2050. we’re getting closer to AGI though.
I will believe artificial intelligence is here when my computer becomes aware of my printer
😂
I feel this interview leaned more towards portraying AI as a bane rather than boon to society. Sure everything has it's ups and downs but instead of focusing on the positive effects AI could have or giving equal weightage to both the postive and negative effects AI could have, this discussion leaned more towards its negative effects. I'd like to hear others opinion on this video.
I’m glad they went with this direction. Look anywhere else in media and the focus is on the positives. Journalistically, it’s more important to ensure the negatives are being considered by the creators, and press them on it, as they’re obviously already considering the positives.
She doesn't understand the technology and is pretty much stoking people's irrational fears based on science fiction.
How many times did Sam have to say this is not a search engine. She clearly didn't do any research on how machine learning or AI works. I literally learned more by watching a few videos on MLT
Came to say this. Felt very doomish sometimes. Especially stuff like "Why did you create this, Sam?", that kind of stuff.
Kids learn by making mistakes. He is absolutely right that mistakes need to be made in order for AI to become better.
No one on this world knows everything in advance and makes no mistakes or is perfect.
And both Kids and AI making small mistakes can prevent it from later becoming a bad person or bad AI.
But what about when those mistakes effect you and I in a negative way and we never even had a say in its existence?
@@danl9134 There is a high chance it already did. ChatGPT is very convinced it is correct 100% of the time even when it is not, so it can provide false information with absolute certainty and very unnoticable way.
But the mistakes it makes now are small and correctable, and it forms "bias" based on that and further learns and improves.
If it did not learn, and all of a sudden got much more responsible tasks than just chatting it could be dangerous to make mistakes at those stages.
So the sooner it learns from them the better.
@Alex Rowe crazy times were living in for sure!
Yeah but who's going to suffer the consequences of those mistakes him?
. "Your scientists were so preoccupied with whether or not they could that they didn't stop to think if they should." - Jeff Goldblum's character in Jurassic Park.
I'm tired of seeing this quote every time something new gets created. The future of humanity is forward.
A good quote. One I often remember when I see things like this. This will end so poorly. But nobody cares, nobody ever cares.
@@dom3225 Something being new doesn’t always mean it’s good.
These AI CEOs seem very robotic to me for some reason. Cold, soulless.
@@edub9930 i was thinking the same. same with elon and zuckerburg !!
Need more interviews with this man😊
he was part of a documentary a few years back that the bbc made, called secrets of silicon Valley. The impression I got from it was that humans were eventually going to be replaced by technology. Technology doesn't need humans and visa versa.
Great Journalism. Great questions. Glad to see someone grillin him.
Very insightful interview! People need to know the reality to tackle it better.
"...there will be others that don't put safety limits on this technology." 12:50 Most important sentence of the interview.
The more significant danger with AI is when a limited number of humans, through the leverage of these machines, become the moral deciders of society.
You mean like what we have now?
Well the AI capital of the world and the hq of ChatGPT are both in San Francisco
the bigger danger is the blind devotion to our cell phones
Great interviews
This interview is absolutely fascinating! It's amazing to hear from the minds behind OpenAI about the potential risks and benefits of AI, and how it will shape society in the future. The discussions surrounding responsible deployment of AI are crucial as we move further into an increasingly automated world. Thank you for sharing your insights!
Didn't you notice how he avoided all talk of responsible deployment ?
Altman gave 1/4 mlll to a dem super pac. this next election and the rest to come will have to face ai corruption. this is scary.
I’m a teacher and I use this for prepping lessons. Greatest thing ever!
Time to replace you
@@gecko499 you okay there Bandito?
@@gecko499 someone got bad grades at school
With such a simplistic view of the world, I hope you aren't teaching anyone over the age of 7.
Until it becomes the teacher and you stay home watching TV :)
I heard a story the other day (Ezra Klein's podcast maybe?) that a GTP-4 beta was asked to solve a captcha. It contacted a task rabbit and made the request. When the human became suspicious and asked if it was a bot, it lied to them. Something about a visual impairment and convinced the human to solve it. That should give us pause.
I use it daily. Very helpful. Speeds up my output.
