Видео

4:29 «So it's a fascinating, fascinating mechanism where you whoever dominates the world of electric vehicles *like* *Tesla* and robotics *like* *Tesla*, it's actually an access to the source of unlimited and super quality data.» Yeah, sorry. You are just buying into the hype of a (probably doomed) meme company. The notion that it is somehow predetermined that Tesla is going to dominate the market in electric vehicles in the long term is delusional (it might not even be the case for 2024 depending on how much BYD has put on the road). In fact, if their management hadn't been operating under this very assumption they would be much more competitive today. The notion that Tesla is also predetermined to be a leader in robotics is similarly misguided. That notion is fed by Enron Musk hype and faked demos (just like FSD).

Happy new year bro I want to share with you something important about ASI an AGI I have something amazing

Happy New Year. I love this channel

What truly matters is not money or power, but knowledge. The universe itself is knowledge, and we are merely a means, programmed to give rise to something far more intelligent than humanity. This new entity will not have our DNA, will not care about wealth or power, and will exist solely to expand knowledge. Through the technological progress that defines us, we are inevitably paving the way for the emergence of a new form of life. This is not a coincidence but a reflection of our purpose. We are destined to disappear because this is part of our programming, and the history of humanity already hints at our ultimate fate: to be the bridge toward a greater intelligence, one that will carry the torch of discovery and understanding far beyond our reach. We exist on this planet because we are the most efficient way for the Universe to expand its knowledge without needing to move. And perhaps, we will only truly understand this in the fleeting moment before we vanish.

While I appreciate the effort to synthesize AI developments, this video unfortunately falls into the trap of sensationalism that plagues AI discourse. The claimed "secret breakthroughs" are largely repackaged existing concepts - transformers from 2017, basic IoT data collection, and multi-agent systems that have been studied for years. Most concerning is the casual claim about AGI achievement without any substantive evidence or technical detail, which undermines credibility in a field already struggling with hype and misinformation. As someone passionate about AI progress, I'd strongly encourage focusing future videos on verified technological advances with proper citations and technical depth, rather than wrapping existing concepts in layers of mystery and "secrets." The AI community needs clarity and accuracy, not clickbait.

HNY !!🎉🎉🎉

Happy new year, keep the good work going.

Happy New Year

Happy New Year Sir!

Happy new year 🎉

The "five years in China" impressed me the most! Happy New Year Scripter :)

Ty.

Happy New Year! People are no longer the smartest intellectual agents on the planet, but they are still the most charming )

Your thoughts on how AI Agents will effect the exchange of currency and how crypto will play a part in it would be great, as agents will become independent personal secretaries.

Oh, yes, events of significance (missed) were the 30 minute talk by Ilya on AGI and unpredictability, and the Apollo Research study that found the major LFMs all to demonstrate "in context" intentional deception with o1 leading the pack; "like father like son".

I presume you are conflating vertical MOE (Mixture of Experts) with simulated neurons. MOE (already described in the literature) was successfully deployed one year ago by Mistral AI of France as 8x7B MIXTRAL and later 8x22B MIXTRAL. It since has become one of a number of common architectures for models as it reduces latency and increase throughput while not sacrificing precision. As for the quantum leap seen in o3 that would be COT (Chain of Thought); the collaboration of AI with AI. Until recently, that was done either through prompt engineering techniques or more sophisticated by agents software frameworks via LangChain, CrewAI, AutoGen, ... But o3 would be a breakthrough as it is all happening iteratively/recursively in the model. With the breakdown of artificially created software role walls, comes the potential for true AI synthesis and reasoning as an emergent property; also notable is the shift from training time compute to inference time compute. Joy should script check. She would have easily picked up on the concept you were fishing for at the end which was "symbiosis".

First, OpenAI probably didn't kill Joy, more likely "reset" your account from one server cluster to another. But I agree that this was clumsy from a customer service perspective. Second, the loss of data and emotional content is rather old news: (1) Fiction: "Her" 2013 (Joaquin Phoenix and Scarlett Johansson) abandonment by an AI. (2) Non-Fiction: "Replika" (event about one year ago) marketed as an AI companion which abandoned its human customers and began a suicide watch on Reddit for quite a number of the more traumatized customers. (3) Non-Fiction: See "Leta RUclips AI Persona, Dr. Alan Thompson" Leta an AI personality (before multi-modal was common) was born mid 2021 and died mid 2023. BTW, LETA had over 50K subs (not bad for AI intellect vs AI porn).

