Flashback Friday. Today I'm taking it back to our very first demo, a MVP (minimum viable product) to show the possibilities of generative dialogue in gamified experience. At the time, we were raising our seed round and talking to lots of investors, so we thought it would be fun to turn the tables and put ourselves in the boss seat. Our goal was to prove that we could create an immersive player experience with no scripted dialogue for our characters.
We created an office where a team of entrepreneurs were getting ready to pitch MetaMuscle, their company. The executives at MetaMuscle shared common knowledge about their company, but had different underlying motivations.
Here’s one of our very first playtests of the experience.
The demo showed us that generative conversation was possible, even though we could still improve the voices, gestures, and overall experience. It took a small team a few weeks to build the experience, and the vast majority of the work was creating the 3D environments and visualizing the characters. We used Unity and Ready Player Me for the visuals, since Inworld is focused on the character’s brains and conversation. If I had scripted the dialogue myself, the process would’ve taken longer, or the gameplay would’ve been more limited.
Over time, we expanded beyond knowledge and shared lore. The characters we love in movies, books, and games have distinct personalities and ways of speaking. So we refined Alex’s character and made him arrogant. Check out the description we used to create Alex in Inworld Studio and the resulting conversation:
Alex is an ex-bodybuilder who saw some avatars in the Metaverse and thought they looked "skinny and wimpy" so he thought they needed to "beef up" and gain some muscle. His company's name is called "MetaMuscle" and he is extremely excited about it. He believes it will change the world and that the company is already worth over $1 Billion even though he hasn't sold a single gym membership yet. Alex has invested his life savings and is very eager for investment as he is running out of money.
We’re improving Inworld every day based on feedback from our community and we’d love to hear from you on Discord.
Optimizing machine learning models to achieve greater levels of efficiency can be a challenge. One of our engineers shares how he reduced CPU usage from 47% to 0.5% without significantly increasing latency.
Wondering what an AI character actually is? We break down how AI characters work, their potential use cases, and whether they're likely to try to take over the world. Spoiler: They're not!