ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)

Noitom Robotics released the HiPHI-Hand dataset on November 9 at the CoRL 2026 conference in Austin, Texas. The collection contains 11 hours of egocentric stereo video with marker-based hand ground truth totaling 69.5GB. NVIDIA used this data to train its EgoHand model, which runs at about 9 milliseconds per frame and achieves state-of-the-art zero-shot accuracy on public benchmarks.
Tap to vote and see what everyone thinks.
Summary by ByteBrief