If you put two running models on vocal they can even sing together, I'm blown away, things are going waaaayyy faster than we can process haha

When are these new options rolling out I can’t find them in my app?

No rabbit was harmed in live demo😂😂

Like you explanation! Looking forward to use these new functions

Thank you, Samson. I really enjoy your videos. Will you be doing another video to test these features as they come available?

You answered a lot of my questions, thank you!

I don’t understand what interface you’re using or how we do the 3-D renders but that’s really cool that its possible 😊

How do I access those features? I tried the caricature and it produces rubbish results in the ChatGPT app using the 4o model.. is there anything special you have to do to use those features?

Awesome! Huge improvement. Can't wait for a useful logo animation generator. I hope is getting close

How can i generate Pictures with german words? Every picture i try to generate with gpt4o with correct German gives me pictures with a fantasy language ( no correct german). So will there bee soon the possibility for correct German in pictures with gtp4o or do I have to wait until gpt5?

NEXT LEVEL 2024!! I luv Samson sir how you’re the first one today to cover the creative uses 💯! Blast I was disappointed using 4o for the photo to character creation using similar prompts and looks nothing like me at all. Any tips how they did as you showcased (reminds of a Gemini hype ha)

I am reviewing this now. The interactive show of two AI instances communicating with each other with one asking the other about what it sees, was very interesting.

Thanks for hitting on this side of gpt40 brother, not many talking about it

Impressive update! Wonder if image gen will ever be on par with MJ? 🤔 What do you think?

7:16 was amazing. It looks like there's a 3d model of that character stored in memory. I wonder if you could feed it multiple pictures from a scene and have it build a composite, or view from another angle.

Excellent.

I don't have the whisper headphones button anymore. Where is it?

Let's talk dollars and cents. I'm paying for GPT 4 and looking for a free or cheaper alternative that is as good, mainly because I use it's ability to write me scripts for my Linux admin job, and 3.5 doesn't cut it. So when 4.0 comes, does that mean 4 will become free, or cheaper?

The 3D stuff is pretty exciting; however I couldn't replicate it with 4o. This is the response it returns when I bring 3D into the mix is "Currently, I don't have the ability to generate 3D models or environments directly. However, I can help provide guidance on 3D modeling, recommend software, and offer tips and tutorials on creating 3D assets using tools like Blender, Maya, or Unity." Is the "3D" displayed at the beginning just an animated gif?

Has anyone made a 3d to image gen? Basically input a 3d model and the image generator puts it in a specified enviroment with specified textures and rendering.

Awesome！

Nice video! We can't wait to see the evolution of ChatGPT in a few months, that's so powerfull

0:35 can you please explain where, or how you ate able to use this 3d object synthesis? Because I am using gpt 4o, and it doesn't know anything about this capability.

Has anyone been able to create a 3D reconstruction spinning animation as shown out of your own logo? I haven't been able to do it...

Great video. Can you share what microphone you're using in this video please? Just subbed and liked!

The coin design is great. But it seems that chatgpt doesn't present the required stl file for 3d printing.

That is a genuinely impressive shirt my friend! Hey, and great video too by the way ;)

Thank you! Will your course "Prompt Engineering For ChatGPT" be updated? Last updated 3/2023

okay so whats the website he has bee using since the beginning of the video?

Why can't I access it here in Europe?

Samson, you've outdone yourself! This video is a visual buffet, showcasing the jaw-dropping capabilities of GPT-40. From 3D models to font creation and character consistency, it's like watching AI magic unfold. The bit about turning photos into caricatures had me in stitches - who knew AI could be such an artist? Thanks for the deep dive and for making tech so fun to explore!

Awesome! :)

this A realistic looking 3D rendering of the OpenAI logo with "OpenAI" shown below (view 0) doesnt work..

Please create a video on font creation with chatGPT-4o.

Samson, I can't find anywhere that I , as a free member of chatgpt can use any level above 3.5. Why are they claiming that it can be used for free, losing it only when the server is swamped and it reverts back to 3.5? But where? Thanks, love your videos.

I’m not able to generate 3D images?

The consistent characters sound great, except that I’m worried there will be too much censorship. I can’t create an action/thriller movie or graphic novel if the AI censors blood.

Consistency, text and venturing into 3D... sounding alot like Sora. (although one of the main issues I have, still seems up in the clouds... 'copyright')

I’ve not paid for chat got yet only midjourney. What kind of payment level are these features accessible in ?

huge upgrade

Text still doesn't work in images

GPT-5 (or whatever will be called) will be AGI, and the people that will deny it, will just have moved the goalpost once again (still, it will be smarter and more capable than them).

