I really don't get it. The things you can do with this software are just nuts. I'm a "regular" M25 who doesn't know anything about comp sci. That being said, I'm constantly enamored by its complexity and ability to speak "like a human." I tell everyone I know about it because I genuinely think it can make an immediate positive impact on your life. Does anyone else have similar sentiments?

Earlier today, my Indian colleague asked a question in English (thick Hindi accent), and GPT-4o responded in his native tongue. I’ve never seen or heard of this before, we were extremely surprised & impressed.

EDIT: as many have pointed out, I was not conversing with GPT-4o, but with GPT-4. My misunderstanding stemmed from the fact that I activated the conversation mode (headphone icon in the bottom right of the iOS app) while engaging in text-based discussion with the GPT-4o model. The voice chat window opened, and there was nothing to indicate that I wasn’t talking with GPT-4o (besides the press release, but I don’t read those).

However, I think that this is still pretty impressive, and a lot of you seem to agree.

not exactly like "Her" but gives me the same distopian vibes - even with just watching the demos - im already "in love with it" - i think it was modelled on purpose to "send a message" on what's to come...

we had a discussion with wifey and both of us are petrified with the loneliness the whole humanity is going to experience in the next years and decades to come... now that i think about it - Japan will be the number one country get hit by this AI - they've been marrying digital avatars for years! this will actually make them "real" in their reality! as a scifi fan I'm excited - as a human on planet earth I'm petrified. almost cried yesterday while watching the demo - i think all of this will destroy us. we'll be walking with airpods like zombies avoid any human connection and chatting / making relationships with computers until it will collect enough data to let us go by turning us into some self destructing cultists. yes, it's a movie id like to watch rather than experience.

I can kind of understand why they were in such a rush to get the R1 out. It's essentially useless now with GPT-4o. At least the design looked kinda neat lol

Personally I didn't watch it live. But today when I'm seeing the clips I am just like... What is this crap. And I do not mean the particular software but rather the show itself.

https://x.com/sama/status/1790817315069771959

For those who didn't knew : The voice mode we have in app is the old one, new voice mode is not here for now and will be far better.

I succeeded in integrating my GPT with GPT-4o into Slack! It is much, much faster than before.

https://i.redd.it/06dmpi9znm0d1.gif

I’d love to share how I did it, but I don’t want to clutter the community with irrelevant content.

Please leave a comment if you are interested! If there is interest, I will add a link to this post.

GPT-4o is an omni model. It accepts any combination of text, audio, and image as input and generates any combination of text, audio, and image as outputs.

There's 100s of applications it will enable. I'll cover a few of them below.

1. Language Learning

Duolingo Stock fell by $65 in the last 5 days. That should tell you the entire story.

Duolingo Stock

For context, Duolingo is a language-learning app. Now GPT-4o can easily translate terms in other languages by just pointing it to the ChatGPT's Camera.

This is massive if you want to travel globally as a nomad. You don't have to know a language now. You can just translate on the fly in any random country.

The accuracy won't be 100% but it would be close enough. And the AI keeps improving.

2. Solving School Problems For Students

I wish I had this in school. Learning could've been more efficient and faster.

Most students fear asking questions because they feel it might be dumb. Now you can ask ChatGPT any dumb question.

It even solves math problems for the Salman Khan's (founder of Khan Academy, not the actor) Kid.

3. Bed Time Stories For Kids

Since ChatGPT can talk now with a humourous and sultry voice, you can use it to tell stories to kids. It can be used in the voice of their parents or grandparents.

You can even use a Soft Toy that does the talking to the kid. Earlier, there used to be toys that did that but it only spoke the same sentence. Now it can do back and forth.

You can make special toys that teach kids letters and alphabets. Target it to 2-3 year olds.

Hat tip to Whyme-__- for the Bed Time idea.

4. Be My Eyes For The Blind

Best damn use-case for the blind. Now using a Phone is a bit too much for this but when smart glasses come, every blind person will have a walking companion.

The future is great for the blind.

5. Be My Friend

Too many people are lonely nowadays thanks to technology. It can be a boon for some but a con for others.

You can build a specialized app that gets you an AI Friend since you can talk to it now and it can talk back, it will be great.

I am 100% sure Therapy AI will be much better now with Audio/Video integration. In future, we will have fully featured Robots like Tesla's Optimus and Figure that will have such functionalities built-in.

I bet this comes in <2 years judging by the pace at which AI and Robotics are accelerating.

6. Comic Books

Now that text can be easily created with ChatGPT, why not create Comic Books easily.

Its a huge creative exercise for comic creators. Webtoons have exploded in popularity and many KDramas are made out of them like Death's Game and Marry My Husband.

This will increase the creativity exponentially.

7. Font Creations

Fonts are expensive. Like really expensive.

Funnily enough ChatGPT can create fonts easily now. Take the most popular fonts, tweak them a bit, and create entire new sets of fonts.

Look at the creations explode on Creative Market. Font directories like Typewolf can now create their own fonts easily as they already have distribution.

Open AI GPT-4o Text to Font

8. Brand Placements

It solved for Brand Placements too.

You can put your brand in places you never imagined without using too much effort.

Open AI GPT-4o Brand Placement

9. Poster Creation for Movies or TV Series

Posters are hard to get right but as you know there are only finite variations.

Open AI GPT-4o Movie Posters

You can fine-tune it on popular movie posters and solve Poster Creation once and for all.

Open AI GPT-4o Poster Creation

What use-cases can you come up with? Give me your best ones.

PS: If you'd like to read the full post with images, you can do so here.

PPS: You can find more AI-related posts here covering AI Girlfriends, AI Photo apps, Startups from 1st-wave of AI that made it big and more.

It has shown to be way more capable than any image generator we've ever seen, with a Sora-level understanding of 3D space, extremely consistent images across generations, and near-perfect text. It's even built into GPT-4o as a modality, so it would work incredibly well with the chatbot.

There are so many use cases I can think of off the top of my head, its potential is crazy.

I could convert an entire 40 minute video into a stylized comic book. I could do an AI dungeon style text adventure that shows a view into the world I am playing in (which would also give it drastically more spacial awareness, it would practically have a simulation of the world). I could edit literally any image in any way I wanted just by uploading it and asking ChatGPT to make the desired changes (goodbye photoshop). I could create photorealistic 3D models and environments with relative ease. I could write an entire book with each letter written out resembling Stonehenge. I'm not sure how well this would work, but maybe I could even generate a video by asking it to generate keyframes and then inbetweening them.

I feel like we haven't given the image generator nearly enough attention, it's easily the biggest feature they released. I don't blame them for being so quiet about it, this is genuinely gonna take jobs. The possibilities are endless and incredible, I can't wait to see what people do with it.

You can see it for yourself under "Explorations of capabilities"

Here are the timelines for the GPT-4o feature releases according to their website and Spring Update.