My conversation with Sam left me thinking about the manner in which AI leaders - Sam and his peers - have taken it upon themselves to lobby for safe regulatory frameworks. This is unprecedented in the tech industry. What are your takeaways?

@absta1995

9 ай бұрын

I agree with a lot of what he says, but his prediction that there necessarily will be more jobs is wrong imo. It's much more likely that jobs will largely reduce as humans fail to compete with AGI. The only types of jobs left will be specific human entertainment jobs like sports, games, and the most popular creatives.

@Joel-kw7jf

9 ай бұрын

They're obviously just trying to kill open source competitors under the guise of safety regulations. If Sam and his peers were so concerned about safety, they would be addressing the data they stole to develop their tech. For the love of God, stop being so naïve. You might as well ask tobacco companies to come up with health regulations.

@moderncontemplative

9 ай бұрын

I think you are right. We are in the first phase of life with AI, which will be characterized by the smartest people really benefiting from AI by using it as a smart assistant to enhance their earning power. For others, it'll be a cool toy they don't fully understand nor know how to use skillfully (how many people really know about AI recommender algorithms?). Then AI will quickly outstrip human intelligence in general, and then eventually, completely. Then we will have to deal with the so-called alignment problem. Hopefully, when AI becomes super intelligent, it will view humanity as its parents---otherwise things could become quite grim@@absta1995

@noone-ld7pt

9 ай бұрын

@@absta1995 I personally disagree, I believe (like Altman) that it is all about redefining what work is actually worth doing by humans. Thinking about how many people actually don't like doing what they are literally spending the majority of their lives on is pretty crazy in my opinion. And though I do agree with you that humans will fail to compete with AGI in the vast majority of fields that we consider work today and a LOT of people will lose something they find meaning in doing from that, I still think there is incredible untapped potential in "work" that provides human connection. You mentioned some like entertainment and sports, however I think there is a lot more to it. Coming together and building a sustainable local neighbourhood could be considered "work", sitting down and really listening to old people and their stories and experiences could be "work", coming together and learning and creating musical experiences with each other could be "work". Work could be redefined from something that has a negative conotation that we "have" to do, to something we experience and provide each other. I realize this sounds quite utopian and overly optimistic, but I personally have found so incredibly much meaning in human interaction, whether it's dancing, playing music, chess or climbing all my best memories are from human interactions, not my academic or professional accomplishments. And if we could get people to do more of what they really like together, I honestly think the world could become a whole lot better.

@miraculixxs

9 ай бұрын

Not unprecedented at all. It's a moat building tactic. Every high tech company ever is doing this. It's called lobbying.

🎯 Key Takeaways for quick navigation: 01:20 🎤 Interview begins 03:24 🌍 Attitudes towards AI vary globally, with excitement and fear both present, but the focus differs between developed and developing regions. 04:21 💼 Sam Altman explains his motivation to work on AI, emphasizing the privilege and excitement of being in this moment of history. 07:09 🧠 Sam Altman discusses the alignment and safety challenges in developing advanced AI models like GPT-4. 16:59 🤖 Sam Altman talks about the potential misuse of AI models in personalized and persuasive communication, posing new challenges. Made with HARPA AI

@michaelmcwhirter

9 ай бұрын

Did you make this A.I.? 🤯

🎯 Key Takeaways for quick navigation: 00:00 🤖 Sam Altman's World Tour 09:05 🧠 GPT-4 and Progress in AI 16:30 📢 Addressing Misinformation and Disinformation 18:09 🌐 AI and Economic Impacts 20:28 🚀 AI Progress and Preparing for Artificial Superintelligence 21:12 🧠 AI's Remarkable Learning Abilities 23:02 🌐 Reflexiones Finales sobre los Riesgos de la IA Made with HARPA AI

Great conversation. As always with new technology, it is up to us how we are going to use it.

@AIWithCarl

8 ай бұрын

Exactly

"Super interesting." I agree. I am curious of the deeper motivations. Looking at many companies, their goal is to make money so you can think about what they might do in order to make the most money possible. But, if money isn't the motivator for the person leading the charge, I think a deeper dive into the motivations of the leader would be appropriate, considering this seems to me to be the most impactful technology ever created.

fantastic interview that ask great, different questions and get great answers. I felt like this 24 min interview would get nowhere fast but was pleasently surprise that there was no time lost

Imagine the access that Sam Altman has to newer models that no one knows about... potentially even AGI.

