What is Apple up to?

Apple just one-upped Google and OpenAI…

Apple just published some incredible research around its GAUDI artificial intelligence (AI).

Technically speaking, GAUDI is a neural network that creates 3D representations of text. But rather than just a text-to-image generator, GAUDI is capable of incorporating verbs/actions in a way that creates 3D environments.

We have explored text-to-image technology quite a bit in The Bleeding Edge. We recently had a look at Google's text-to-image AI in June. And we talked about OpenAI's DALL-E back in April and again last month.

If we remember, both can create incredibly lifelike images just based on text input.

For example, here's the image that Google's AI created from the following prompt: "A photo of a raccoon wearing an astronaut helmet, looking out of the window at night."

AI-Created Graphics

Source: Google

It's incredible to see the AI produce such a creative, lifelike image just from text.

But Apple's GAUDI takes it to a whole new level. Instead of 2D images, GAUDI generates 3D videos based on text.

Here's a look:

GAUDI's 3D Capabilities

Source: Arxiv

Here we can see Apple's AI generating video based on the prompts "Go through the hallway" and "Go up the stairs." This is happening in real time. What a powerful generative engine.

And it doesn't take too much imagination to realize what Apple's up to with this.

As we know, Apple is gearing up for a big augmented reality (AR)/virtual reality (VR) launch. We can expect more information about Apple's AR/VR headset next month. There's a chance that we'll hear some news at Apple's major annual iPhone product release event currently scheduled for September 7.

If we think about the process of building 3D virtual worlds for immersive AR/VR content, historically it's taken years and millions of dollars just to produce one game or application.

Well, a text-to-video generator like GAUDI can largely automate the process. This is an incredible tool that could crank out content for virtual worlds extremely quickly – and cheaply.

So this research shows us that Apple intends to be a major player when it comes to producing programming and content for next-generation games and applications in virtual worlds.

This is right up Apple's alley.

Jeff Brown