Repository navigation
Request to acknowledge NeuralOS in the repo/blog and clarify code lineage #5
Description
Hi all,
This is a very cool work, but I'm writing to request a public acknowledgment of NeuralOS in the Neural Computers repo/blog.
For context, NeuralOS is here: https://github.com/yuntian-group/neural-os
Online demo is here: https://neural-os.com
I understand that NeuralOS is MIT-licensed, so my concern here is not reuse itself. Rather, I was surprised that neither the current repo nor the blog post appears to acknowledge NeuralOS, even though the released data-collection code appears to substantially overlap with our earlier data-collection pipeline.
If you feel it would be appropriate, I would appreciate a short acknowledgment of NeuralOS in the repo README and, if possible, in the blog post as well. That kind of acknowledgment matters a great deal to us who built these systems over the course of almost two years, and it helps maintain an accurate public record of technical lineage.
Thanks for your consideration,
Yuntian
Activity
mczhuge commentedon Apr 11, 2026
Hi @da03
It is fair to add it to README! It is truly my negligence and I add all of the repos I have refered.
We indeed build our data collection (GUIWorld part) pipeline from your original codes (anonymous GitHub, August 2025?). However, we also conduct improvements of the codes (and solve bugs, because that version has not conducted code cleaning), and added CUA ourself (your previous codes only support random trajectory). We also hope our improvement of codes could also help yours.
Noted that I am really happy we may share many insights. Will you join ICLR? Happy to talk in-person.
Thank you,
Mingchen
--- after check again your newest repo ---
Your current repo still does not support CUA collection, which I think is the most important for the community. In our experiments, random collections are not very useful for improving performance; in our current repo, we provide that.
There is an issue that you have not answered before: yuntian-group/neural-os#1
I also think you could refer to our work to improve your current version.