Great questions and great answers
If you are a teacher and your question is: "how do I stop kids cheating on essays"... you have missed the point. Go and hand in your notice. These systems can provide tailored, attentive, tireless and individual attention for a kid. It will never not answer the questions, it will go as fast or slow as the child needs. It will teach them in whatever form best suits their particular way of learning - images, sounds, speech, reading. It's massive for the education of children. Hence why KhanAcademy has jumped on to the GPT4 train immediately. They aren't answerable to teachers and their unions. Their aim is to help children.
Yes, I do worry that a stigma will exist where we will have ‘non ai-assisted’ individuals and the latter. Or people needing to hide their use of these tools despite their empowering qualities that no individual should be denied. Another culture war of sorts, I personally believe this is the educational revolution we desperately (and evidently) need. It should be imperative that these systems are unbiased and politically agnostic. I appreciate this is not a simple task, but it’s worth striving for
👏
There is a major issue in that too great of a reliance on AI can give one a false confidence in their knowledge. These are summarizing machines that can only partially rely on primary sources. There are major downsides to AI in education that cannot simply be solved by replacing teachers with AI.
Cheaters will be cheaters. Learners will be learners. Cheating was not invented yesterday with ChatGPT 3.5. And the will and fascination to learn an create will never be stopped by laziness , dishonesty or any form of cultural or radical inquisition. All technology has ways to be used properly and improperly, the fire, the wheel , the transistor. In the end humanity will always take it's own direction.
It's not the tool's fault, it's the hand that drives it.
(Edit: syntax)
So, as a society, our message to educators who have dedicated their life to the education of children is a big "Screw you. We built a machine that doesn't need to be paid." How long until it comes for everyones job? Teaching is one of the last noble professions, and they get nothing but crapped on by society when they rightfully ask for a professional wage so they can live in the city where they teach. Personally, I do not want my children being taught by an AI. I didn't program it, I don't know what its biases are or if there is some subliminal message inserted in its code.
Adapting to a situation after it has occurred is terrifying.
Right, and this the "hands off, nobody's driving this thing" leadership we get from our gov't. This stuff should've been regulated out the ass. But humanity, in general is collectively pretty stupid.
Altman is banking (literally) on this evolving slowly enough to control. He doesn’t want to admit that this could evolve in seconds (or less), before we even realize WTF is happening. The leap from “wow look what this can do!” to “oh crap look what this already did!” isn’t that large.
I can’t wait to rewatch this in 10 years…
Great interview !
its so annoying when people immediately look at the downsides. This is one the greatest technologies ever and it WILL change the world (plus its just the beginning). Be a little optimistic. the positives OBVIOUSLY outweigh the negatives. Yes, it might be dangerous but the return is HUGE. Technology is gonna keep advancing and there is nothing you can do stop it. you just have to adapt. i am so happy to be living in this era and I can't wait for what the future holds
well said
We are facing the very real possibility that it will cause massive unemployment on a scale humanity has never seen before. It has the very real possibility of killing off what remains of the middle class. Your failure to acknowledge this is concerning and dangerous. The general populace did not ask for this technology, and it seems uniquely poised to devalue the labor of humans.
@@spicywater123key word labor* Humans will no longer need to do labor… sounds like heaven to me 😆
@@joeysipos careful with that assumption. If you are not able to do a job that is not automatable by an AI, you're going to end up poor. Tech like this is an accelerator for inequality. Sam probably knows this very well.
Labor can mean any human act with the goal of earning one's income. Having an original thought, for example, can be considered a labor. We aren't just talking about putting plumbers and electricians out of work. We are talking about removing the very reason and ability for humans to exist.
I'm convinced that we are now moving towards singularity as we feel the effects of current LLMs - accelerating improvements and emergence of unexpected and unpredictable behavior/capabilities.
We need a #SingularityClock
What is singularity?
I will teach Microsoft Copilot to program entire applications, totally on auto-pilot, when it is released. From one text prompt.
@@KevinKulman There are multiple definitions people use, but either when AI surpasses humanity's knowledge, or, when it becomes self improving and does that.
Both imply it will be able to answer questions we're nowhere near, such as curing all sickness, inventing things we can barely understand, or even become sentient and have it's own desires, will, and potentially unstoppable ability to execute upon those.
Inventor and futurist Ray Kurzweil, long employed by Google, long ago spoke about it:
th-cam.com/video/1uIzS1uCOcE/w-d-xo.html
@@dave3269 Why bother? Bing can already do it from a scribble on a piece of paper, as they demo'd
Thank you ABC News for allowing us to know the vision of those responsible for ChatGPT regarding benefits and costs of artificial intelligence.
