📚
@saturdayrobotic Robotics & World Models Reading Club 12:
@QuantingX7410 on Dexterity — 06/13
👉🏻RSVP:
luma.com/5w7c1t2a
Keynote:
@DanielXieee (Co-Founder,
@QuantingX7410 YC W26): Dexterity Benchmark We Need
2026 is the year of dexterity claims — frontier labs advertise “human-level” and “dexterity-first” manipulation, yet this spring alone three separate benchmark initiatives launched from industry, academia, and standards bodies. None are comparable. Definitions remain vague, taxonomies cover only narrow slices of manipulation, and demo reels stand in for protocols. The rigorous verification and reproducibility machinery that worked for vision, NLP, and even 1940s occupational therapy has yet to arrive in robot manipulation. This talk traces a century of attempts — from the Purdue Pegboard to today’s fragmented benchmarks — and argues that every piece of a proper dexterity benchmark already exists, just scattered across communities that rarely talk. Highly interactive: bring your own definition of dexterity and we’ll see whether the room converges any better than the field has.
Pre-Readings
Definitions & taxonomies:
Napier, The Prehensile Movements of the Human Hand, JBJS 1956 Elliott & Connolly, A Classification of Manipulative Hand Movements, Dev. Med. Child Neurol. 1984 Cutkosky, On Grasp Choice, Grasp Models, and the Design of Hands, IEEE T-RA 1989 Ma & Dollar, On Dexterity and Dexterous Manipulation, ICAR 2011 Bullock et al., A Hand-Centric Classification of Human and Robot Dexterous Manipulation, IEEE ToH 2013 Dafle et al., Extrinsic Dexterity: In-Hand Manipulation with External Forces, ICRA 2014 Feix et al., The GRASP Taxonomy of Human Grasp Types, IEEE THMS 2016
Human dexterity assessment:
Tiffin & Asher, The Purdue Pegboard, J. Applied Psychology 1948 Mathiowetz et al., Box and Block Test, AJOT 1985 Light et al., SHAP: Southampton Hand Assessment Procedure, Arch. PM&R 2002
Robot-hand dexterity benchmarks:
Zhou et al., 50 Hand Dexterity Benchmarks (HD-marks), 2020 Coulson et al., The Elliott and Connolly Benchmark, IEEE-RAS Humanoids 2021 Elangovan et al., Modular Dexterity Test Board, 2022 Liconti, Zhou, et al., POMDAR: A Benchmark of Dexterity for Anthropomorphic Robotic Hands, arXiv:2604.09294, 2026
Task suites & object kits:
Calli et al., YCB Object and Model Set, ICAR 2015 Kimble et al., Benchmarking Protocols for Small Parts Robotic Assembly (NIST task boards), IEEE RA-L 2020 Heo et al., FurnitureBench, RSS 2023 Luo et al., FMB: A Functional Manipulation Benchmark, IJRR 2024 Liu et al., LIBERO, NeurIPS 2023; Nasiriany et al., RoboCasa, RSS 2024
Evaluation methodology & infrastructure:
Li et al., SimplerEnv: Evaluating Real-World Robot Manipulation Policies in Simulation, CoRL 2024 Zhou et al., AutoEval: Autonomous Evaluation of Generalist Robot Policies in the Real World, arXiv:2503.24278, 2025 Atreya, Pertsch, et al., RoboArena: Distributed Real-World Evaluation of Generalist Robot Policies, CoRL 2025 Agia et al., CUPID: Curating Data your Robot Loves with Influence Functions, CoRL 2025 Chen, Kimble, et al., ManipulationNet, arXiv:2603.04363, 2026
Location: San Francisco (Downtown) (tentative)
Time: Saturday, June 13, 2026 | 2:00 PM – 5:00 PM
Hosts:
@junfanzhu98,
@aurorafeng_01
Agenda
2:00 PM — Doors open & social 🍓 Unlimited strawberries (official Reading Club fruit!)
2:30 PM — Keynote by
@DanielXieee (
@QuantingX7410)
4:00 PM — Q&A open-floor roundtable (10–20 min per topic; spotlight any paper you’d like to highlight)
Come ready to discuss what “dexterity” actually means, how to build rigorous and comparable benchmarks, hand-centric taxonomies, robot manipulation evaluation, and the missing reproducibility layer for embodied AI!
Past sessions brought together researchers & engineers from Boston Dynamics, Google DeepMind, NVIDIA, Stanford, UC Berkeley, Dyna, Physical Intelligence, Tesla, Generalist, Rhoda AI, and leading Bay Area robotics startups.
👉🏻RSVP:
luma.com/5w7c1t2a
#Robotics #WorldModels #EmbodiedAI #Dexterity #RobotManipulation #SFTech