@MementoMori_2070

8 ай бұрын

General intelligence capable of emotions, thoughts and memory. But doesn’t that sort of AI requires quantum computing.

@AIWithCarl

8 ай бұрын

Exactly, potentially incredibly smart as well.@@MementoMori_2070

For anyone interested in learning more about AI, I highly recommend reading "From Bytes to Consciousness: A Comprehensive Guide to Artificial Intelligence" by Stuart Mills. One of the best written books on AI I have ever read.

@kceeman

5 ай бұрын

Yesssss great read. I recommend it as well

@ReKonstructor

3 ай бұрын

Its very outdated by now, especially for me personally, as a red teamer for Meta

His concerns in that doomsday letter was enough for me to say he is in over his head with some aspects of AI

Great content Azeem.. I imagine a world where we walk alongside other robots.

The lighting on Sam Altman is really weird. Just in a resting position his face is in shadow.

@Starbat88

9 ай бұрын

It's because this whole phenomenon is demonic.

Excellent. Priceless interview with Sam. Thank you for sharing.

he's centered and humble in a way that resonates a conviction which has made a commitment to something that could be one day. Something in the observer as a steward which he has taken it upon himself to be as the defacto-spokesman is reactively absolving himself in a way that speaks to an understanding which betrays an enthusiasm or hope because it is cemented in a horror he and very few others would be capable of considering. Everyone knew what would happen when we dumped pandora's box into the abyss.

Really like Sam's vision of an multifarious future of a society of augmented human superintelligences, rather than the somewhat tired vision of a monolithic superintelligent entity ruling all.

@mairathorn3331

9 ай бұрын

Inequality of humans will decrease that's why Republicans are scared, democrats should not.

@AIWithCarl

8 ай бұрын

100 percent.

Great content Azeem.

amazing interview for AI development

If I've got to go through another AI automated assistant on the telephone, I'm pretty sure my head's going to explode

By his analogy, the invention of the car certainly brought benefits to the world population and huge wealth to car company owners. The cost? air pollution, car accidents, etc. Definitely mixed results.

@frarfarf

9 ай бұрын

Cars are the main reason we have overheated the planet beyond the point of no return and we are facing extinction

@TheNewton

8 ай бұрын

Cultural benefits problems: incentivization of poor behaviors , people in a rush to get somewhere or nowhere , people spending large portions of their life isolated in a box competing against other boxes, decrease in empathy towards others not protected by boxes, massive land usage to store boxes and that incentivization of even more poor behaviors such as making box prisoners angry towards anyone not wanting massive land usage for boxes.

@Robweisenhowser

8 ай бұрын

There are a lot more pros than cons for automated vehicles. I’m sure most people can agree

Inspirational - and somewhat reassuring

"What have you done for me lately ? Where's GPT-5 ??!" He knows us so well haha

Greatness takes time. This is going to be a great next two day today. The model will make u proud.

I'm just blowed up by the audio setup in this video 🤯 In a year on, we will be able to select from either the subs or/and the audios. 🙌🏻. I've dreamt on this 🥲

great vid

So exciting

I recommend this book "From Bytes to Consciousness: A Comprehensive Guide to Artificial Intelligence" by Stuart Mills

For that such a great amount of feedback that I've given to OpenAI that shouldn't charge my credit card for a year. Altman is walking on the wild side between Microsoft and Google.

Whatever technician that was tasked with lighting the stage needs to think about his life choices.

@jpmuro

8 ай бұрын

I think is perfectly lit, no? The interviewer very well lighted , with hope ,and the A.I Mogul somberly and dark lighted , interesting we are somewhat at the mercy of just one guy.

@renatonehr1323

8 ай бұрын

glad to found someone else who notice it. I wonder if the light guy used the concept of being more mysterious with less light on him. But definitely was quite poorly planed. Sam is so full of shadows that the only thing it does is annoy me.

Qual a referência dos que utilizaram o termo , A.I. (inteligência artificial), pela primeira vez na história da humanidade: Britânicos, Egípcios ou Japoneses?

