Baovy06
Baovy06
• HODL through thunderstorms, reaping fruit at moonrise. • Position makes it all. • Calm before the wave, steadfast in front of the chart.
1KFollowing
1.1Kfollowers
Feed
Feed
Suddenly missing Hanoi
It's been a long time since I visited Hanoi
The gentle West Lake breeze carries the scent of lotus
The café glows with green and red lights
Cars pass by, leaves fall along the roadside trees
Walking through each street step by step
In midsummer, flamboyant flowers bloom brightly
Soft sunlight falls in strands under the eaves
So many memories stir my heart with longing……
@quipnetwork @NucleusCodes @sleepagotchi
If you’ve been active in the @sleepagotchi ecosystem, now’s the time to check your eligibility
The project has officially launched its Creator Leaderboard on Nucleus Codes with a massive $120,000 reward pool in $SLEEP for creators and active community members.
What makes Sleepagotchi interesting is that it’s not just another web3 game. The team is building around sleep-fi, gamification, and AI wellness, creating a unique ecosystem that stands out from typical GameFi projects.
They’ve also raised millions in funding and consistently pushed community campaigns with strong engagement and solid rewards for users.
If you’ve done tasks, played their mini games, or supported the project on socials before, go check your eligibility now
Join now:
This week, the team at @axisrobotics continued improving many important parts of their robotics data system, from automatic task generation and simulation environments to failure recovery and object data augmentation.
But for Zy, what makes a project truly go far is not just the technology itself, but also the support and contribution from the community.
And the longer Zy stays with Axis, the more that feeling becomes real.
Here are a few updates that really stood out to Zy this week:
• The task generation system has been upgraded to better understand available objects, environment layouts, and long-horizon workflows across different robot embodiments.
• The simulation infrastructure is now much more stable, especially when running long tasks or handling multiple objects at the same time.
• Robot controls were improved based on community feedback, especially for grasping, robotic arm movement, and the control interface.
• One of the most interesting things is how the team is using failed or near-failed robot actions as new training data. This helps robots learn how to recover from mistakes instead of only learning from successful attempts.
In addition, Axis is also collaborating with research groups to develop object-level data augmentation technology. From a single object, the system can generate multiple realistic variations to improve training quality.
Zy has been experiencing the project for more than a week now, and honestly it has been pretty enjoyable. Some tasks can still feel a bit laggy at times, but it also teaches patience and persistence along the way
As for the community, there’s really nothing to complain about. Everyone is active, supportive, and always willing to help each other.
Especially this morning, Zy and a few community members joined Discord and started a small karaoke session together 😂 It was just for fun, but somehow it created a really warm and connected feeling. Moments like these are what truly give life to a community.
Hopefully all of us can continue keeping this positive energy, supporting each other, and staying together until the very end of the Axis journey ❤️
Thank you everyone so much.
Axis AI
Axis Weekly
This week, we continued strengthening our closed-loop robotics data pipeline, from TaskGen and simulation infrastructure to failure recovery and asset-level augmentation.
Key updates:
- Task generation: We completed asset scan and merged it into TaskGen, helping generated tasks reason over available assets, scene layouts, long-horizon workflows, and multi-embodiment settings.
- Simulation infra: We improved MuJoCo verify, replay, and scene-variant workflows, with fixes around repeated downloads, caching, compatibility, and long-horizon multi-asset task stability.
- Robot controls: We cleaned up gripper behavior, IK, teleoperation, and the control panel based on feedback from longer-horizon and multi-asset tasks.
Failure recovery: We continued building a pipeline to turn failed and near-failed grasping states into reusable data for recovery learning.
- Asset augmentation: With academic collaborators, we advanced a shape augmentation direction that can expand one seed asset into many physically plausible object variants.
A closer look at this week’s progress 🧵
Baovy06 reposted
Melody Vietnam National Ambassador Program｜Recruiting 1 Leader
Melody is launching the Vietnam National Ambassador Program to grow local communities, content outreach, and offline impact. We are recruiting 1 National Ambassador to represent Melody across Vietnam and connect with music fans, Web3 users, and creators.
Application channel:
#RWA #MusicFi #MELO
Baovy06 reposted
Axis Weekly
This week, we continued strengthening our closed-loop robotics data pipeline, from TaskGen and simulation infrastructure to failure recovery and asset-level augmentation.
Key updates:
- Task generation: We completed asset scan and merged it into TaskGen, helping generated tasks reason over available assets, scene layouts, long-horizon workflows, and multi-embodiment settings.
- Simulation infra: We improved MuJoCo verify, replay, and scene-variant workflows, with fixes around repeated downloads, caching, compatibility, and long-horizon multi-asset task stability.
- Robot controls: We cleaned up gripper behavior, IK, teleoperation, and the control panel based on feedback from longer-horizon and multi-asset tasks.
Failure recovery: We continued building a pipeline to turn failed and near-failed grasping states into reusable data for recovery learning.
- Asset augmentation: With academic collaborators, we advanced a shape augmentation direction that can expand one seed asset into many physically plausible object variants.
A closer look at this week’s progress 🧵
