Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/PEOPLE/CHENG CHI
// PERSON

Cheng Chi

ROLE RESEARCHERMENTIONS 2LAST SEEN JUNE 12, 2026
// BIO

Researcher at Columbia/Stanford and lead author of the Diffusion Policy paper (RSS 2023), foundational to robot learning with diffusion models.

// RECENT MENTIONS
// SIGNALS
2 SIGNALS
01
mention·arXiv Physical AI·JUNE 12, 2026

Chi et al.'s Universal Manipulation Interface [10] is the foundational prior work that HyVLA-0.5 builds upon and substantially extends — from SLAM-based to optical motion-capture tracking, and from hundreds to 10,000+ hours. Chi's ongoing work (DexUMI [48]) on morphological extension of UMI rigs is also cited.

Source
02
mention·arXiv Physical AI·MAY 21, 2026

Cheng Chi et al. (Columbia/Stanford) — Not authors but foundational: their Diffusion Policy paper (RSS 2023) is the starting point for all experimental comparisons.

Source

AI-extracted from podcast / newsletter / paper summaries. May contain errors.