(3/3) đź“„ Tasteply: Externalizing Stable and Situational Preference as Manipulable Spatial Objects
Lasting and in-the-moment preferences become spatial objects people can see, move, and adjust to steer how AI personalizes for them.

(3/3) đź“„ Tasteply: Externalizing Stable and Situational Preference as Manipulable Spatial Objects
Lasting and in-the-moment preferences become spatial objects people can see, move, and adjust to steer how AI personalizes for them.
(1/3) We're heading to Detroit for #UIST2026 (Nov 2-5) with 2 posters on interfaces that adapt to people đź§µ
→ CORAL: generative interfaces that keep learning from collective use
→ Tasteply: preferences you can see, move, and shape
New #UIST2026 paper led by @yuxuanli1225.bsky.social
WhatIf lets policymakers steer, inspect, and compare large-scale LLM-powered social simulations in real time. It helped emergency professionals recognize previously unrecognized planning vulnerabilities.