←
DATA Ecosystem1,998 posts
DATA Ecosystem avatar

DATA Ecosystem

@Data_Ecosystem

Home of the Data Foundation ecosystem.

◷ Joined September 2024▧ 179 Photos and videos
DATA Ecosystem avatar
DATA Ecosystem@Data_Ecosystem·Oct 28

24 hours left in the DATA on Sui: Content Quest. To enter: ▸ Create: original content about DATA on Sui ▸ Amplify: quote 3 of our DATA on Sui posts with your take. Entries close Oct 28, 11:59pm PT. Last call ↓

0:06
Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·21h

Mentally, we're still at @corl_conf. Our highlights: ▸ Robots are more accessible than ever. Deployment is the next frontier ▸ Real-world deployment data is becoming robotics' most valuable asset ▸ @RussTedrake reinforced the importance of deployment data, showcasing robotics on a Toyota factory line ▸ Dexterity, tactile learning, and real2sim2real are gaining momentum ▸ Human video is increasingly being converted into executable robot actions More from @jdubpark ↓

CoRL takeaway: Robots are now fully accessible... deploying is next. With off-the-shelf hardware and agentic robotics (Astra), you can put together a working robot. It will be slow and make mistakes. But this turns the question from “can anyone build an e2e robot pipeline?” to “how do you improve one?” The race is now on to get them working in the field. So the visible moat is deployment knowledge. WAMs/VLAs, sim, hardware, policies, etc. are all being commoditized. And every architecture and method chases the same goal: robots in prod at scale. Ofc, to get there you need a closed feedback loop that includes deployment, like Optimus in the Gigafactory or Amazon FAR in warehouses. Deployment is the back-propagation for robotics. Despite the accessibility, the gap between research, hardware, and real-world usage remains wide. Walden Robotics founder @RussTedrake's keynote impressed many with “deployment data is the key” and a demo on a Toyota factory line. Tbh, that’s common knowledge. But in our minds, the path from software to robots working in production lines is still very murky. And this conference is a great wake-up call for roboticists and companies to be more practical -- get your hands dirty in factories and homes, break things and learn to recover in prod. Perfection matters more than generalization in most deployments - failure is unacceptable. Agentic robotics is emerging fast! Astra and Fable can control robots with no extra training decently (pre-trained on ego videos :p). It’s slow for now, so most people treat it as long-horizon planning. The stack is settling into System 2 (what to do), System 1 (how to do it - agentic robotics), and System 0 (real-time actions). More notes: Progress to a functioning robot in the wild depends on connecting three capabilities: - Learning from human demonstrations - Reacting to contact and the environment with hands and whole body - Retargeting data & skills across different hands and bodies Human video is becoming a source of executable demonstrations, not just for training for generalization. EgoScale, Do as I Do, and Video2Sim2Real all convert video into actions a robot can execute. Lots of imitation learning alongside RL (though it's doubtful whether it's better than RL for prod). Deployment data is the gold standard, since everything in robotics exists to make deployment work. Walden’s 4.7B “Sage” model trained on 100k hours from a Toyota line and performs very well on that task (unsure for other tasks). Teleop interventions catch failures and record them to improve the model. Synthetic data is growing too, ultimately for real2sim2real: bring reality into sim, train and test at scale, transfer back. GRAIL, Play2Perfect, etc. Dexterity got a lot of attention. T-Rex used 100 hours of tactile data to sharply raise task success. Some papers drop vision and rely on touch alone... but again, why? Feels very detached from practicality. Unlike Walden, tactile research is still far from industry deployment.

Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·2h

Failed segment ≠ failed recording. Failure-aware QC can salvage usable footage, without discarding entire videos. In our #CoRL2026 paper, we audited 500 hours of footage to understand the failures that affect Physical AI training data. Full paper: openreview.net/pdf?id=rLGRN1…

Post attachment
Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·23h

Better data is the breakthrough. Today, we're publishing a case study on how our 10,000-hour Bengali speech dataset reduced Character Error Rate (CER) by ~32% on Whisper Large V3. ↓

Post attachment
Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·Nov 11

We're honored to present research at @corl_conf. Our team studied 500 hours of crowdsourced egocentric video to better understand where data collection goes wrong, and how to improve quality control for Physical AI. One key finding: 90% of footage in flagged clips falls outside labeled failure intervals. Rejecting entire clips can mean discarding a lot of potentially useful data. Tremendous work by @iamrashminagpal, @sebliu, @jdubpark, and Sandeep Chinchali. First look at our research poster below. Catch us tomorrow at the 'Oops, I Erred' #CoRL2026 workshop ↴

Post attachment
The DATA Foundation avatar
⟳ DATA Ecosystem retweeted
The DATA Foundation@datafdn·Nov 9

Verifiable data infrastructure is coming to enterprise AI and finance with @spaceandtime. Proud to be a launch partner as we continue building toward a future where data provenance, privacy, and verifiability are foundational to AI.

