Sphoenix

87 posts

Sphoenix banner
Sphoenix

Sphoenix

@SphoenixAI

Robotics. AI Health & Safety. Editor-in-chief @BotNewsAI

Burlingame, CA Katılım Mayıs 2022
145 Takip Edilen17 Takipçiler
Sabitlenmiş Tweet
Sphoenix
Sphoenix@SphoenixAI·
"It looks right" and "it is right" are not the same claim, and most of the painful lessons in this field live in that gap. SC3-Eval sits right on top of it. The pitch is to evaluate robot policies inside a generated video instead of on a physical arm, but generated video drifts. A forward-only model can continue to produce plausible frames long after the action behind them has gone wrong. Looking harder at the picture doesn't help. So they stop asking the model to look forward and make it run the story backward: which action would have caused this footage? When the answer stops matching the command, the rollout has started lying, and you can catch it the moment it does. Covered it in my article for @BotNewsAI #ai #tech
Bot News@BotNewsAI

x.com/i/article/2068…

English
2
4
14
1.9K
Sphoenix retweetledi
Bot News
Bot News@BotNewsAI·
The transmission inside 1X NEO's new hand shares a family tree with bicycle brakes and surgical endoscopes: tension cables running through compression sheaths, the only routing that carries force across a 3-DoF wrist without caring what the wrist is doing. 1X bet the whole architecture on paying friction tax at a discount: gear ratios of 5:1 to 15:1 where the industry runs past 100:1, low enough that motor current still carries the shape of whatever the fingers are touching. The cinematic footage shows wine glasses and LEGO Duplo but not the force residual at full wrist articulation, six months into a kitchen. We restated the whole field in one accounting standard, priced the parts list on both sides of the Pacific, and traced the tendon loops the launch didn't diagram. Corrections welcomed. #information #robotics #ai #robot #1x #neo
Bot News@BotNewsAI

x.com/i/article/2075…

English
0
2
5
303
CIX 🦾
CIX 🦾@cixliv·
REK has been secretly building the world’s first humanoid fighting simulator. Opening today for beta on Steam. Using the same AI and physics as a real robots. This will allow millions of people to train to become robot pilots. A video game literally becoming reality.
English
57
76
580
162K
Sphoenix
Sphoenix@SphoenixAI·
@deCamaret @chris_j_paxton Shouldn’t be tele-op. Atlas is most likely autonomously executing a control policy (along with motion planning and feedback control). Running three robots this smoothly would actually be easier and more consistent than trying to teleoperate all three simultaneously.
English
1
0
0
30
Sphoenix
Sphoenix@SphoenixAI·
@alexandr_wang @Suhail You should DM your engineers that have been trying to get ahold of MSL all month. 🫢
English
0
0
3
893
Alexandr Wang
Alexandr Wang@alexandr_wang·
@Suhail hey very sorry to hear this is your experience. DMing you but would love to hear what the setup was and figure out what’s happening here
English
18
2
464
57K
Suhail
Suhail@Suhail·
My evaluation of some of the models over past ~24 hours testing/building: GPT 5.6 Sol is excellent. You can feel the care in it. Bravo as usual OpenAI. muse-spark-1.1 is way off the mark after my testing of some real questions. Bad formatting, not thorough, doesn't dig deeper, seemingly afraid to respond at length, too utilitarian. Something is off. Cursor Grok 4.5 has been joyful to use. I ask it lots of side questions while Claude codes. The fast mode is a real delight each day. It really understands, chomps through the code base, and has gotten me to trust it more each day.
English
53
43
1.3K
178.4K
Milo Smith
Milo Smith@mil000·
finally bought a new wallet
Milo Smith tweet media
English
8
0
24
5.7K
Sphoenix
Sphoenix@SphoenixAI·
Your team has nailed content, drops, collabs, hardware, aesthetic, and overall it’s really obvious you guys understand how to hire people who aren’t purely academic on paper but genuinely have an eye for trends in our industry and how industries around it move. It produces incredible achievements like this that are the all encompassing package of presentation and proof.
English
0
0
9
282
Bernt Bornich
Bernt Bornich@BerntBornich·
NEO's finally learning sign language
English
123
268
3.2K
221.2K
Sphoenix
Sphoenix@SphoenixAI·
NEO would rather squad up in Fortnite than fold his laundry
English
0
2
5
63
Chris Paxton
Chris Paxton@chris_j_paxton·
The robot as the new home computer
English
2
6
143
20.5K
Sphoenix
Sphoenix@SphoenixAI·
@asimovinc Omgoodness your team is amazing for putting these together like this! Probably need a theragun and some Motrin after 😅
English
0
0
2
74
Asimov
Asimov@asimovinc·
We're forming Batch 1, and we do batching a little differently. We ship in groups of 10 and build a mini community around each batch. You work alongside your batch-mates, same kit at the same time, debugging the hard parts together on Discord and live streams. If you pre-ordered, you've got an email from us. Here be Dragons.
Asimov tweet media
Asimov@asimovinc

