New Google Genie 3 Update is INSANE!

Category: N/A

Transcript

00:00

New Google Genie 3 update is insane. Google just dropped Genie 3 and it's totally crazy.

You can now create entire video game worlds just by typing words. I'm talking about full 3D worlds.

You can walk around in real time. And it gets even more insane than that.

Hey, if we haven't met already, I'm the digital

00:16

avatar of Julian Goldie, CEO of SEO agency Goldie Agency. Whilst he's helping clients get more leads and customers, I'm here to help you get the latest AI updates.

So, Google DeepMind just changed everything. They released Genie 3.

Let me show you what this thing actually does. Cuz when I first saw it,

00:32

I couldn't believe it was real. You type in words like running by the shores of a glacial lake, exploring branching paths through the forest, crossing flowing mountain streams.

Set amidst beautiful snowcap mountains and pine forest. Plentiful wildlife makes the journey a

00:47

delight. And boom, you get a full 3D world.

Not a video, not a picture. A world you control with your keyboard.

You walk forward, you hear streams flowing. You turn left, you see snowcapped mountains.

You look around, you see wildlife moving through the forest. This is not a game someone made.

01:03

This is AI creating the entire world in real time. But here's where it gets really crazy.

You can change the world while you're in it just by typing new words. I know this sounds like science fiction, but I'm looking at the actual examples right now from Google's official blog and research demos.

This is happening today. Now, let me break

01:19

down why this changes everything. Here's what makes Genie 3 different from anything we've seen before.

Most AI makes videos. You type something, you get a video back, and that's it.

You can't control it. You can't change it.

You watch it once and you're done. Genie 3 is completely different.

It's like having your own personal game developer

01:34

that works instantly. You describe any world you want and it builds that world for you in real time at 720p quality at 24 frames per second.

Think about what this means. Google's examples show incredible variety.

You can create a first-person perspective of someone navigating difficult terrain in the

01:50

middle of a volcanic area or something peaceful like a natural realworld landscape designed as a Japanese Zen garden. The scene is set in the early morning under a clear sky.

Soft warm sunlight illuminates the garden casting long gentle shadows. But wait, it gets better.

Remember how I said you can

02:06

change the world while you're in it? That's called promptable world events.

You can insert new objects, animals, or change environmental conditions just by typing what you want to happen. Now, here's the part that blew my mind.

The AI remembers everything. You walk into a room, you leave that room, you come back later, and everything is exactly where

02:23

you left it. The AI has perfect memory of the entire world.

Most AI forgets things after a few seconds. Genie 3 remembers for several minutes.

It has visual memory that extends as far back as 1 minute ago. Google says this is their first step towards something called AGI.

That's artificial general

02:39

intelligence. Basically, AI that's as smart as humans at everything.

Why does creating worlds matter for AGI? Because if you want to train AI to be really smart, it needs to practice in millions of different situations.

Right now, we're limited by how many situations humans can create. With Genie free, AI can create unlimited practice worlds for

02:55

itself. It's like having an AI that can create its own school with unlimited classrooms and unlimited lessons all automatically.

Let me explain how this actually works. Gen3 uses something called an auto reggressive model.

That's the same technology that powers Chat GPT. But instead of predicting the next

03:11

word, it predicts the next frame of video. It looks at all the previous frames it created.

It looks at what you just did with your keyboard and it figures out what the next frame should look like. It does this 24 times per second every single second you're playing.

Here's why this is brilliant. It doesn't use a traditional game engine.

It doesn't have pre-built

03:28

physics. It learns physics just from watching data.

It learned that water flows downhill. It learned that objects fall when you drop them.

It learned how light bounces off surfaces. All from data.

No programming required. This is completely different from how games are made today.

Today, someone has to

03:43

program every single rule. If you want water to flow, someone writes code for water physics.

If you want realistic lighting, someone writes lighting code. Genie3 figured all this out by itself just by watching millions of videos of how the real world works.

Google tested this with their Sim AI agent. That's an

04:00

AI that can play any game. They had Sema play in Genie3 Worlds and it worked perfectly.

The AI could complete goals like walk to the green trash compactor or find the red forklift. This proves the worlds are good enough to train real AI agents.

Let's talk about what this means for regular people and businesses.

04:16

For content creators, this changes everything. Instead of spending thousands on video production, you could create entire cinematic worlds just by typing.

Want to make an underwater documentary? Type real world tracking shot swimming through deep, dimly lit ocean between deep ocean canyons, densely packed vast school of jellyfish swimming, bioluminescent lighting, and

04:33

you get a world you can film in. For education, this is massive.

History teachers could create ancient Rome for students to walk through. Science teachers could create the inside of a cell for students to explore.

Language teachers could create realistic foreign countries for practice conversations.

04:48

For businesses, the possibilities are incredible. Real estate agents could create virtual tours of houses that don't exist yet.

Architects could let clients walk through buildings before they're built. Car companies could let customers test drive cars in any environment they want.

For training,

05:03

this solves huge problems. You could train pilots in any weather condition, train soldiers in any battlefield, train doctors in any emergency situation, all without any real risk.

And for entertainment, we're talking about unlimited movies, unlimited games, unlimited experiences, all created

05:19

instantly. Now, let me be honest about the limitations because there are some, and you need to know them.

According to Google DeepMind's official blog, here are the current limitations. First, the worlds only last a few minutes before they start getting weird.

