Carter Ness

[email protected]

[Carter Ness @CarterNess9]

Current building a SaaS. Looking to connect with robotic engineers and developers.

Inbox is open 👀

[Carter Ness @CarterNess9]

@axhiii88547 building before you understand it leads to a completely slower process in the long run.

[Carter Ness @CarterNess9]

@axhiii88547 Planning and having a true understanding before you build.

[Carter Ness @CarterNess9]

@IliaLarchenko @FlexivRobotics @ieeeiros Would a system to manage runs and evidence be helpful for this problem? It can enhance the efficiency of finding errors.

[Carter Ness @CarterNess9]

@BoehmeMargarete @axisrobotics Does this include a accurate evidence tracking and run comparison tool?

[Carter Ness @CarterNess9]

The modern day shift to hardware brings countless opportunities. A software to bridge the gap between the hardware and software world. There is no undo button within hardware. How do you know what changes when a robot breaks? Is this the holy grail for AI?

[Mustafa. @oprydai]

if you are in software, 2026 is the best year to make the pivot to hardware

[Carter Ness @CarterNess9]

No doubt on provenance, 3 important additions:
-the robots setup, how did we get to this hardware?
-Comparison, what changed between a success run and a failed run.
-Testing changes before they were deployed.

[Hosam 🟣🟢 @hosam_web3]

One aspect of @axisrobotics that deserves more attention is data provenance.

For Physical AI, collecting massive amounts of robot data isn’t enough.

The real value comes from understanding the context behind each trajectory:

• How was the task performed?
• Did it actually succeed?
• Where did the data come from?
• What happened to it before entering the training pipeline?

As data collection scales through the community, having a clear and verifiable history becomes increasingly important.

Axis connects contributions with on-chain provenance, adding a layer of transparency and traceability to robot data.

That could lead to better datasets for training and evaluation, while making results easier to understand and reproduce.

The future of Physical AI won’t be defined by how much data we collect, but by how well we understand and trust the data behind it.
https://t.co/N39yxNvpb1

[Carter Ness @CarterNess9]

Robotics engineers trying to figure out what changed between the last successful run and today’s failure.

[Carter Ness @CarterNess9]

Robotics engineers: what’s the hardest robot failure you’ve had to troubleshoot, and what finally helped you find the cause?
https://t.co/WwpdU5Q362

[Carter Ness @CarterNess9]

@Jony3016 @Naz_Web3 My guess is most are still looking through folders, searching for that data within the robotics space especially.

[Carter Ness @CarterNess9]

@badbuta21 @0xfablo Agreed. What made you say that? Is there a case where you couldn't trace a dataset back to where it came from?