Reward AI

Frontier Intelligence for Any Robot

[Mengyue Yang @Mengyue_Yang_]

Efficiency will be the next trending?

[Ilir Aliu @IlirAliu_]

.@RewardAI_ came out of stealth yesterday with OM-1, a robot foundation model trained entirely from human demonstrations.

The part I find interesting is how they collect the data.

Instead of teleoperating a robot, a person wears a device called the Omnibody Hand and simply does the task themselves.

Pick something up, fold laundry, mix a drink, package a phone, or plug and unplug cables.

The system records the human hand motion, tactile information, proximity, vision and position. That data can then be used to train OM-1 to control different robot bodies.

Reward calls the idea:

“One Model, One Data Interface, Any Body.”

According to the company, a new task can be learned from less than 30 minutes of human demonstration data, without collecting additional robot data.

The videos are worth watching.

They show four robot arms packaging a phone, robots folding laundry and mixing cocktails, an arm unplugging an Ethernet cable, and early experiments with humanoids.

Some of the longer tasks finish in under 30 seconds, in real time.

There is also some history behind this.

Reward AI builds on DexCap, work that came out of Stanford in 2024 around capturing human dexterous manipulation and transferring it to robots. @chenwang_j , C. Karen Liu and Li Fei-Fei were among the researchers involved.

I think the bigger bet here is about the data layer.

Robot hardware is changing quickly. If collecting training data means teleoperating every new robot body for thousands of hours, scaling gets expensive fast.

Reward is betting that you can collect much of that knowledge from humans instead, then transfer it across different embodiments.

That could become quite valuable if it works.

For now, we only have the launch material. There is no paper or open model yet, and Reward has not published the kind of large-scale success-rate data needed to judge robustness.

Still, this is one of the more interesting approaches I’ve seen to the robotics data problem recently.

Source: https://t.co/tYyhp90zGn

——

Weekly robotics and AI insights.
Subscribe free: https://t.co/9Nm01QUcw3

[CyberRobo @CyberRobooo]

Honestly, watching robot bartenders, robot installers, and humanoid assistants actually get work done autonomously is way more interesting than sitting around worrying about the dangers of AI.

They’re already moving fast and smoothly. Desktop manipulation tasks seem almost effortless, and they look pretty close to being ready for fast customer order fulfillment.

[Raphael cohen @cohenrap]

Why isn't laundry folding the first frame? 😍😍

[Afiz ⚡️ @itsafiz]

Turning a nut calls for a different kind of control than shaking a cocktail. The range from quick arm movements to careful finger work is pretty compelling here.

[Reward AI @RewardAI_]

RT @aliteracy: if the electromagnetic trackers are much faster as they claim, then their umi operators can be free to move a lot more quick…

[Peter Stone @PeterStone_TX]

Congrats @yifengzhu_ut and colleagues on impressive progress!

[Minghuan Liu @ericliuof97]

The speed is seriously impressive — I can only imagine how much hard work went on behind the scenes. Huge congrats to the team on the release!

[Shaoxiong (Shawn) Wang @ShaoxiongW]

Super impressive! The co-design of mechanics, sensing, and data are fantastic, and the policy is incredibly dynamic and smooth! Living the dream toward human-level dexterity! Big congrats to @chenwang_j, @zipengfu and the team!

[Shengjie Wang @ShengjieWa34067]

So amazing work!👍 It challenges the point that we need more robot data.

[Sudharshan Suresh @Suddhus]

So snappy, and such an underrated bit of co-design!

[Mark Grinev @markgrinev]

crazy seeing arms move so fluidly

[Shuo Yang @ShuoYangAIR]

Amazing demo from @zipengfu’s team!

[Siyuan Huang @siyuanhuang95]

Achieving human-level speed execution is something that Astra can never achieve. Congrats to @zipengfu and the team!

[Thomas Wolf @Thom_Wolf]

Wow

[anand iyer @ai]

First GEN-1, now OM-1. {roof is mounting that pre-training on human wearables without teleop or on-robot fine-tuning is a massive step forward for robotics.

The cross-embodiment dexterity (zero-shot!) from tabletop arms to full-body humanoids is wild to see!

[Yen-Jen Wang @wangyenjen]

Learning from human data could let robot learning scale beyond robot data collection. Cross-embodiment models make this even more interesting - shared skills across different robot bodies. A very promising direction!

[ヤマリキ@AI_robotics @rikiya_y]

ロボットがバーテンダーの手捬きを再現しているこの映像、人間の作業データだけで学んでいる。しかも映像は等倍。

Reward AIのロボット基盤モデル「OM-1」。一言でいうと、ロボットの学習データをロボットから取るのをやめた。
・センサー付きグローブを着けた人間が普通に作業する、そのデータだけで学習
・テレオペも実機データもゼロ
・卓上アーム / 産業用アーム / ヒューマノイドにゼロショットで転用
・複数ロボットの協調作業まで射程に入っている
ポイントは「ゼロショット」のところ。
従来のロボット基盤モデルは、テレオペで実機を動かして集めたデータが前提だった。だからデータが機体に紐づく。機体を変えれば集め直し。OM-1はデータの入口を人間の手に置いたので、まだ存在しない機体のためのデータを今日から貯められる、とのことだ。

[Tongzhou Mu 🤖🦾🦿 @tongzhou_mu]

The handheld gripper market just gained a strong new competitor. Congrats!

[Peng Xu @sippeyxp]

This is huge! 🤯

Reliance on real robot data and cross embodiment are big issues of robotics action models! Super excited to see OM1 solving both of them!

The motion is also fast and looks smart!

Great work @zipengfu @chenwang_j!