TASTE
INTELLIGENCE
Scaling the AI Product Taste
tasteIntelligence is a community research lab dedicated to scaling the AI product taste. We bridge the gap between foundational AI agent research and real-world AI application innovation, creating a space where cutting-edge research meets practical implementation.
While AI capabilities have advanced dramatically, the gap between research breakthroughs and product excellence continues to widen. The knowledge of how to build truly exceptional AI products remains concentrated within a few organizations, limiting both public discourse and people's ability to leverage AI effectively. To bridge this gap, we're building tasteIntelligence to make AI product development more widely understood, customizable, and impactful.
Look at Manus, Doubao GUI Agent, Cowork, OpenClaw — the underlying tech stacks behind these products have existed long before they shipped. This proves something crucial: in the agentic era, product taste matters more than research or engineering prowess alone. Taste is the new intelligence. We see this unfolding across IP-enabled toys, AI glasses, otome games, elderly companionship — vertical agents that actually connect with users. We need more playful, delightful agents.
We are researchers, engineers, and builders who believe that the future of AI lies not just in scaling models, but in developing the taste to create products that genuinely serve human needs. Our work spans from foundational agent architectures to user-facing applications, always with an eye toward what makes AI truly valuable.
Research is better when shared
Scientific progress is a collective effort. We believe that we'll most effectively advance the field by openly sharing our research, collaborating with the broader community, and building tools that empower others to innovate.
Our Focus Areas
Foundation
Agentic systems, RL infra, context learning, continuous learning, self-evolve,datageneration,self-evolve evaluation,environment generation
Application
User experience design, product-market fit, and AI-native application patterns
Community
Open research collaboration, knowledge sharing, and developer empowerment