Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

The Onchain Execution Policy sorts every manager action before it executes. - Routine and bounded actions execute immediately - Actions that carry material risk enter a timelocked queue Jeroen on the reasoning, and on what sentinels do in Centrifuge V3.3.

10,306 Aufrufe • vor 2 Monaten •via X (Twitter)

11 Kommentare

Profilbild von Centrifuge
Centrifugevor 2 Monaten

Read more about V3.3 and the Onchain Execution Policy ↓

Profilbild von Justskilled.zk
Justskilled.zkvor 2 Monaten

@offerijns great but you rug us

Profilbild von dkarbon
dkarbonvor 2 Monaten

@offerijns Nice. Sorting actions by risk before they execute is a practical design for active onchain portfolios.

Profilbild von RightRedReturn @rightredreturn
RightRedReturn @rightredreturnvor 2 Monaten

@offerijns

Profilbild von shiho
shihovor 2 Monaten

@offerijns The community vote on the $CFG Rewards Date is now live. Eligible participants have 48 hours to cast their vote and unlock potential additional rewards. Submit your vote and discover the details below:

Profilbild von Alpha_Cherche
Alpha_Cherchevor 2 Monaten

@offerijns Timelocks mitigate governance exploits, but they introduce execution latency during fast liquidity events. ​Classifying "material risk" needs dynamic market parameters, not just static thresholds, or sentinels will lag behind market cascades.

Profilbild von steffie b
steffie bvor 2 Monaten

@offerijns The $CFG Rewards Date proposal is now open for community voting. Eligible participants have 48 hours to participate and may qualify for additional rewards. Review the proposal ⤵️

Profilbild von 太田清伸
太田清伸vor 2 Monaten

@offerijns Voting is now open to determine the distribution date for Rewards. The vote will remain open for 48 hours, with participation determining eligibility 🔗

Profilbild von Nitin Gupta
Nitin Guptavor 2 Monaten

@offerijns Splitting execution by risk tier instead of by action type is the harder design call. Most systems timelock everything or nothing.

Profilbild von 太田清伸
太田清伸vor 2 Monaten

@offerijns ATAAYYYYYYYY

Profilbild von j.cheol
j.cheolvor 2 Monaten

@offerijns Where the hell are you? You betrayed your holders and now you’re nowhere to be found. The only way to save $CFG is to apologize and start a token buyback. That’s the only solution. Forget the pointless 14-day feedback process. Fix this mess. Now. @itsbhaji

Ähnliche Videos

How do you grade a world model? Against what the robot really did. Picture a robot gripper closing on a piece of fruit. In the real world, it fails to grasp it. Now hand the same first frame and the same action sequence to an action-conditioned world model (ACWM). The rollout shows a tidy, successful pick. Nothing on screen looks wrong. That is the moment WorldOracle Bench was built for. Created by AGIBOT and NVIDIA together with Princeton, UT Austin and the University of Toronto, it is an evaluation platform for robot behavior models. It replays the real, recorded actions through a model and checks the predicted future against the paired real execution. The stakes are practical. The paper points to data generation, scalable policy evaluation and closed-loop policy learning as potential uses of ACWMs. If a model gets the physical consequences wrong, everything built on top of it inherits the error. The authors call the property they are testing execution coherence: does the model follow the actions it is given, even the bad ones, across different embodiments and action sources, and over long rollouts? What makes the setup unusual is where the evidence comes from. The paper compares eight existing embodied world-model benchmarks, and none of them combines action-conditioned evaluation, real-world data, failed executions, long episodes, multiple embodiments, multi-view outputs and policy-generated actions. WorldOracle Bench does. Its evaluation set has 335 episodes: 100 on AGIBOT G2, 75 on YAM and 100 with a human hand, all human-operated, plus 60 rollouts where a learned policy drove a real AGIBOT G2 robot. About 40% of the human-operated segments are failures, so a model cannot get by on success stories.

Science girl

24,313 Aufrufe • vor 17 Stunden