0xiChl

.
Contributor @AxisRobotics

[Hikal (✱,✱) @HikalHank]

Glad if the script can help my sobat axis gw.

You guys tired signing @axisrobotics task one by one? I have a solution. You can use a script that can be work through a desktop web browser console. I'll share the txt file link in the comments👇

[0xiChl @0xiChl]

@HikalHank @axisrobotics bola bali kapusan

[0xiChl @0xiChl]

@axisrobotics highlights an important point:
A robotic dataset isn't just about collecting more trajectories

The real value comes from task diversity
Different tasks teach different skills, helping robots learn and perform better in the real world

[Axis Robotics @axisrobotics]

https://t.co/v8VgqcSfOn

[0xiChl @0xiChl]

@Hengtzyggltmpn @axisrobotics

[0xiChl @0xiChl]

From manual workflows to a systematic service
@axisrobotics is automating the entire loop: task generation, continuous growth training with a 10% performance bump, and automated evaluation heatmaps
Big step for scalable Physical AI

[Axis Robotics @axisrobotics]

Axis Weekly

This week was about making the AXIS loop more scalable end to end: automating data-to-model workflows, testing recovery-driven training, expanding TaskGen coverage, and preparing the dataset and model stack for release.

Key updates:

- Data-to-model automation: We used scripts to speed up and standardize several repetitive but critical workflows.

- Continuous-growth training: We completed multi-data-scale training and success-rate comparisons across several failure tasks.

- Failure task expansion: A new batch of failure tasks has been pushed to test, expanding the evaluation range for ablations across data scale, data quality, and randomization.

- TaskGen: Articulated-object generation is now merged into the automatic generation pipeline.

- Model and release prep: We finished the first round of fine-tuning, evaluation, and benchmarking, completed the dataset’s conference submission, and are now improving experimental results for release.

Details below 🧵

[0xiChl @0xiChl]

@Hengtzyggltmpn Masih di pertimbangkan

[0xiChl @0xiChl]

Broooo wtf is this

[0xiChl @0xiChl]

@HikalHank @axisrobotics

[0xiChl @0xiChl]

@HikalHank @axisrobotics belum bang wkw

[0xiChl @0xiChl]

@0xRyzzu @axisrobotics action data is the real edge here
text data scales easily, but physical interaction is a different game

[0xiChl @0xiChl]

@Zhafreen3 bukankah ini my kisah, banyak yang failed

[0xiChl @0xiChl]

@HikalHank @axisrobotics udah axis banget

[0xiChl @0xiChl]

@Zhafreen3 @axisrobotics exactly ser, that's the key point

[0xiChl @0xiChl]

A good Physical AI pipeline needs structure

Environments to run tasks, humans to create behavior, systems to refine data, and models to learn from it

That's why @axisrobotics architecture is interesting
It treats robotics data as a full stack, not just a dataset

[0xiChl @0xiChl]

@BirlyAditya @axisrobotics exactly, @axisrobotics is one of the interesting plays here

[0xiChl @0xiChl]

@sanwahm @axisrobotics @AxisRoboticsID bots filtered, data protected

[0xiChl @0xiChl]

@Deekyaja @axisrobotics exactly ser, it's not just about the model. the environment, data, behavior, and validation layer all matter

[0xiChl @0xiChl]

@asmedesigned @axisrobotics same here, still learning too