So OpenAI just announced AND released GPT-4o to the public this Monday.
If you haven’t watched the announcement or read the news, here’s what was announced (written by me, not an AI):
- GPT-4o, new model, “o” stands for omni. Kinda weird name IMO, most people will get confused with “four-point-oh” or “forty”. Whatever, not my problem 😄
- GPT-4o and 4 features are now available on FREE accounts, including browsing, dall-e, code interpreter, AND YES: custom GPTs too!
- Current estimates state that there are 100M+ total users, but only 1-4M of them are Plus or Team users. Meaning that 99M people will be knocked out by GPT-4 capabilities in the next months. All custom GPTs can expect an uptick of 50x users. Our potential customer base also grew by 50x overnight, because most of our customers were subscribers and used GPT-4. Excited, and also stressed about it. Will we be able to piggyback on this big jump? Will see. Anyways.
- GPT-4o is a lot faster, almost real time conversation latency, even in the voice and vision modes. This means faster turnaround times for responses and new use cases that were impossible because of the latency.
- GPT-4o voice model can also sing, laugh, and change its voice to robotic, dramatic, etc. ➔ New creative opportunities.
- GPT-4o can also detect things from your voice, like if you are sad or anxious ➔ New opportunities for therapeutic use!
- GPT-4o can generate lot better looking typography. Text you can actually read.
- GPT-4o can use your mobile camera and give you real-time video guidance on the things you see.
- and now the biggest thing in my opinion:
- With the desktop app (coming soon), GPT-4o can WATCH YOUR SCREEN as you work! This means that a well configured custom GPT could help you write prompts, build AI automations, respond to emails, etc.
I think this is the real life changer.
This is the level of seamless integration that we need to truly have AI assistants.
I really encourage you to watch the full announcement here, or our reaction livestream + thoughts here
So let me explain why this will change your life.
Previously, voice chat with ChatGPT was slow. If I asked a question, it took 3-4 seconds to begin to answer. Now it’s less than 1 second (I don’t know the exact latency). Researchers state that 400-500ms is the threshold for human conversation.
Anything above that, you start wondering whether the other person heard you, whether they’re still there.
But with this low latency, and not the dumb GPT-3.5 level, but GPT-4 level of reasoning and capabilities, this is big.
Now we can have private teachers that are there with us 24/7, and know everything.
Now we can have private consultants that can help us 24/7.
Or AI girlfriends available 24/7. Every coin has two sides.
We are getting insanely close to an AI that can create AI agents.
Imagine this:
You start a video call with ChatGPT and share your screen. You say:
“Hey ChatGPT, I’m looking to automate this task that I do. Watch carefully, I’ll explain what I do, and then help me build an automation.”
It goes: “OK, show me what you got, fam.”
And then you just do your work.
It will see and understand that you are opening Excel, or ConvertKit or whatever software you have.
If you think out loud, it will understand your thinking.
And by the time you finish the task, it might give you an entire Make scenario blueprint that automates that task. Or give you suggestions on how to do it.
That’s the dream state. For now, I think it will probably be only able to help you put together the workflow documentation, but if you start building in in Make, it can also watch you and give you tips, help you debug, etc.
I have so many thoughts about how this will change people’s lives, that I could probably write a book about it.
But they’d all be just ideas. The problem is not that we don’t have ideas. The problem is that we can’t build all of our ideas.
Maybe you have 100 ideas per day, but you can only properly execute on one of them.
That will change.
Soon, you’ll be able to execute two of 100 ideas.
Then 3. Then 5. And before you notice, you are offloading your entire work to an AI so you can spend more time with your kids.
That’s the whole idea.
To stop working so much, and still get things done.
That’s the transformation our Prompt Master course students achieve. They either spend less time on the same tasks that they do, or they spend their time on higher level tasks.
Instead of writing emails to clients, they just review them.
Instead of creating content, they just manage the content system.
For example, one of our students, Shae could review hundreds of research documents relevant to her company in minutes using ChatGPT. Just like how According to researchers at Harvard, Wharton and MIT, if you’re skilled at using AI, you can do any knowledge based job 40% better and 30% faster.
Or how we like to call it: you can THINK FASTER.
Prompt Masters think faster. Become one by clicking here and enroll in our course today!
Talk soon,
Dave Talas