Luda
On building a full-stack AI Lab focused on delightful consumer experiences & our investment in Luda
I met Vijay Sundaram many years ago in San Francisco. We embarked on a long meandering walk throughout Russian Hill in which we talked about the desire to find increasingly weird and technically ambitious companies. Over the years we’ve invested together, hosted dinners surrounding a variety of frontier tech areas and even recorded random (and now embarrassing) conversations on AI & Generative Celebrities and Holograms & 3D.
In all of these conversations, Vijay had a nagging feeling that the ambition and joy in deeply technical products was lacking. He tried to satisfy this feeling by partnering with those trying to bring this intersection to life as a VC and angel investor, until he eventually joined Google Research to focus on AI and gaming at a time when many still believed AI was only going to impact super non-creative workflows.
After creating groundbreaking interactive AI systems at Google, Vijay and I spoke through the pandemic and sat for multiple months in 2021 talking through his vision for what a full-stack AI lab focusing on the next-generation of consumer digital experiences could be.
At Compound, we often talk about technical inflection points and how applying breakthroughs that progress from one domain into another allows for multiple very different companies to be built around similar premises. Vijay’s vision was the perfect embodiment of this. As we talked through the research his team had done, the long path we had seen with what is now Generative AI (and what we were calling Computational Creativity for years prior) with Runway and Shadows1, as well as what breakthroughs we had seen in simulation and reinforcement learning from AI Reverie and Wayve as it related to embodied intelligence, we became increasingly excited about Vijay’s idea for a new company.
The premise for Vijay’s company crystallized years of these conversations with a simple yet perfectly weird vision; AI would enable novel consumer experiences beyond our current imagination, and these advances in reinforcement learning, large language models, and browser-based training would let Luda build joyous products at the scale Vijay had always dreamed of.
Vijay co-founded Luda, a full-stack AI lab in 2021, to build delightful interactive consumer-focused “agents”. Powered by proprietary models, native infrastructure and breakthroughs like real-time reinforcement learning, these agents are a fundamentally new "AI-native" creative medium.
Luda’s early agents are called Mels, built from composable Lego-like parts that users can easily shape to bring their imaginations to life. To start, this manifests itself in a sandbox game environment, with much more to come over time as the Luda platform expands to allow users to create a variety of games, stories, and social experiences as well.
We were thrilled to co-lead Luda’s Seed round years back and support the team’s journey into pioneering playful AI. We couldn’t be more excited to see the team start to talk about what they’ve been up to and to continue to execute on their long-term vision of creating new experiences in consumer and perhaps entirely new categories.
Vijay was an angel investor in Shadows