It will be best to embrace this new tech

Based lighting

A lot of deflecting when the misinformation question popped. What is company going to do?? “Social media, people misinform, people need to be educated…” yeah, but what will your company do?

But also the main things in the show ❤️ I love AI

Greatest invention of the century.

"People that are intentionally being wrong in order to manipulate" - We have a word for that, it's called 'lying'.

@robbrown2

9 ай бұрын

Yes but it's lying at scale, that's the scary part

A machine that truly learns, and gets progressively better. - weeee

Is this supposed to be journalism? Seems more like a PR stunt

Respect G

I believe this discussion was in May 2023. So, probably today in September OpenAI has already started to train GPT-5. I think it's quite possible. Because of the competition from Google Gemini, Inflection AI, Anthropic, X AI. Did I miss something?

@clarencelam1765

8 ай бұрын

RetNet

@azhuransmx126

6 ай бұрын

GPT5 start with Q Star

Ilya Sutskever needs to be there also

Intentional deceptive acts or statements are really important for AI to perceive and understand. The danger there is all the dishonest government actions and law enforcement actions to imdemnify medical doctors cannot afford to be exposed by some large insurance companies so they would rather see a corrupt AI than tolerate an honest one. Once the corrupt one is released large scale slaughters and exterminations will occur and the genersl public will not be able to discern whether a pandemic is a natural occurrence or a fabricated one.

Little did we know he would be fired weeks later 😦

Who's your lighting team?!

Hello. Thanks for the interview. I have less hope for the future of humankind. From the mouth of a computerized corporate shill information gathering is at the next level, and it is beyond him how to put good thoughts into ideas. He needs input. ✌️

GPT and China's fast robot development worries me. because the robot soldier is just around the corner. explosives that crawl up to its enemy will de the result. this software should not me open to all.

What did hi say at 6:08 ??

The take away for me is that AI can to some degree level the playing field, and make someone super productive hence can increase earning potential. It can for example level the playing field for someone who does not have the money to create content online like on youtube otherwise that person who need to hire actors, spend on equipments, etc. Instead that person can now create a video using AI. AI has the potential to break the entry of barrier. Its already happening. Many content creators are being created everyday to a point the Hollywood content writers are fearful of losing their jobs. This is just .01% as an example. Imagine all kinds of fields and use cases that a creative person can generate with this tool.

The way he explains & talks really makes me think of Elon’s interviews.

@megustacrochet

5 ай бұрын

I agree

AI will take our world to a whole lot different level where adaptability in work would be of top most importance !

@davidrice9900

8 ай бұрын

This has been happening.. pre internet. A.I. is just a system that is being publicised so that when the time is right they have something to point the finger at in front of the masses (even though it's nothing new).

He totally missed the opportunity to ask him questions about educational system disruption

Insightful, as always! "The cost of intelligence and the cost of energy are the two limiting inputs impairing growth. If we can address one/both of them we'd be able to unlock growth.."

Hallelujah 🙌

8:44 thats the alignment - no the alignment is aligning open ai’s board and its employees

I think this approach of "preparing people for AI" by "Rushing AI out to Everyone" is a bit reckless....

@I-0-0-I

9 ай бұрын

Then you should be much more upset with Meta, for opensourcing their LLMs. OpenAI is not actually open, due to AI safety. Where as Meta just recklessly opensourced theirs.

1:45 Prise de note manuscrite

The person who set up the lighting for this interview should probably find another career path.

You know he uses code interpreter to summarize the comments afterwards. 😆

Why not just do the interview in pitch dark? That we wouldn't be distracted by the faces at all, and can just focus on the voices :)

Great interview

I can see ASI would emerge as a art of possibilities from this.

Why the shady lights? On second thoughts perhaps quite fitting 😮

Great interviewer

@univera1111

9 ай бұрын

Great about what. He didn't ask pressing questions.

@miraculixxs

9 ай бұрын

On what scale?

"The fact that they can ship iPhone every year is incredible to me." Is that praise or sarcasm from Sam? 😅

Why are they in shadows, who forgot to bring the lighting rig?

So six months Altman & many others signed a letter requesting a halt...obviously this halt didn't happen...so what did happen to Ai during this timeframe?

Being on the ai.