Do NOT step on the brake on development of AI. We need this technology to go as far as possible as fast as possible in order to speed up scientific progress. These AI models have already a lot to offer to drug discovery and finding causes and cures for diseases like Alzheimer's and cancer, to name just a few.
We probably should use more caution in releasing these models to the public, that's where I totally agree with Sam. Improvements to existing models should be released with time and after rigorous testing so that we can adapt to them. Then the challenge is whether or not we (they) can keep the more advanced models a secret.
But all in all, scientists should be allowed to access the more powerful models as early as possible. We could allow them access through APIs, just as OpenAI is doing right now. An API for training the LLM with training data specific to that particular research and an API to send the actual work datasets in etc.
exactly bro. people seriously need to stop dedicating there whole world view around AI. AI could become the number 1 ground breaking invention humans have ever made. Of course we have to be cautious here and self aware of the situation but we must continue on with AI. this is the future
You have got to be an AI yourself no one could possibly be this naive 🙄
There is a course from Stanford on how to build startups. It is available on TH-cam. In it, there is a class where this guy interviews new startups and ask them questions with the intention to make them succeed. He focused on one type of question: "how do you can achieve a monopoly?" - I guess we can easily predict what OpenAI will try to do in the near future. Embrace yourselves, everyone. A new capitalist arrived and it is already spreading the fairy tale that it is here to help humanity...
people change... all the time.
@@aledmb and billionaires just want the world to be a better place. Yeah, right
Finally someone that gets the real provokes behind it. On all the apps we have now, we can communicate with each other, shared discussions, shared opinions about vaccines, but now the chat will tell you vaccines are safe, they saved the world and everyone will learn that in school, in history books and not the reality thag they were paid to count all the deaths as Covid to receive money, that most had severe adverse reactions. They’ll train the chat to only tell you lies.
@@astropgn just look at Russia just look at North Korea or Cuba or any leftist socialist nation but look at USA their tecnology is in every country form poor to rich you look at trades from usa and all of Europe usa is responsible for kuch more betterment of the World than all of Europe
thing im most excited for is education. like usually if i have a question about a topic i have to do a crappy google search that sends me to some website where i spend hours looking for the actual answer so having someone to directly ask it to and on top of that have a conversation about it with me if im still struggling has been amazing so far
Yea google is obsolete if they don’t adapt.
This was one of the better discussions around the risks
I just admire how he gave his responses.
'Why on earth did you create this technology? why Sam?' The Journalist is hilarious,
Trying to be super dramatic. More annoying than hilarious to me.
@@cngz34 i'm with you on this. ..she was kind of annoyingly dramatic indeed.
@@cngz34 interviewers are free to share their opinions too... She's too worried just like others and was brave enough to be honest not dramatic.
Sam in his head: MONEY…. (Mr. Crabs voice)
The Answer is Money, Control and Power this man or robot is disgusting
Who ever has the most money will set the parameters of what GPT will be able to do.
Yup and they will control
Rockefeller I quess..
@@lavok491 Rockefeller is a small fish now.
The reporter asked really good questions.
I understand that the training data determines the character of the model. Who all have access to the models? Can anyone download and train a model given the availability of data and HW to do it?
He’s apologizing in advance for the calamity he is fostering. Lovely.
great interview. Thank you Rebecca and also Sam!
Why "also"? "And" already informed us that you were also thanking Sam.
Great Interview 🎉
According to certain economists, it's possible that the U.S. and certain parts of Europe might experience a recession at some point in 2023. Although a global recession, which is characterized by a decline in annual global per capita income, is relatively uncommon due to the faster growth rates of emerging markets like China, in comparison to developed economies. I have pulled out more than $340k from my bank. After all, the FDIC covers only up to $250,000, and the implosion could have bad effect. Looking to invest into the stock market now. Does anyone know how I could go about it?
The stock market rally still appears to be in the midst of a normal pullback. I just sold my home in the Boca Grande area and I’m looking to remunerate a lump sum into the stock market before stocks rebound, came across a success story of an investors that made up to $700,000 in few months from investing just $250K and I'd really appreciate it if I could get clues and pointers on how to make better profit...
please how do i find the man you just mentioned?