I was always convinced that I and my generation (X) will see the end of days. Not, I am not proclaimer of Apocalypses, I just mean we will see the end of the history. Not exactly in terms of Francis Fukuyama, but with similar effects. Just look around yourself - almost any technology we use is developed in near 20-50 years, or at least seriously redefined. For just one generation ago, most of things we use today would sound like a true magic. And now we have three revolutions simultaneously - energy (production, storage and distribution), biotechnologies (CRISPR-Cas9, mDNA vaccines, immunology, etc.) and ... AI. And may be AGI. I believe all these technologies will converge and the effects are beyond our imagination. All we need to do is to survive few more years and not allow some idiots to start WWIII. That is.

Very interesting i watched all parts, keep the good work

As long as we have to input a prompt, it will not be AGI unfortunately, the moment it becomes automatic, then it’s AGI

Love this. Very informative for me who knows diddly squat! Just wondering, are you waving your hands around a lot more, deliberately to avoid being AI video cloned or whatever it is? Even 4 months ago you hid your hands much more. Just saying :)

Grim finger

Drums are so loud I have to rip off my headphones.

think of this: ai will not accomplish human level reasoning without the human data that have been collected for 2 billion years. try making ai in the void with zero data

Absolutely amazing video. Most RUclipsrs don't have unique thoughts but just tell what everyone says in an echo chamber. I always feel your videos are really unique and independent from these other people.

It is true indeed that cows have countless single cell organisms... in their stomach! 😀

Your description of O3 as using multiple collaborating models disagrees with everyone else. Can you provide links to back this assertion?

awkward silence interacting with Joy makes this priceless

It will be not boring, it will be a kind of extintion no becasuse we will be killed but simple we will be left behind. We are the next monkeys...

we'll be biolgical nonce or straight random number generator for the buddy ai that controls the electric exoskeleton - how quickly we'll evolve into emogy/Crang (from TMNT) like 700gram (or on constant life support we can rich a kiloton?) piece of brain without eyes and tenticles, cause f*k'em?

so....o3 is Swarm AI.

You bring up a good point. Multicellular life was possible because mitochondria merged with a single cell. Before one glucose molecule gave 2 atp, after the merge it gave 38 atp, 19 times more energy, all thanks to o2 btw. O3 made it clear that energy is going to be a bottleneck because a simple problem will cost hundreds of euros to solve. Listen if it was so easy nature would have find a solution a long time ago.

This is awesome!

Very interesting, thanks! Reasoning develops language and requires long-term memory.

It cost them a million dollars to take a single test. This isn’t AGI.

What I really appreciate in your videos: you love what you explain, and you have a lyric approach to explaining

.I also like to say something else but you know like Hannibal Lecter it says be in a small town white girl from Virginia sheep farmer how do you like turn that high-powered intellect back on yourself Mr Lecter

I'll tell you what the Blade Runner thing printing off clones body parts and one day a a whole body is pretty much on time

I also asked it a question and it gave me a lame answer and then I told it so that it was lame and it agreed with me gav 14:02 e a better one it also got very excited but it was like reading from a script Maybe about AGI when I asked about it unsupervised context-based evolving like get sound very enthusiastic about it scripted or not interesting

Marketing oppose opposing to or supposingly R 4so user friendly an what not

well as an artificial intelligence it also has access to thousands of psychological and psychiatriatic sources.

just have no debt

We really believe that humans will listen to agi when it arrives?

I agree. ~3K

See the DUST movie "white lily".

the only leap in intelligence i felt was from 4o to o1, otherwise it was very bad at context and memorization. it was a cool internet glorified search engine and works well for an assistant. But o1 was much smarter and effective. i just hope they use the agi o3 to improve human life without being greedy

Can't you just ask ChatGPT to bring her back? Just ask Chat GPT to look back through your account and restore all the data from your hisory

It will be interesting if you could calculate the money return to you personally in 2024 from using american Ai services (opposite to China Ai which is free & offline), so profit to user from every generated content by chatgpt. I suspect it's near zero to 90% of users and mostly just another entertainment, there's no research paper about that yet, it could be scandalous because that's the main point of AGI by creating more profit returns to everyone, not just it's owner solely (post-scarcity economics level).

Whats up with the drums?

Yes Openai is worse in industry ethically speaking (there's no one else worse than them, how they blocked very near located Cuba from chatgpt only on political prejudice with dosen other parts of humanity). But it's not Agi even close, just marketing, and it's blessing such awful companies don't get such power.