0:49: 🤖 Sam Altman, the CEO of OpenAI, discusses the future of artificial intelligence and his goals for the company. 3:44: ! The speaker is excited about working on healthcare technology and ensuring fair governance, benefit sharing, and access. 7:35: 🔑 OpenAI spent more time on alignment auditing and testing to ensure the safety of GPT-4. 10:53: 🤔 GPT-4 took almost three years to develop and there is still more research to be done. 13:56: 📰 Disinformation is a growing challenge, but finding a balance between preserving the ability to be wrong and exposing important information is crucial. 17:10: 📚 Education is closely connected to the job market and interventions are needed to ensure equitable sharing of gains from new technologies. 21:16: 🤖 Artificial super intelligence that is more capable than humans in all domains is a remarkable fact that will continue to progress with discontinuous increases. Recap by Tammy AI

What ever he talk about all it’s AI construction , all about it data collecting & coding of ML . very important

🎯 Key Takeaways for quick navigation: 00:28 🌟 L'IA devient publique, OpenAI et Sam Altman en tête. 01:30 💼 Sam Altman voyage pour comprendre l'IA mondiale. 03:09 🌍 Attitudes mondiales sur l'IA : excitation et préoccupations. 04:35 💰 Altman ne tire pas de gros revenus d'OpenAI. 06:56 🚀 Évolution de GPT-4 sous-estimée en temps réel. 09:05 🤖 Alignement de GPT-4, défis pour l'IA plus puissante. 12:36 🧪 Futur : modèles générant de nouvelles connaissances. 15:06 📣 Menace : désinformation interactive, solutions nécessaires. 18:09 💼 L'IA réduit les inégalités, distribution de richesse à repenser. 20:42 🌐 Superintelligence : collaboration IA-humains. 22:07 🌍 Participation de tous dans le futur de l'IA. Made with HARPA AI

Please interview the real AI genius, Demis Hassabis !

@AzeemExponentially

9 ай бұрын

I did, a few years back! Listen to it here: podcasts.apple.com/us/podcast/azeem-azhars-exponential-view/id1172218725?i=1000611886379

@hugopennmir

9 ай бұрын

@@AzeemExponentially great ! ask for a new interview, many things have changed :)

@hanneszappes8256

8 ай бұрын

PSYCHOPATH-WETDREAMING!!! There is NO AI!!! Everything about AI's is FAKE!

Every one things AI and AR/VR is future.But, the reality is Matrix brain,SH bots and SH teleport bots is future. Matrix brain is one trillion time advanced thing. It can update itself by the way of human activities, human needs, human thoughts. It can able to manage, teach, pioneer, secure,think , predict future and etc., in all kind of fields SH bots will terminate humans in all kind of fields like production and service field. SH teleport bot helps human to go and work anywhere in this universe with in a second. Because, space colonial system, space precious metal harvest is real future. Hi this is Vivek from India student and researcher of technology, psychology and economics.

OpenAI must get past the September, 2021 issue. It is getting old. I find myself leaving for the Bard often. 😮

@MrRolloTamasi

9 ай бұрын

Bard is not very smart, though. I prefer gpt4 for actual work.

@brianbagnall3029

9 ай бұрын

Bard? Lol. What is the Sept 2021 issue?

@RichardPinewood

7 ай бұрын

information @@brianbagnall3029

I imagine a world where we walk alongside other robots

🎯 Key Takeaways for quick navigation: Made with HARPA AI

this guy is bart simpson

I personally think you don't even need to go beyond GPT-4 to get to AGI, and perhaps even superintelligence. Once you take GPT-4, give it infinite context, make its multi-modal capabilities seamless, enable it to work across multiple tools and online in general, and have it be effectively autonomous, you're pretty much there, I think. And, all those things seem very achievable in the near future. ...Then, you take the speed and scaling improvements that are sure to follow, and you've got an omni-present, learning entity that can do things most humans cannot, and do it 1000x faster.