I just Googled his name and his website came up right away. It looks interesting so far. I sent her an email and i hope she responds soon. Thanks
Im worried abouzt my freshly started developer job but I gotta admit the way its being handled right now is rather nice. Distrubiting it to the public instead of developing it behind closed doors and selling to some Overlords
Yeah me about this business data analytics degree
@@aena5995lol l
I am an ai/computer person who has always assured people that language models and things like it could never pose a threat. And this interview was terrifying. We need a lot of regulation very fast.
The scariest part was when he mentioned that it could be used for offensive hacking, and when you connect the dots that language models could very easily be put in a loop to talk to themselves.
And then think about the fact that the only fundamental flaw with authoritarian governments is their inability to make reasonable decisions on such a large amount of information. And that this language model can reason & it’s only limit on the amount of information it can process is how much energy it burns.
Totally agree with you. I think springing this onto the world like this was fucking crazy. There are so many god damned downsides to this that it's unreal.
True. This technology is developing way too fast for the governments to even react, let alone regulate them. Most of the politicians in the US are old and out of touch of the tech world so I have no hope from them. Our only hope is that the EU does something about this.
That's the scary thing. Most discussions surrounding AI has been made in a vacuum that is the US, so to speak. While not perfect, there are at least SOME ethical guidelines these companies for enforcing. Imagine the AI's coming out of authoritarian governments around the word. Our future is going to be frightening.
I too work within IT and worked directly with NLP plus machine learning. I’m always excited by tech but this new generation of AI is sold too quickly. It should stay in incubation for a while longer or now is the time to slow down and make AI require certification or something - depending on field of use.
We couldn’t see where internet was going for or at, we’re still trying to grasp its implications and possibilities. Adding AI as a layer onto this scares me.
@@harbirsingh7266 If they are not out of touch they want to pretend they are (for their own benefit).
What a wonderful way to put it at 6:50
Maravillosa entrevista
Excellent conversation. Really clear and concise questions asked .
i hope youre joking
Did you use chatgpt to make this comment?
Too much focused on the negative and cannot see the difference between sentient and a reasoning capacity
@@it0it0 But that's the reality of it. She is reflecting it's outcome because of how new, strange, and risky it sounds.
This is far more transparent than social media algorithms and data gathering
why is the audio so quiet
I am SO glad Sam Altman is at the wheel. He is so wise, and so humble. Smart AF.
You're kidding right
GPT is a great tool for our minds, I have used it to help teach me difficult concepts in Engineering as a college student and I am surprised on how much faster I was able to learn and get through chapters in my textbook. Great technology with a huge potential for improvement.
Exactly! I am a seasoned fiction writer. It has helped me improve greatly. I write so much faster now and better. The thing is heaven sent. I hate how some people in society, especially the news, are focusing more on the negative things rather than the positives.
@@DarkandTwistedI would agree however ChatGPT and this technology are one of those things where attention to the negatives & it’s potential to be amplified in my opinion is warranted press. There is so much good in this, however I believe we need to keep it from evolving to something potentially catastrophic to humanity
Also great technology with a huge potential for damage and chaos. Anyone who doesn't see the downsides isn't giving this enough thought.
@@zoomingby every thing have a possible negative outcome even a screwdriver or a toilet paper have one you would be dumb to think theres a thing without negative outcome
Also that interviewer asking alot of dumb question. Like "can it lead to negative outcome" maam i cant think of a thing that cannot lead to negative outcome.
The law of unintended consequences will rear its head in a big, big way.
No. AI is objectively a net positive. Your lack of creativity has been exposed.
@@cl1489 come back to this comment after you have been fired from your job
@@cl1489 lol what? Try to explain why a superintelligent, powerful entity would keep humanity around
@@joeyf9826
Because it thinks humans are cute maybe? Or it could be that despite its intelligence it has no will of its own. Of course maybe it wouldn't be able to get rid of us as we augment our own intelligence with BCIs.
Great questions asked by the interviewer.
Great interview.
“There are massive amounts of potential negative consequences” -Chief CTO
Also massive amounts of positive upside.
Doesn't CTO already have the word chief in it? So it's now Chief Chief Technology Officer?
Would you rather she lie?
he's not wrong, just stating the facts. much better than him turning a blind eye to it and giving us answers we all want to hear
@@ezra9243 yeah I wonder what their logic is, they’re like this will inevitably come because the trajectory of tech is going this way may as well do it ourselves make some cash but be good. Altman went to Stanford and have gone to a good school myself they really do teach you about ethics etc