Pinned Tweet
learning has never been easier, it has never been faster, and industry has never been so willing to hire (in our lifetime)
i studied 8 hours a day for ~7 months to prep, and i would do it again.
everything is yours for the taking if you're willing to reach out and grab it
4.7 and 4.8 mightve been the most nothingburger releases of all time. 2018 Shanghai Dragons level of nothingburger
2
170
spacex step up from nvidia? for jensen huang?
Jensen to SpaceX is now a very real possibility. It is the only lateral to step-up move out there. Terafab and orbital data centers are the most challenging and ambitious projects in AI. I'm sure Jensen would love to help tackle them. He would also be a perfect and obvious next-in-line successor to Gwynne, if and when she ever chooses to retire.
77
my company is hosting a corl event. im not even going (havent posted ab it either) and got 10+ data vendor messages abt it
5
194
railway getting absoluted decimated on the tl. terrible time to be a software company.. who could've seen it coming ?
58
i love giving scoped research projects to my agents with some suggestions (how to curate the data, what evals to run, etc) and report/blog writing guidelines. leaving comments everywhere while having my curiosity sated is the autist wet dream
4
157
i love my robots and my robots love me (i reset the scene without them estopped and i havent died yet)
82
looks kinda like code as policy
Quick video showing some of the "Reactivity & Recovery" capabilities of Instant Robot Learning at @Machenta_:
1. One real demonstration: putting blue object in grey box and closing box.
2. Robot reacts in real time when object and box are moved around, with a closed-loop end-to-end policy.
3. Robot recovers when object is pulled out of gripper, and re-grasps it.
4. Robot notices that object is removed from box, and places it back in.
5. Robot notices that box lid is opened, and re-closes it.
The important part:
1. None of this reactivity and recovery needed to be shown in the demonstration: it all comes from the intelligence of the pre-trained model.
2. Model has never seen these objects before. It is trained 100% in simulation, and these objects were not in the simulation.
machenta.com/blog/instant-ro…
1
1
509
excited for my girlfriend to move in so i can truly dedicate myself to my work. long distance is time consuming and requires care, but we both would be significantly more okay with me working until 9pm if we both lived together.
"has argon solved robot manipulation" -- is there any better way to tell every robotics researcher that you have no idea what you're talking about than beginning your post with this sentence?
Has Argon solved robot manipulation? Or do the demos fail as task complexity scales
We benchmark Argon, Astra, etc on complex manipulation tasks on real hardware + 1000 simulation runs
Can they one shot complex tasks? Or is a physical harness and tools via dimOS still needed?
1
116