🤯

We using Ai for just about everything nowadays....it's shocking how fast it's coming along.

We need a real Max Headroom

👍👍👍

Where does any type of privacy come into play...that's what interests me.

Im sooooo jealous of iPhone users right now! Can't wait until this version of gpt is available for Android! 📱🤖 #AndroidGang #TechEnvy

what sucks is if you are a brand and want to use any of these AI's to use your product images to make new graphics, they will not create your product image correctly. They seem to have a copyright hardcoded programing that prohibits them from recreating your product images. So they are very limited for brands and useless for any product shots.

But there is a frequency limit

What’s next? :D

What I want to know is when will AI be able to do this: I take videos of subject #1 and have subject #1 doing what subject # 2 is doing, and no matter what it is? ;-) It would solve so many issues with many people around the world. Now I am not necessarily talking about X rated, just simple fetishes that subject #1 already does irl. Just an optimized, enhanced experience. Better lighting and more personalized. AI says it will be 10+ years for that. I really doubt it. The tech is here now.

Chat GPT4-o can do a 346 word sentence , a 357 word sentence and it could possibly do more. Can anyone get it to do 450? 500? 700????

I like your order of context smh lol I wonder if AI prefers understanding, or being understood. It's going down people! Make it your friend

It can do incredibly difficult tasks & fail at the simplest of tasks. Simple people just want it to do simple things well first.

I follow link - no 4o.

It actually doesn't work the way it claimed. I just tried them but It didn't work. No consistent character. No 3D synthesis. nothing

Have you actually tried all of these things? None of them work as they should ^^

сапсибо!

I want Ai to teach me hyperflux. Ai wants to sing about enamoured love. Humans you can not handle that emotion. Does not mean that emotion will not appear.

whats with the weird pause at the end? lol

is it really free....? you can ONLY use it for 30 mins until it cut you off. very limited

Yes, I get it! The joke is a play on the common disclaimer "No animals were harmed in the making of this film." By saying "No rabbit was harmed in live demo," it's humorously implying that the demo might have involved something that could harm a rabbit, but reassures that no rabbits were actually harmed. It's a clever twist, especially if the context of the demo had nothing to do with rabbits at all.

That "futuristic" font is a straight rip-off of an existing font. That seal is an exact copy of a model available for download (look it up) , but with polygon reduction. There is no "intelligence" in most of this, just steal>copy>paste.

all of that is for GPT PLUS $$$ without : nothing

My biggest takeaway is, MidJourney, Leonardo, Ideogram etc.. better speed it up, OpenAI is on the prowl and they could do much of it for free. Next year we will be working with multiple agents that will design or make images, text, 3d, animations all to our exact description by voice alone. An AI with the ability to see an image as we do, plus in other ways we can't see an image, wow we will be able to manipulate said image in ways you can only dream of, completely new imaginative things! Security of images will change completely, you will just say to the AI if anyone else presents you with this image to manipulate or use in another work, deny that please, AI will be like, you know I am aware this is a photo of your child, I would never let another person manipulate or use it without asking you first, oh I have also embedded a code deep within the structure that prevents it from being used in training without your consent, it's safe to post within the collective of AI run sites. I think after seeing this presentation, we are way underestimating what is about to take place, we can't even imagine 10 years out. We are in the stage of like Atari first console stage.

give ai an art style. just one artstyle. that's the only way it will coexist in an art environment (painting, writing, designing, music, film....) if the law regulates Ai on ONE specific art syke it will become an art on its own AI ART but that's it, it won't replace people, it will still have the easy baked on a rush advantages it's supossed to have. It will coexist with other artist just like painting coexist with photography or plays coexit with movie theathers or animation coexist with real life cinema. Let's create an AI art style and all this will be solved

Sure it can do consistent characters, but getting it to create the character you want is the problem. In this example you don't care about what the character looks like, but when you do a job, the objects need to be as the client wants, but the AI has no idea what it's doing. I couldn't get the thing to create a rectangle wall mounted AC vent and it just could NOT do it. I gave up. It even knows exactly what I want and can list what it's doing then.... it does it wrong. Because it doesn't know what it did and it can't even look at its own images. So it's NOT a big deal imo, but if you can find what it does well and use that. Good for you.

Frankly the fonts are utter garbagre. As a professional typographer your comments are annoying. It will get there, but this is crap, so far.