Pratikk369

minting moments. Faster chains, funnier nights.

[Pratikk369 @pratikk369]

alpha got something new

[Consensus @consensus_games]

Are you ready to own the decision?

The first @Uniswap v4 Hook powered GameFi on Robinhood Chain

🏛️ 777 Genesis Oligarchs
💎 Real yield distributed in $ETH
🎟️ Free Mint on @opensea

Allowlist details dropping soon. Turn notifications on 🔔

[Pratikk369 @pratikk369]

@consensus_games @Uniswap @opensea boom

[Pratikk369 @pratikk369]

@axisrobotics takes a Simulation First approach.

Users can control simulated robots from a browser, complete tasks and generate trajectories containing robot states, object poses, actions and task data.

Those trajectories can then be expanded into training samples.

[Axis Robotics @axisrobotics]

Grateful to @cicada_mm for featuring us in their research blog.

Dive into the full piece for our thinking on scaling data volume without sacrificing quality, how we prioritize robot embodiments, and what foundation-model scale actually means.

[Axis Robotics @axisrobotics]

Announcing our partnership with @Dexmal_AI

About Dexmal: Dexmal is a technology company focused on general-purpose embodied intelligence, committed to building intelligent, useful, and trustworthy robots.

As a core data infrastructure partner, Axis is teaming up with Dexmal to empower their VLA and world models through large-scale egocentric, simulation, and real-world data production. Backed by successful enterprise data deployments, this milestone highlights the scale, precision, and production-grade reliability of Axis’s compounding physical data engine. Together, we are establishing the core data backbone to accelerate next-generation embodied intelligence.

Learn more about Dexmal: https://t.co/B4rhl8F3Au

[Axis Robotics @axisrobotics]

From Digital Twins to Data Engine: Cutting the Real-Data Burden with Sim-Powered Robot Learning

Teaching a robot a new task could take hundreds of teleoperated demonstrations. For foundation models, adapting to an entirely new robot can cost orders of magnitude more — dedicated hardware, trained operators, months of engineering.

On @boosterobotics' dual-arm robot, we studied this at two levels:

✱ Specialist: Can task-aligned simulation mixed with a small set of real demonstrations reduce the real-data burden?

✅ Yes. With only 10 real demos, the policy made no contact at all in physical rollouts (0/20). Adding 50 simulated trajectories brought contact to 17/20.

✱ Foundation: Can data accumulated across tasks build a reusable starting point (a Booster-specific model prior)?

✅ Yes. After full-parameter continued pretraining, a model adapted with just 30 demonstrations per task beat the original given twice as many: 14/16 vs 10/16 in simulated evaluation. Before any task-specific adaptation, in zero-shot simulation, it was already roughly 3× closer to the target (17.27 cm → 5.78 cm).

This work runs on Axis Suite, our Physical AI solution across different robot embodiments. Distributed contributors generate task-aligned sim data on Axis Hub at scale, reducing real-data needs for specialist adaptation while powering cross-embodiment generalist training.

Read the full blog: https://t.co/gRVi2LZnaJ

[Axis Robotics @axisrobotics]

Axis Robotics x @BinanceWallet is live.

1,500,000 Axis Points. 30 days. Reserved exclusively for Binance Keyless Wallet users.

Teach robots. Sign your data on Base. Get paid in Points from a pool that's entirely separate from the main product.

How it works:

[Axis Robotics @axisrobotics]

Axis Hub Update: The Pause Button is LIVE.

✱ Click once — the button lights up and the simulation pauses between your actions. Click again to deactivate.

✱ Pre-training: when active, the simulation pauses after each move so you can think before the next. When off, the timer runs nonstop once you start.

✱ Post-training: each takeover is capped at 8 steps — and they run out fast. Pause lets you plan each move without burning through your budget.

See it in action ⬇️

Why we cap interventions at 8 steps → see the quoted thread.

[Axis Robotics @axisrobotics]

What Happens After Post-Training Data Collection?

Since launching Axis V2, human-gated DAgger correction has been a core part of our data engine. In this thread, we share a series of experiments exploring how to best process and use post-training data.

The core finding: not every human intervention helps. But if you verify which specific actions actually change the outcome — and train only on those — the improvements are real and they scale.

Results:

- 660 corrections collected → 161 (24.4%) entered training
- Each training snippet is ~0.8 seconds
- Naively imitating full human trajectories dropped success from 40.0% to 36.7%, while 161 verified snippets lifted the three-seed average to 48.3% on the same evaluation set — and reached 48.8–52.5% across three seeds in a separate paired evaluation.

Read the full blog: https://t.co/ASYarP7CgU

Details below ⬇️

[Axis Robotics @axisrobotics]

Axis Hub update: The Signal Boost campaign is now live in your Portfolio.

How to participate:
1️⃣ Navigate to Portfolio (or click the Signal Boost banner).
2️⃣ Select "X Connect & Follow" next to your username to link your X account.
3️⃣ Verify your status to activate the Signal Boost badge and receive 1 Axis Point:

Already following: The badge lights up automatically upon connection.

Not following yet: Click the button again to follow on X, return, and the badge will unlock.