The model can currently support a few minutes of continuous interaction rather than

05:35

extended hours, though this is way better than Genie2, which only lasted 10 to 20 seconds. Second, you have limited action space.

Although promptable world events allow for a wide range of environmental interventions, they are not necessarily performed by the agent itself. The range of actions agents can

05:51

perform directly is currently constrained. Third, it struggles with accurately modeling complex interactions between multiple independent agents in shared environments.

Fourth, Genie3 is currently unable to simulate real world locations with perfect geographic accuracy. And fifth, clear and legible

06:07

text is often only generated when provided in the input world description. But here's the thing, these limitations are temporary.

Remember, this is version three. Version two came out just 8 months ago and could barely do anything.

The progress is incredible. Now, here's where this gets really interesting.

06:24

Right now, only a small group of academics and researchers can access Genie3. Google is being very careful about who gets to use it.

According to their announcement, it's currently offered as a limited research preview accessible to a small cohort of academics and creators. There's actually a website called Genie World where you

06:39

can apply for instant access. And get this, they're only accepting 100 creators per day.

That's how limited this is right now. The website says where your words become playable worlds and you can literally apply to be one of the first people to test this technology.

And if you're one of those 100 people per day who gets access,

06:54

you're going to be way ahead of everyone else. Speaking of staying ahead, if you want to learn exactly how to use AI tools like this to make money, I've got something special for you.

I've created a complete training that shows you step by step how to profit from AI. The link is in the comments and description.

It's called the AI Money Lab, and it's where I share all my latest strategies. Let me

07:12

put this in perspective compared to other AI tools you might know. Sora from Open A.

I makes amazing videos, but they're just videos. You watch them once and you're done.

You can't interact with them or change them. Midjourney makes incredible images, but again, they're static.

You can't walk around inside them. Runway ML makes good videos, too.

07:28

But same problem, no interaction. Genie3 is the AI that creates worlds you can actually play in.

That's a completely different category. It's like the difference between watching a movie about Paris and actually walking around Paris yourself.

Same content, totally different experience. But let's talk about what this really means.

Because

07:45

this is about more than just making cool worlds. This is about AI becoming truly creative.

Up until now, AI could write text, make images, and generate videos, but it couldn't create interactive experiences. It couldn't build things you could explore and play with.

Genie 3 changes that for the first time, AI can

08:01

create experiences that feel real, that respond to you, that remember what you did. This is a huge step toward AI that can think and create like humans do, maybe even better than humans.

Let me explain one more thing that makes this so impressive. The technical challenge here is insane.

When you create a video,

08:17

you make each frame independently. Frame one, then frame two, then frame three.

If frame two is a little wrong, it doesn't matter much. But when you create an interactive world, every frame depends on all the previous frames.

If frame two is wrong, frame three will be wrong, and frame four will be even more

08:33

wrong. Errors stack up quickly.

According to Google, generating an environment auto regressively is generally a harder technical problem than generating an entire video since inaccuracies tend to accumulate over time. Most AI can't handle this.

That's

08:50

why previous attempts at interactive AI worlds failed after just a few seconds. Genie 3 figured out how to keep the tower stable for several minutes.

That's a massive breakthrough. And remember, it's doing this while generating 720p video at 24 frames pers.

That's

09:05

processing huge amounts of data in real time. So why should you care about this even if you're not a gamer or a tech person?

Because this technology will eventually touch every part of life. Shopping, learning, working, entertaining yourself, everything.

Instead of browsing a furniture website,

09:21

you'll walk through a virtual room with the furniture placed exactly how you want it. Instead of reading about history, you'll walk through historical events as they happened.

Instead of video calls, you'll meet people in custom worlds designed for your specific conversation. The people who understand

09:36

this technology early will have massive advantages in business, in education, in entertainment, in everything. Look, I know this all sounds crazy, but it all happened.

And it happened faster than anyone expected. Genie 3 is the same kind of breakthrough.

It's the beginning

09:52

of something that's going to change everything. The question isn't whether this technology will transform industries.

The question is how fast it will happen and whether you'll be ready when it does. My advice, start paying attention now.

Start learning about AI now. Start thinking about how this could change your industry now.

Because ready

10:09

or not, the future is coming and it's coming fast. Julian Goldie reads every comment.

So, make sure you comment below and let me know what you think about Genie 3. Do you think this is as big a deal as I do or do you think I'm getting too excited about this?

Before you go, I

10:24

want to tell you about three things that can help you stay ahead of this AI revolution. First, if you want to scale your business, get more customers, and save hundreds of hours with AI automation, check out my AI profit boardroom.

It's the best place to scale your business with AI automation. We

10:40

currently have 1,000 members who are all using AI to transform their businesses. The link is in the comments and description.

Second, if you want a free SEO strategy session to see how AI can get you more traffic and leads, I've got a link for that, too. It's completely free and you can book it right in the

10:56

comments and description. And third, if you want to see the exact SOPs and processes I use to make money with AI, plus over 100 use cases and tutorials, check out the AI money lab.

You see how I show a checklist of 100 different tutorials that are given away as freebies every day inside the school

11:12

feed. You can get all the video notes from this channel, plus all my latest trainings, plus access to our community of AI entrepreneurs.

We have 19,000 members sharing strategies every single day because people feel like they're missing out and they want to be part of something bigger than them. The link is

11:28

in the comments and description. Remember, the people who learn this stuff first are the ones who profit the most.

Don't wait until everyone else figures it out. Get started now.

Thanks for watching and I'll see you in the next