gentleman scientist 🤖 @armnet_dev building the world's largest open arm farm 🇫🇮🇬🇧

London, UK
Joined May 2020
A month ago @pravsels and I set out to reproduce @physical_int’s RL Token paper. Today, I am sharing our research notes from the journey we’ve been on. xitter.cf/VilleKuosmanen/status/…
"RL Token" looks like a great and surprisingly simple post-training methodology for optimising robot models for dexterous tasks in the real world! Over the next few weeks, me and @pravsels will be attempting to reproduce the results (& open source the code) Stay tuned 👀
1
12
2
129
22,580
Always love being welcomed home by the beautiful English weather 🏴󠁧󠁢󠁥󠁮󠁧󠁿😍 California was too hot anyway
12
541
SF is awesome but I will make a hot take claim I think we might have passed peak SF? Last summer it felt like the only place you could build a successful AI company was SF. Accelerators rushed to relocate their portcos here But is SF the right place for the age of deployment?
6
26
2,277
Is this how @sundayrobotics transport the robots? Or is Memo driving the van? @chichengcc can you confirm 😄
3
4
123
8,378
next mistral model spotted in the wild 🐈🇫🇷
"How's your diet going?" "Fuck off"
1
9
502
annoying thing about sitting in a cafe next to the window is having the sun move, requiring you to adjust your seat every few mins to stay in the shade of the tree blocking it
1
367
not a problem in london, we never get proper sunshine!
1
1
127
awesome project from Eric!
1/ I trained an AI to be top 1% in the world at Super Smash Bros. Melee by watching 80k hours of human replay. Meet HAL.
1
4
596
Ville🤖 retweeted
thinking of starting a whatsapp group chat to discuss robot finger and gripper designs, who should I invite?
11
1
24
1,704
It’s remarkable how fast the opinion in the ground at SF has shifted since I last visited in August. All everyone wants to talk about is Astra and agentic robotics - world models are out
14
5
1
108
9,579
It feels like hardware is still under-indexed in SF. If you’re looking for unique ideas that won’t get crushed by the next wave of model releases, going deep into the hardware layer feels like the safest bet to me
2
1
1
15
813
It’s maybe hard to remember now, but world models was THE thing in August. I wonder how many startups and researchers changed their entire research directions to fit the trend, and how many now regret it?
10
668
one of the most unintuitive but valuable sales tips I’ve also learned over the past year or so
Founders: Are you on text with your business customers? Met with a startup founder who had just been texting his largest customer (which informed him over text that they were rolling out the product more broadly). Texting with your best accounts is a sign of strong, trusted relationships. In fact, if you're NOT texting, it means that the relationship is not strong enough. Get on text!
2
576
many such cases 🍿
3
31
658
Startup idea - batch inference for AI tokens 90% discount on tokens if you are happy getting back results in 24 hours
4
9
983
Inference bros who wants to build this
1
185
Roboticists want this for data annotation
2
156
cool benchmarking project!
🤖 The RoboEval Diagnostic Bimanual Challenge is live! Track 2 at the #CoRL2026 workshop "From GPU-Accelerated Simulation to Scalable and Generalizable Real Robot Policy Learning". Scored on behavior, not just success. ⏰ Deadline: Nov 1 (AoE) robo-eval.github.io/challeng…
17
1,108
Had stiff shoulders and headache after an 11h flight and not the best sleep. Fixed with 2km on the rowing machine 🚣 Sometimes exercise is all the medicine you need 💪
7
377