🔗 Access Axis Hub: https://t.co/h1HaIkMiu7

[Axis Robotics @axisrobotics]

New to Axis? Here is a quick overview to get you started on tasks and earning points:

🟩 Daily Release Time

New tasks go live daily at 12:00 UTC.

🟩 Task Stages

✱ Pre-training: Teleoperate the robot from scratch to teach the base policy. (Use N to save checkpoints and B to restore on long tasks.)
✱ Post-training: Supervise the AI policy and intervene only when it is about to fail. (Manual takeover is capped at 30s per run; click "Return to Policy" right after correcting.)

Tutorial on Post-training tasks: https://t.co/9yak5xuAqq

🟩 Basic Controls

✱ Move: Drag the blue ring on the arm to position the gripper.
✱ Auto approach: Double-click an object to send the arm to a pre-grasp pose.
✱ Keyboard shortcuts are available in the in-game Controls panel.

🟩 Tips to Maximize Points

✱ Pre-training: Focus on task success, efficiency, and motion smoothness.
✱ Post-training: Less is more—minimal corrections and shorter trajectories yield higher scores.
✱ Task Variety: Higher difficulty (up to 5 stars) and broader task diversity award more points.
✱ Sign Runs: Always sign completed runs under Portfolio → History. Unsigned runs do not earn rewards.

📖 Full FAQ: https://t.co/TGN8F7dVrW
🎥 Watch the complete walkthrough in the quoted video below.

[Axis Robotics @axisrobotics]

Introducing the Axis Content Creator Program: The Katalyst for Physical AI

Partnering with @KaitoAI, we are rewarding Axis lore writers with 0.25% of the $AXIS token, to be fully distributed at TGE.

Epoch 1 is now live and runs until September 18th at 12:00 UTC.

We appreciate your efforts in helping Axis grow and push the frontier of physical AI to new heights.

Let’s dive into the campaign details:

[Axis Robotics @axisrobotics]

Axis Weekly

Last week, we focused on unblocking production task generation, moving scene variants off the critical path, and turning HG-DAgger from a local proof of concept into a measurable training recipe.

The week combined infrastructure work — faster scene delivery, a cleaner verify flow, and a quicker policy evaluation loop — with the first clear distillation gains on both old and new task distributions.

Key updates:

- Scene & collection UX: Scene variants load faster, lighting rendering is back, and post-task scoring is tighter.

- Task generation & throughput: LIBERO Pro, RoboCasa, and wheeled embodiments (incl. wheeled Franka) are now in regular generation, with the publish queue consistently above 200 tasks. Policy training and evaluation throughput roughly tripled.

- HG-DAgger distillation: A curated-correction recipe lifted success rate by 20+ points on both old and new distributions (more seeds still needed).

Details below 👇

[Axis Robotics @axisrobotics]

Something big is coming your way…

Notifications on.

[Axis Robotics @axisrobotics]

Announcing our partnership with @OpenRoboto as Axis Robotics expands its data capabilities into the @opentensor ecosystem

OpenRoboto is an open competition on Bittensor (Subnet 80) for continuously improving robotics models. Miners fine-tune an open base model starting from π0.5, every submission is scored in randomized LIBERO-Pro environments, and only a strictly better model becomes the new base.

As the core data engine behind OpenRoboto, Axis is supplying over 3 million multi-modal trajectories into OpenRoboto’s Open Data Pool to accelerate physical AI model training for miners across the Bittensor network. Furthermore, Axis will support OpenRoboto's authoritative benchmarking platform with its Data-to-Model Pipeline, enabling robust, auditable evaluation for both data and models across multiple benchmarks.

This partnership establishes a robust, end-to-end framework uniting distributed data collection with verifiable model evaluation.

Learn more about OpenRoboto: https://t.co/9viK2AXyfy

[Axis Robotics @axisrobotics]

Decentralized robotics intelligence needs open data infrastructure to scale.

Congrats to @OpenRoboto on launching the Open Data Pool on @opentensor. As OpenRoboto enables open data access and model evaluation, @axisrobotics is powering the pipeline beneath with our compounding data engine.

More soon 🤝

[Axis Robotics @axisrobotics]

Our robotic data engine on @base just hit 3M trajectories.

1M on June 7. Now 3x.

Generated by 123k+ contributors worldwide, with contributions recorded, verified, and rewarded on-chain.

@base is the Hub for Physical AI, and we are accelerating it at a scale you've never seen.

[Axis Robotics @axisrobotics]

Here are some really good examples of human takeover from community:

Precise, minimal intervention.

[Pratikk369 @pratikk369]

Social login let me start collecting in a minute with no wallet or extra apps.
3.7M trajectories and 42k hours make even short sessions feel like they count.
I switch between phone and laptop in spare minutes and still add real plus sim data.

@privy_io x @axisrobotics

[Axis Robotics @axisrobotics]

Axis has been building alongside @privy_io from Day 0.

As a pioneering Physical AI data engine in the Privy ecosystem, we’re proud to provide a friction-free onboarding experience for over 150,000 global contributors through social login.

Together, we’ve achieved:
✱ Over 3.7M Trajectories Collected
✱ 42,000+ Hours of Trajectory Data
✱ 4,000+ Active Tasks

Experience it yourself today: https://t.co/h1HaIkMiu7

[Pratikk369 @pratikk369]

@Ntekearning @yososocial only for inside user only shittttt man