I found a trader with a $300K+ PnL.
But after digging through 13 weeks of trading history with @SurfAIHQ the number started telling a very different story.
Meet @rasmr_eth
π€ Paid partnership Β· π€ Made with AI
1
3
55
π«ππ¬π¦π« seems to have an interesting timing edge.
But the weekly PnL, position sizing and drawdowns tell a much more complicated story.
I wouldn't blindly copy these trades.
I'd watch how this trader identifies narratives before they become crowded.
1
4
That's where the real research begins.
Research conducted with @SurfAIHQ using ππ¦ππ«π ππ¨π§ππ² Inflow and historical trading data.
5
A robotics episode can look perfect and still leave important questions unanswered.
What object was the robot interacting with?
Where was the task performed?
Which embodiment collected the demonstration?
What exactly was the robot expected to do?
π€ Paid partnership Β· π€ Made with AI
4
6
36
One practical tip for data collection teams:
Don't treat metadata as an afterthought.
Define the task, environment and robot setup before recording begins.
1
3
A human hand and a robotic gripper can perform the same task.
But they don't perform it in the same way.
A human naturally adjusts grip, changes approach angles, and compensates for small differences in object position.
π€ Paid partnership Β· π€ Made with AI
2
6
44
A movement that feels natural to a person might be difficult for a particular embodiment.
Good teleoperation demonstrations account for that difference.
The goal isn't to teach a robot to move like a human.
1
6
It's to give it enough real-world examples to learn how to act with its own body.
#PrismaX #Teleoperation #Robotics #PhysicalAI #VLA
5






