Founder @ Narsil | Building robots for a better future | @lifeatpurdue engineer

Joined December 2024
no one is doing 3d printed tires like we are.
86
99
8
1,413
48,588
I used to pray for texts like this
8
299
chat, do i open or leave it sealed?
3
19
741
who said robots don't get nervous during demos
A humanoid robot cutting a cucumber with a knife 🤖🔪
255
ali retweeted
We had so many people asking us “why build the robot, why not just build the software?” When we started @AMRwelding Glad we ignored them
Technical moats are mostly over at this point. If you’re building something that is meant to be durable, the only moats now are: • Proprietary data • Network effects • A physical component
9
3
109
3,573
We're hiring, and this time we're on WaterlooWorks! If you're an exceptional person looking to work on one of the hardest open problems on earth - building reliable, deterministic, and verifiable AI that actually works in the real world - we're probably the team you're looking for. If being on a team with: - Chess Grandmasters - ARCT Level 9 pianists - Debate champions - Patent holding Engineers - Software engineers writing papers at 15 - Physicists, Geneticists, & Engineers (yes, even our admin guys fall in this bucket) - Niche micro-celebs (bonus if you get this one) - Ex Pro Gamers - Pro Balisong flippers (this one is a bit of a stretch) Is a crowd that reminds you of home, check out the roles below! (debatably higher talent density than @cognition) Hardware Engineer (Intern, Winter 2027): -> Turn an F150 Raptor autonomous -> Draft CAD designs for a retrofit kit that turns ANY vehicle autonomous -> Work on cutting edge technology both in commercial, and defence applications Software Engineer (Intern, Winter 2027): -> Help optimize programmatic policies to control real world machines -> Run experiments on various RL methods like VPG, and PPO (we do real RL here) -> Build out software and embedded systems to deploy on real world hardware powered by our AI Software / Machine Learning Engineer (Fulltime): -> work on novel evolutionary algorithms, genetic programming techniques (e.g CGP) and neurosymbolic methods -> work on search algorithm discovery, and optimization -> leverage techniques from topology, differential geometry, and physics to guide program evolution the right way -> hop in the trenches and work miracles alongside the rest of the team Feel free to dm, or apply on WW!
🤖 Made with AI
8
3
60
4,444
This is a clear violation of Asimov's third law
F.02 Decommission
4
489
Great engineering is indistinguishable from art
Liftoff of Starship's first orbital flight
1
7
395
The machines are becoming better at understanding and responding to the physical world. That is the technology story. But technology alone will not determine where robotics spreads first or fastest.
20
14
2
85
11,144
Robotics companies have all accepted The Bitter Lesson and we now see awesome demos, but most have no idea about The Brittle Lesson and boy are they in for a surprise when trying to productize.
23
28
9
405
45,839
The world is in a “space race” for energy that started about 2 years ago. As AI and robotics improve, the fundamental input to the economy becomes energy. The 20th century was won by steel. The 21st century will be won by energy, tokens, and robots. Valar Atomics will deliver.
43
107
13
1,999
50,107
For most industrial applications, robots don’t need legs. The reality is that most industrial spaces are already designed and optimized for wheeled equipment (carts, dollies, forklifts, pallet jacks, etc.), so legs on robots add a lot of unnecessary complexity and cost: - Higher constant power consumption (robots use energy just to stand and balance) - Lower stability and less precise end-effector positioning. - Greater safety risks around humans (e.g., if power fails, the robot can collapse) - More expensive (leg) actuators, since leg joints require significantly higher torque than arm joints
Do humanoids need legs? An estimated 85% to 90%+ of U.S. manufacturing facilities are single-story, and over 90% of operational floor space features step-free, ramped, or ADA-compliant transit paths. Advocates will say wheeled bases are cheaper and reduce tipping risk What do you think?
6
3
1
40
9,075
r2d2 was far more useful than c3po in literally every situation
Wheeled robots are an utter dead end
14
5
1
201
19,027
insane that you can just buy a whole robot dog today for the same price as a macbook pro
5
2
15
872
we don’t talk nearly enough about boston dynamic’s sand flea robot
159
372
84
7,322
694,834
robotics is great because everytime you solve a problem, you get 5 new problems
18
25
2
312
10,647
highly recommend applying if you want to work with a cracked team and build sick drones in America. pictures don't do justice to how awesome these drones look in person.
Man who runs US drone factory (me) just closed fresh funding to build high performance AI drones. Senior EE, new grad EE, and Andriod engineers (kotlin) are urgently wanted. Order book is packed and our customers need new capabilities fast. (1/2)
2
31
2,129
robots in the real world coming soon
4
2
45
1,205
chat, we’re cooking
sim training today to reduce the sim2real gap. isaacsim's renders are incredible
2
87
7,547
sim training today to reduce the sim2real gap. isaacsim's renders are incredible
3
1
1
23
8,779
dont have a great history with running sims but we keep trying xitter.cf/robotsailor/status/202…
im never doing robot simulations again
3
562