Sam Altman FINALLY Reveals GPT-5 (GPT-5 Explained)

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?

Music Used

LEMMiNO - Cipher
CC BY-SA 4.0
LEMMiNO - Encounters

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

[ElijaAltenhoner]

It feels like the AI News waited for you to return

[vaibhavwaykole]

Your content is always top notch. Lot to learn from you. Thanks for sharing the latest updates as always.

[MarkoKraguljac]

"Unification" for them probably means hidding even more what is happening behind the "Open" curtain. We wont even know which subsystem is handling what and when.

[greatwazzoo]

GPT-5 will be euphemistically known as “The Old Man in the Cave”, based on the same named Twilight Zone episode.

[icegiant]

What they need to fix is the chats getting into a non-responsive state. I use o1, o1 pro for coding, and after about 6-10 prompts, there is a 1 in 5 chance it will start thinking, and then just die. I can try again, wait another 2 minutes... no answer. I have to switch it to other models to get it to snap out of it. I have been having to start new chats for rather complex work flows, and its a massive PITA. And BTW, I think o1 and o1 pro are better than 03-mini-high. o1 can now take images, and o3 cannot. None of them can take a normal file. Give me o1, let me upload files, solve the hanging problem, and get out of my way. Thank you! BTW here is your $200 for the second month.

[TheChromePoet]

Thank you Deepseek for moving the needle.

[MatthiasRMoore-obni]

I missed your videos 🙏 So glad your back

[RLTW]

My wife and I use 4o constantly. At this point it knows our preferences quite nicely. It knows our dogs name and stats so we don’t have to say it all each time, it is functional. Whatever updates happen, that needs to be integrated. That is easy of use.

[とふこ]

Everybody (openai/claude) gangsta until deepseek V5 and qwen 3.5 max release.

[techbizcanada]

Having 5-7 options causes analysis paralysis? It's never been an issue for me.

I prefer the options. I want to have the control over what model I use.

The primary reason OpenAI and other companies want to move away from giving the user the option is to save money on compute by automatically routing your prompt to the version/AI they see fit.

I don't like that. I think it will probably cause MORE people to try different prompts in order to reach the right route, which may increase the need for compute, not lower it.

[suncat]

A question or statement might seem simplistic at first. It often requires high intelligence to even recognize that a question or statement has embedded within it a deeper meaning or humor. Automatically handing it off to a simple model for processing is not always the smartest thing to do.

[mehdihassan]

Sam said before that GPT 5 is smarter than him. But GPT 5 is coming in months? So by the end of 2025 if their coding agent would be number 1? If 2025 was AI agents, it makes sense that 2026 is innovators.

[isajoha]

The AI edge on top of evolution channel, well done. Great analysis also.

[CollegeWithCorey]

I can’t wait until they make this thing be able to call places for me on my behalf and schedule appointments, order pizza, etc.

[marrty]

I don't hate the picker. It lives on o3 mini high.

[pkfautchoisrunnom]

You basicly said "basicly" more than 10 times

[richardrombouts]

How will this influence pricing? Will we pay for the highest model, while our requests are relayed to the simplest model?

[mostababa]

I love the model picker. As a technical user, I would hate to have it dumbed down or auto-select a model for me which I don’t want to use

[danypell]

Would be nice to have not just Pro at $200 but even more power for $5, 000, $10, 000, $100, 000 / month etc. As long as more compute did mean a lot more intelligence for sure!

[DiceDecides]

It should have been simple from the start, GPT-3, GPT-4, GPT-5

that's what I expected at first tbh