Google Deepmind’s Genie 2 might be able to create 3D worlds: Here’s what it means

Google Deepmind’s Genie 2 might be able to create 3D worlds: Here’s what it means
HIGHLIGHTS

DeepMind explained how users can easily generate detailed environments with simple prompts.

Genie 2 operates by processing user inputs through Imagen3, which is another generative model.

Genie 2 even has long-term memory which allows it to recall unseen parts of environments.

DeepMind, which is Google’s AI research arm, has launched Genie 2. It is an advanced AI model and is capable of generating infinite and interactive 3D worlds. This succeeds Genie, which transforms single images into playable environments. Now Genie 2 works on this itself and offers the creation of dynamic 3D virtual spaces from text or image prompts.

DeepMind explained how users can easily generate detailed environments with simple prompts, in a blog post. For example, typing “a warrior in snow” can produce a snowy battlefield where players can interact with objects and perform actions like jumping or swimming. All of this is governed by real-world physics and lighting effects. Further, it supports diverse perspectives, including first-person, isometric, and third-person views. With this, the virtual worlds are adaptable for various experiences.

Let’s take a look at how Genie 2 works.

How Genie 2 works

Genie 2 operates by processing user inputs through Imagen3, which is another generative model. It uses auto-regressive techniques to simulate environments frame by frame. Like if you press directional keys, it moves a robot character instead of unrelated objects like trees or clouds.

Genie 2 even has long-term memory which allows it to recall unseen parts of environments when you ask it for. This feature ensures that the generated worlds remain coherent and realistic over time.

It is not really a gaming platform and has been created as a creative and research tool. It can be used in video game design, virtual training simulations, and beyond. It can even transform concept art into interactive spaces as it has out-of-distribution generalisation capabilities.3D content creation will become a lot more seamless with this, whenever it is rolled out fully.

Mustafa Khan

Mustafa Khan

Mustafa is new on the block and is a tech geek who is currently working with Digit as a News Writer. He tests the new gadgets that come on board and writes for the news desk. He has found his way with words and you can count on him when in need of tech advice. No judgement. He is based out of Delhi, he’s your person for good photos, good food recommendations, and to know about anything GenZ. View Full Profile

Digit.in
Logo
Digit.in
Logo