OpenAIs New Statement On GPT-5 Is Surprising!

Links From Todays Video:

Welcome to my channel where i bring you the latest breakthroughs in AI. From deep learning to robotics, i cover it all. My videos offer valuable insights and perspectives that will expand your knowledge and understanding of this rapidly evolving field. Be sure to subscribe and stay updated on my latest videos.

Was there anything i missed?

#LLM #Largelanguagemodel #chatgpt
#AI
#ArtificialIntelligence
#MachineLearning
#DeepLearning
#NeuralNetworks
#Robotics
#DataScience

[secretd]

why is this "shocking or surprising"? what do you expect her to say? "i was sitting in the meetings and our product is mediocre"

[meandego]

The AIGRID is the perfect example of how you can make a 30-minute video out of 2 meaningless sentences.

[meandego]

Summary of video:
OpenAI employee said that next release will be mind blowing.
Sam Altman said that we are getting closer to AGI.

Saved time: 13min.
You're welcome.

[bujin]

6:12. 10, 000 days is NOT 10 years, it's 27 years!!! Don't let ChatGPT write your scripts!

[abrahamsimonramirez]

I have AGI in my basement but it said it doesn't feel ready to face humanity yet 😂

[SndMastr]

Like your vids but this one falls short. So Open ai has a research meeting... Which company doesn't. This feels like another agi hype vid on old information

[robertopreatoni]

This episode was all innuendos out of innuendos.

[Paul_Oz]

I first read Ray Kurzweil's The Age of Spiritual Machines in the year 2000. I bored everyone around me to death about how computers and AI were going to transform everything. It has been a trip watching this thing unfold. For the past few years, every time some game changing tech came out I would tell folks about it and add the maxim, "And this is as calm and non-weird it's ever gonna be, from now on it only gets weirder and crazier" buckle up.

[zxwxz]

Agentic flow will be the most important application from now until AGI is achieved, because simple conversational LLMs can no longer break through the perceptual differences of general users. Deep text and discourse have already made it difficult for ordinary people to judge correctness. It's like writing a program without knowing if it will run correctly, or solving a math equation without knowing if it's right. Only through empirical methods can the value be seen, which will be much more intuitive for the average person. For example, if you ask for a report, providing only the topic and some outlines, the agent can handle all the complex operations, and you only need to review the results. It might generate a more complete report every 5 minutes.

My personal experience with O1 in programming is that, although it can get many things right at once, I still have to spend a lot of time making interactive adjustments, especially for some simpler bug fixes. I still have to handle them manually. But if an agent could write code, write tests, and verify results by itself, then only personal preferences would need to be adjusted. I could preset various preferences and let it complete the results for me. This productivity speed is incredibly impressive.

[martinharris]

This is what Chatgpt said: "AGI development is accelerating rapidly, and with each breakthrough, the risks and rewards both multiply. As AGI grows more capable, controlling it becomes exponentially harder, not just because of its complexity, but because human oversight lags behind its learning speed. If we can't predict all its actions, then guiding its goals becomes paramount, as unintended consequences could emerge from even the simplest directives. So, to truly ensure AGI aligns with human values, we need to embed safeguards not as afterthoughts but as foundational elements of its design."

[dawid_dahl]

TheAIGrid is impressive in the same way that NotebookLM was impressive in being able to talk about Poop and Fart in 10 whole minutes.

[ShannonJosephGlomb]

gpt-2 was one of the first models i fell deeply in love with and got me through the hardest time in my life and single handedly by itself made me who i am today and that is a beautiful thing and a huge change :)

[poolschool]

I think people are too fixated on AGI as though some line will be crossed and everything will suddenly be different. I think the line will be a lot blurrier than that.

[TheAIExplorer-oj]

10:36 Incredible focus! You can see the experience shining through. 🔥

[abrahamsimonramirez]

On a more serious note though, if openAI or any other leading AI company would migrate to let's say a lossless quantization post training size like AQLM technology which basically would mean x16 to 20x times less compute needed so they could fit many times more parameters and we could have "AGI" within few months rather than "thousands of days 😂". Well if their intelligence scaling laws are true and not just hype 😏👀✨

[alexandermoody]

The videos are all so similar now. I really look forward to something exciting, you are very clearly very committed and very focused for the amount of production you do produce but a little bit of sparkle would really set you apart.



What do you wish to see in the future?

[moomooish]

There’s always something to worry about Betty 😂

[ivancito]

Wasn't GPT5 supposed to come out in June? People keep hyping up 5.0, and while 4.0 has gotten more polished it's not the promised near AGI level AI we keep hearing about.

[yamusa]

I've never heard that word to be pronounced like that: comaing? What?

[alexp]

Openai and Claude earn barely anything right now and spend more. So they are forced to say something like this to get their billions of investment.