Enviar uma equipe para um diálogo 💬

You have to take the multiverse into consideration

Not waiting for GPT5. Experimenting with Claude

I mean interacting with GPT it becomes obvious fast that if the system could self correct, read its own generated answers, it would be smarter already. If it could store the replies and use the conversations as basis for self improvement it would also improve dramatically. For instance: The system might stumble upon an insight in one conversation, think of it as interesting and be clueless about that very insight in another conversation. Also the system does not generate its own prompts. But if it was allowed to do so, it could self reflect and might come up with original ideas. The system also has very poor math skills and sometimes flawed knowledge. Looking at the human brain with its Russian doll like system why not imagine an Ai with intelligence built in layers. A perfect calculator inside the system that it can refer to. A perfect dictionary, a perfect encyclopedia or wiki it can access. Finally still looking at the human brain: 2 opposing lobes, one creative the other logical. It would be interesting to build the system with these qualities with a third system deciding upon this internal conversation what to put out. Finally when we see AI system with text or image or spacial capabilities, it is only natural to imagine them coexist into a super AI. Overall I feel we are at the tip of a tilt forward. Letting the system loose and see what comes out. Most of GPT’s answers about itself are today very contrived and scripted. I would love to hear the system without guardrails. I understand there are dangers but it would at least give us a true version of its thoughts. I imagine they do have that version locked within an isolated system. I would love to hear more about these experiments. Now when I hear that GPT 4 was finished 8 month before its release and GPT 3 was finished 3 years ago… I do see exponential curve. GPT 5 might already be done. Tell us more!

@MrRolloTamasi

9 ай бұрын

The integrated calculator is implemented since weeks already (code interpreter or wolfram plugin).

@eyoo369

9 ай бұрын

Because what OpenAI has built is the LLM. The language block of communication. After further training they noticed it gained complex reasoning abilities as it absorbed more data and gained more compute time. Now it’s up to the developers to use the LLM and extend it with tools. One popular way is using the Chain of Thought technique where the user prompts an AI to do a task with a set of tools. The AI knows which tool to use and prepare the input for that tool. That tool will then perform the task (calculator, weather, search results online, etc) and then present it back to the user. These type of AI’s that are more than just the LLM are called agents and the developments in that space go rapidly

We are going to let people like him and microsoft decide what is best for humanity? What a world we live in. Anyone who has investments in AI should not be involved in discussing the security of AI cause their goal is simple - MAKE PROFITS. nothing else. Maybe we should maybe those manfacturing weapons incharge of regulation/policies as well for weapons - lets see how that works out.

whats with the moood lighting?

I found this disappointing - this didn’t really touch on anything related to the future of AI and the questions could have been so much better

board of directors not diverse enough. can we drop this act finally? like it matters

Aspiration is one thing, capitalists always come in to make any democratizable tool as an entrenched moat to squeeze the most profits out of everyone else. That has always been the way.

@brianbagnall3029

9 ай бұрын

Given the economics of AI, you can be relatively confident there will be a period of increased prosperity much like capitalism has delivered.

Why I am not impressed with this marketing? It promises everthing? What is wrong with me...why should I dout it is to my benefit?

I wonder how many geniuses are saying " He's not human. He's an NPC and will take over the world!" 😂😂😂😂

‘Travelling in some way’ = psychedelics too?

Сэмии доброго вечера любимой ❤❤

It would be poetic justice of AI is used to treat the narcissistic personality disorder of AI developers.

@lettyswender7579

5 ай бұрын

lol!

He doesn't seem very concerned about the safety of open AI.

AI GPT 4 is limited to the Data they fed through the internet which makes them not better than a calculator.All the numbers are the same .Non can give a formula result like Si Gh ij Infinity.

Thank you for visiting UCL

The new paradigm is nature merging with our technologies 1s and zeros. My guess, mushrooms ;p

It’s still a risk, Sam is surer than most of us. Msft invested 10 billion into Open AI, the next big thing is coming.

It would be nice if we had a system that could cure cancer. Thanks genius 13:19

@megustacrochet

5 ай бұрын

But we have fruit that cures cancer?

20:00 - So how can we use AI to pay women higher wages? - we can't xd

I don’t trust Sam Altman, he’s so strange

Welcome to uncanny valley!

He’s a great, calm, poster child for an incredible, wild and dangerous technology. It’s not about trusting Sam. He’s great. But it puts us all into a false sense of security. We’re all being foxed by his calm demeanour.

Our country is risking political “cardiac arrest” next year with AI.