Post attachment

Introducing Space and Time Enterprise: a new data privacy platform for AI and finance. Deploy Proof of SQL in your enterprise environment, keep your underlying business data private, and share verifiable results with your AI agents, auditors, and stakeholders.

1:51
Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·Nov 10

Come say hello to the team at @corl_conf. @sebliu will present our paper "When Crowdsourced Egocentric Video Goes Wrong: Failure Modes, Metadata, and Quality Control" at the "Oops, I Erred" workshop. 🗓️ November 12 📍JW Marriott Austin

Post attachment
Poseidon avatar
⟳ DATA Ecosystem retweeted
Poseidon@psdnai·Oct 28

See you at @corl_conf 2026. Our paper, "When Crowdsourced Egocentric Video Goes Wrong: Failure Modes, Metadata, and Quality Control," was accepted to the "Oops, I Erred" workshop at CoRL, a leading robotics research conference. If you'll be in Austin, say hi: @sebliu + team will present our research on a critical challenge for Physical AI: how to ensure crowdsourced, real-world video data is reliable and useful for training robots. corl.org/

DATA Ecosystem avatar
DATA Ecosystem@Data_Ecosystem·Oct 27

Our friends at @otodotearth are searching for contributors to support AI Training via video chats. See post for details ↓

We're looking for people to help train AI through video chats with a friend! Up to $50/h per eligible recorded hour, per person. 10-minute chats. US residents 18+. Face and voice used in commercial AI datasets. Hours not guaranteed. Eligibility and rates: fiber.oto.earth/?utm_source=…

DATA Ecosystem avatar
DATA Ecosystem@Data_Ecosystem·Oct 15

Create: Make a thread, video, visual or meme about @datafdn x @SuiNetwork. $1k grand prize, 4 × $500 prizes, and a dedicated KR-language prize. Amplify: QT our DATA x Sui posts with your take on why rights-cleared AI data matters. 4 × $500 prizes. Closes Oct 28 11:59pm PT. Full Details on Discord ↓ discord.gg/datafdn

DATA Ecosystem avatar
DATA Ecosystem@Data_Ecosystem·Oct 15

Help us tell the DATA on Sui story. ▸ $5,000 USDC prize pool ▸ 9 winners total ▸ Every prize $500+ Two ways to win ↓

0:06
The DATA Foundation avatar
⟳ DATA Ecosystem retweeted
The DATA Foundation@datafdn·Oct 13

The lifecycle of data, as told by the DATA protocol → @SuiNetwork.

0:26
Andrea avatar
⟳ DATA Ecosystem retweeted
Andrea@devrelius·Oct 12

This is a big step forward for DATA with a $30M investment from Sui. This makes Sui Foundation the largest holder of DATA, above DATA Foundation and PIP Labs. Our founder SY Lee is also joining Sui as a strategic adviser on AI and training data. Most importantly, bringing Trace and the 275M+ data receipts onto the Sui stack gives us the infrastructure to go beyond provenance alone, with reference data storage, access control, and stronger verification for the real-world data AI increasingly depends on. Excited for what we will build together.

DATA is coming to @SuiNetwork. The DATA Network and Trace are joining the Sui stack, bringing a live supply of rights-cleared AI training data to infrastructure built for AI workloads. With 250M+ receipts already recorded by Trace, AI labs will be able to source real-world data with verifiable provenance, while @WalrusProtocol provides the underlying storage layer. Sui Foundation is investing $30 million in the DATA Network, establishing the largest institutional stake in the network, alongside the continuing holdings of PIP Labs and the DATA Foundation. Furthering the alignment, SY Lee, founder of the DATA Network and CEO of PIP Labs, joins Sui Foundation as Strategic Advisor. In the coming months: ▸ The DATA Network will be upgraded to run as a protocol natively on Sui ▸ Walrus becomes the default storage layer for DATA Foundation apps ▸ $DATA remains the token of the DATA protocol The next chapter of DATA starts here.

Post attachment
Altcoin Buzz avatar
⟳ DATA Ecosystem retweeted
Altcoin Buzz@Altcoinbuzzio·Oct 12

SUI is making a move for the AI data economy. DATA Network is bringing its Trace registry to Sui, with Walrus as the default storage layer, giving AI labs a single place to source rights-cleared training data. Why does that matter? AI models need data, but using data without clear rights can become a legal headache. $SUI is positioning itself right where AI, data ownership, and decentralized storage collide. And that could get interesting fast.

Post attachment