Software for dependable physical AI. Building Fraeno, a pre-deployment security gate for robot software updates.
San Fransisco, CA, USA
Joined August 2025
- Tweets17
- Following126
- Followers60
- Likes251
DeepUbuntu Labs retweeted
I literally just woke up from a dream where @tim_cook approached me and we sat down for maybe one or two minutes talking about Wand @getwand_app, the project I built that lets you control your Mac through an AI assistant like Muse.
At some point, he said he remembered when he listened to me pitch it to someone named Asia Montgomery.
(So, if there is an Asia Montgomery who works at Apple: apparently Iβve either met you before or my subconscious has. Please reach out π)
Tim seemed genuinely interested and captivated by the project, but we never actually reached any final verdict. I finally escorted him out, we said our goodbyes, and then I woke up.
What a ridiculously specific dream.
DeepUbuntu Labs retweeted
Free for Mac, public beta. Install, sign in with Google, then just ask Muse.
getwand.app
DeepUbuntu Labs retweeted
Muse on your iPhone can't see your Mac's screen and apps. But with Wand, now Muse can.
We gave Muse more powers. Try Wand.
getwand.app
Wand lets Muse look at your Mac, and take virtually any actions for you, from your muse app on your iPhone.
DeepUbuntu Labs retweeted
How a robot arm is controlled, explained for ML people new to robot learning. First of the explainers from my own speedrun.
Next one will be on ACT. Follow me to catch it.
Software doesnβt care about physics. But robots have to obey them
Today, weβre launching Fraeno
Fraeno runs robot software updates in isolated CI/CD env's, detects behavioral regressions, & blocks bad deployments
Physics checks coming soon
beta test fraeno.com pls
DeepUbuntu Labs retweeted
Thanks @Eric_Wallace_
Getting to work now. But, how do we maintain a balance between not violating TOS and responsibly testing capabilities?
Am I supposed to test only on my personal or my companyβs systems only?
Today we are releasing GPT-5.6-Cyber.
The model is our first large-scale attempt at directly improving capabilities for advanced cybersecurity tasks such as exploit development.
We are finding it to be really quite strong for accelerating defensive work. We are using it across our stack for red-teaming, and our security researchers have used it to find and patch a huge host of 0-day vulnerabilities in open-source software.
openai.com/index/expanding-dβ¦
True. Reality has more dynamics that physics can simulate on paper.
DeepUbuntu Labs retweeted
itβs surprising to me how many people seem to not understand that great models are built with super high quality curated data
finding novel ways to create / get this data is a huge edge
DeepUbuntu Labs retweeted
I will be attending YC Startup School 2026 in sf (July 25-26). if you 1) want to pitch your startup or 2) you are building autonomous vehicles-adjacent, or 3) you are building monitoring layers for AI agents, dm, let's have a coffee on me.
Also, If you are cracked, pitch your startup in the comments, might refer you.
If you have been looking for the right way to onboard P4 (@perforce Helix Core) into your working ecosystem and you did not understand the existing documentation or @JaseLindgren 's YouTube tutorials, this is for you:
deepubuntu.com/blogs/perforcβ¦
We just published a short blog about The State of Data in Autonomous Systems Today. Check it out at:
deepubuntu.com/blogs/state-oβ¦
DeepUbuntu Labs retweeted
Hi,
If you are a decisionβmaker or any member of the technical team at any autonomousβvehicle, robotics, and industrialβautomation companies, pretty pls fill in our survey.
forms.gle/Jvt9Z6Po2CE3N9wPA
.
.
(pleaseπ)
