Open research groups can’t out-scale the frontier labs. So they're picking problems where exploration can beat exploitation: architectures beyond quadratic attention, on-device+cloud hybrid AI, interpretability to improve world models, an AI for science benchmark, etc.
Introducing Slingshots // THREE: 13 projects from the top researchers and top institutions in US and Canada. Alt architectures, world models, continual learning, hard sciences, HCI, and more. Let's meet the batch 🧵
