10.5.23-Showcase-Tokyo-Common_Sense_Machines

DHRUPAD KARWA: [SPEAKING JAPANESE] My name is Dhru. And I'm the product lead at Common Sense Machines, also known as CSM. Our company was founded by Tejas Kulkarni and Max Kleiman-Weiner, who were both computer science PhDs at MIT.

So the problem we're solving relates to 3D world-building. Creating 3D assets is a really, really difficult process. To produce a single 3D asset, it often costs thousands of dollars, and can take days and weeks of collaboration across different teams. And because of this complexity and prohibitive cost and timing, it means that it becomes very difficult for companies to build large and diverse 3D data sets.

Take Shutterstock, for example, one of the largest content marketplaces in the world. It has just under a million 3D assets listed on its website, which is nothing compared to its number of stock photos and videos. So it's really high time that we solve this.

Meet Cube by CSM, an AI product that lets anyone anywhere, regardless of skill level, generate 3D assets from simple prompts like a line of text, a single image, or a short video. It works on any platform, web, mobile, Discord, and also through APIs. And I think the real beauty and power of Cube lies in its simplicity.

It works in just three clicks. Click one, you enter a prompt. For example, this cool looking sheep in a denim jacket. Click two, you hit a generate button, and then let our AI work its magic. And click three, you export your model in a file format of your choice, and go on to build your project.

This entire generation process today takes around two hours. Not two days or two weeks like the traditional 3D modeling. And we expect this time to go down drastically as our team continues to make improvements to our core models, both in terms of speed and quality.

Here's another example, this time showing a video prompt taken through our mobile app. You can see how Cube has transformed this humanoid object into something a lot more lifelike that's really ready to enter reality. This character could become a feature in an action game. It could become a protagonist in Marvel's next superhero franchise. Or it could even be 3D printed and turned into a lifelike robot filled with intelligence and common sense. The possibilities are truly, truly exciting.

Now in terms of quality, Cube is really state-of-the-art compared to any other company out there at the moment. As you can see, our outputs are much more vibrant than the majority of other products, including the mighty OpenAI and their 3D Shape-E model, which was trained on millions of proprietary data points.

Now returning to the theme of robots, I wanted to share a recent use case that came out of Tokyo right here. So in August, a game development studio-- an artist at a game development studio turned to Cube, and used it to generate 3D characters based on single images. And then he animated those characters to music using programs like Mixamo and Unity, and shared these videos and these dance videos online. And they went really viral on Twitter.

And I wanted to share one with you, which was a cute little robot dancing. Oh, what a shame, this worked yesterday. [LAUGHING] There we go.

[MUSIC PLAYING]

I'm actually having breakfast with this artist tomorrow morning because he wants to turn it into an entire movie and a sequence of different dance videos, including a K-pop number. So that's going to be really fun, especially if we go to Korea on the weekend. What's cool is he actually typed up his entire workflow and published it on his personal blog. And I feel, especially coming from the world of consumer products, that early inspiration and energy like that is often a lead indicator of great things to come.

So it's difficult to believe that we launched just a couple of months ago now. And we've already scaled to over 50,000 users without a penny of marketing spent. Our cohorts in Korea and Japan are particularly exciting. They're growing 20% week over week, with users performing thousands of interactions on Cube every single day. And the average Japanese and Korean user generates 2.6 3D assets every 24 hours.

And even more interestingly, those 2.6 assets are almost always characters. Not necessarily trees or anything in a landscape, but always characters and anime. So that's a really interesting behavioral insight for us.

And we're seeing a lot of applications emerge across different industries with 3D workflows like gaming, animation, simulation, and 3D printing. And we have enterprise plans as well as APIs ready to support these applications and help take these industries into a really beautiful 3D-generated future. So let's partner, build POCs, and create new worlds together. Thank you so much.