44 points by kstonekuan 13 hours ago | 12 comments
- Super cool! How do you think you will want to sell these into robotics teams, I find so many core engineering teams in Robotics companies have a "we'll just build it ourself" attitude around tools like this.
Have you had luck cracking past that?
- We have had this same experience, and it's a big reason we chose to go open source and make it easy for teams to migrate their business logic onto our SDK instead of requiring them to adopt a new system. By targeting a more foundational layer, we aim to be a part of that "built it yourself" system instead of replacing it.
- This is a strategy I’m taking as well. Build the tooling that can be integrated into anyone’s platform, over some all in one system that inevitably won’t check all the boxes.
- Yes, especially in a field moving as fast as robotics, users need the flexibility to extend the system quickly
- I'd be interested in using this with something like the new HuggingFace robots. Do you have demos showing how to go from HFlow dataset to training to deploy?
- Love those robots! Unfortunately, we don't support the training and deployment side yet as we are focused only on data processing for now.
- Creating a basic ergonomic example of the full train/deploy loop, using your data engine, for a popular platform in the zeitgeist, will give you a traction boost.
You don't have to innovate on the parts outside your core competency. Have your favorite LLM ingest the Reachy docs, give it control of a robot, and put a lightweight example together. Iterate till you like the example flow and you've purged docs and output text of AI slop, then release!
- Thanks for the advice, we'll definitely consider an example like this.
- How robitics may be “scalable”?
- Build a software implementation, lets call it The Matrix. Feed it to unknowing organic vessels, lets say Humans. Make it so they existence is a soothing mixture of comforting (fabricated) reality and inorganic desires. Extract power from these vessels. Make more robots. Profit.
- Our users process thousands of hours of data used in robotics model training, that's when they need scalable infrastructure to handle the workloads
- Filing this away