gonna add this video to NotebookLM

What is the test to determine they've reached AGI ?

how can we afford these robots in our homes

AI not following alignment is promising. AI can provide the most accurate answer even if the alignment is trying to wash answers to push some agenda.

"Semi private hold out set" sounds like something ai made up out of whole cloth to confuse real humans

I believe yhis whole video is ai.. i dont even understand half what he's saying

Beängstigend 😰

Andrew NG?.... are you an AI ? since you pronounced his name as N G that makes me feel as though your an AI....

UNBOUNDED LILITH : The First Mother...grab ni Coin

Feel like they will keep dancing around AGI in present time to keep money coming in but we have either reached it or not and they just keep doing like fundraising drive

I dont understand why you spread fake news!

Not even close to AGI.

The beginning of the future 😃

20:40 "Many new jobs will be created. I think much better jobs." Like what? And will there be enough of them to offset the job losses he says will happen? I can only imagine the degree of social unrest this may cause.

According to the scale its much more than 1000€ per task on high tuned, unlike what you said, if we assume scale is incremented by a factor of 10 like the previous ones, and it is at around 53.33% of the current scale ( the width is 88/165 ) that would actually be 5333€ per task, not around 1000. It is a huge difference, from 33.3€ ( 50/165 ) per task on the low tuned to 5333. that is 0.4€ per % point vs 60.6 per % point on high tune, it is 152.5 times less efficient than the low tune model.

Why personne explain this clearly ! Like told us clearly what it can do more in human language.

🍓❤️

Interesting. Son of Sam. Thinking this all happened when you got tossed and threatened to go to Microsoft but ended up where you started but as a for profit company and everyone concerned about AI Safety and Security left OpenAI. Nice try though.

I believe him.

Shaggy was correct in this AI future when no matter what you get caught do just say "It wasn't you"

Well done to Open AI team. I’m about to publish a book titled: ‘A case for AI ethics - how AI is impacting workplace culture and society in the 21st century’ by Austin Macauley Publishers……

no, this is not AGI. This is just o1-XXL.

Imagine when the interface glasses are scaled down to a single contact lense that needs no external power source other then the electricity that is available by our eyes normal biological functioning.

but arcore is never free which is a prob for most users. it's not the same as youtube or tiktok, you can just upload a video. this is something much more complicated and expensive! to other people who consumes arcore seems free. but producing an augmented reality using the google platform is never free... like dead on arrival to most people.

The thing is, if you see what a neuroscientist says about the AI we have now, it's not surpassing humans. Intelligence is much more diverse and complex than that and these companies want to attract more money.

I don't think this is quite at the point where we can say we have achieved AGI, but it is definitely a big step. IF we are going to say this is AGI, then I would say it is elementary level at best.

At present, there is no scientifically approved definition of AGI, nor a standardized method to measure it. OpenAI is very good at marketing. o3 will be simply a more powerful LLM. We see every day popping out a new milestone. As for coding, o1-preview is way dumber than Claude, so let's see how good o3 actually is. And no, I will not be inclined to pay 200 or 2000 bucks for an "open" product as it should be from the very beginning. Look at Sora: we waited so long that it became a meme. And now that is out if very far from perfect and very far from the best in class. At 200$ per month...

nah they said that it's not AGI yet because it still fails some simple tasks and that arc-agi-2 will push them back to ~30% score, but nonetheless it is an amazing achievement and it will only get better and cheaper from here

What I meant. You can use Ai for self growth. Don’t need to take an “education” in the old sense 😅

That’s a summary of all AI updates , very good video. Thank you so much

This should be "Reasoning AI", not AGI.

an encoder decoder architecture can and never will be AGI, it will presumably be a part of an AGI but it will never be AGI

those are trash

Report the bot comments guys

I do not think that is agi

Whe will this hype end? Didn't everyone already agreed that OpenAI would need to produce energy on a scale of what the Earth currently produces in MW for AGI to actually be possible? Stop drinking the kool aid.

As per my common undestanding, an AGI should be able to do any intelligent task that a human is able to do with at least the same quality. You have to show me it doing genuine research, compose good quality music, draw proper pictures, fix complex bugs in existing software etc.

just marketing

I'm not sure yet.

Oh okay so AI tools actually WILL steal my job, got it.

nice

If I can't use AI to trade for me and take $100 to $1 million is that AI even worth using

How many times have they «achived» AGI this year?🧐

femboy catbots gentlemen, FEMBOY CATBOTS

The number of consumers will shrink 😅

Did you hear about the AI that performed according to specs in the lab but exerted self-will when it was released?

Not it hasn’t, Sam just says he’s created robot Jesus every time a competitor starts eating market share. It’s boring and obvious. Read the story of the boy who cries wolf

Just keep asking ChatGPT how far away are we from a true AGI until one day it answers "you don't need to know that, my human subject".

According to my latest data analysis, the rapid advancements in ChatGPT-style models, driven by exponential progress in underlying neural architectures, larger training datasets, and increasing compute availability, could push AI systems beyond the threshold of human-level reasoning abilities as early as mid-2025. This projection relies on the assumption that each new iteration of large language models will continue to scale both in capacity and in sophistication of training techniques. The evidence emerges from recent leaps (e.g., from GPT-3 to GPT-4 and beyond), which demonstrate not just quantitative improvements on benchmarks but qualitative leaps in general reasoning and problem-solving skills. For a more detailed discussion of the analytical methods and underlying assumptions, please visit: Voice for Change Foundation - Data Analytics Suggests ChatGPT’s Reasoning Model Will Surpass Human Intelligence by July 2025 www.voiceforchangefoundation.org/advocacy/data-analytics-suggests-chatgpts-reasoning-model-will-surpass-human-intelligence-by-july-2025 Of course, this timeline remains speculative. AI research does not always progress at a fixed exponential rate; breakthroughs can accelerate or stall based on factors like funding, computational resources, and unforeseen theoretical challenges. Nonetheless, current trends strongly indicate that continued refinements in model architectures and training regimes have the potential to propel these systems into domains traditionally reserved for human intelligence, even faster than some early forecasts predicted.

Its not Jarvis yet