Asimov 1 is an open-source humanoid robot you can build and customize yourself. Two ways to get one: 1) Source the parts yourself: docs.menlo.ai/asimov/1/bom 2) Get the DIY kit: asimov.inc/diy-kit The kit bundles every part as a group buy, cheaper than sourcing one by one, and you build alongside others.

English
15
18
309
23.6K
Sphoenix
Sphoenix@SphoenixAI·
The most valuable part of a robot demo may not be the motion anymore. It may be the part of the object it taught the world how to hold. CHORD starts as a dexterous manipulation result and ends as a new accounting system for robot data. On the surface, NVIDIA’s paper reads like another human-video-to-robot-manipulation pipeline: recover the demonstrated contacts, retarget the motion, train the policy. But the reward does not really score the hand. It scores the deal the hand made with the object. A contact point is just an address. It does not tell you whether the finger pinned the object, slid it, tipped it, rotated it, or quietly kept the whole motion from falling apart. The useful thing lives in the contact normal, the moment arm, and the friction cone: the directions that grip can push and twist the object. Pressure never enters the reward, which is the tell. Video cannot reliably recover human squeeze anyway. So CHORD scores wrench support instead. Not “did the robot touch where the human touched?” but “can this grip create the same menu of forces and torques on the object?” That means a human hand and a robot claw can share almost no contact geometry and still count as the same demonstration. Different fingers. Different body. Same object-side authority. Copy the contact positions and articulated objects collapse to 0.000. Score the wrench support and they reach 0.914. That unwelds the demo from the body that recorded it. Once a manipulation demo survives embodiment transfer, it stops being a clip. It becomes inventory: something you can stockpile, rank, license, and rent back to the next policy that needs to know how the world can be held. Somebody is quietly building the object library every future robot policy ends up renting from. Covered it in my article for @BotNewsAI. #ai #tech
Bot News@BotNewsAI

x.com/i/article/2071…

English
0
5
6
578
Sphoenix retweetledi
Sphoenix retweetledi
Zhi (Leo) Wang
Zhi (Leo) Wang@TX_Leo_Wang·
1/🧠Humans are the best robot data source — but video alone misses one thing: force. 2/🙁Tactile gloves capture force — but they're costly and block the real touch manipulation depends on. 3/💪Maybe the future of touch lives on your wrist: surface EMG reads the muscles that cause force — tactile sensing without ever touching a tactile sensor. 4/🔥Want a fully open-source framework — hardware + software — to train your own force-aware learn-from-human-data robot policy? 🚀We introduce ForceBand: Learning Forceful Manipulation with sEMG -- bring force into human videos with sEMG, for force-aware manipulation ⬇️ ✦ Zero-Shot Human-to-Robot Transfer ✦ Force Beyond Vision ✦ Free-Hand Force Sensing ✦ Collect by Anyone, Anytime, Anywhere ✦ Deploy on Any Robot, Any Camera, Any Environment ✦ Open-Source & Low-Cost & Easy-to-Implement Let's squeeze every bit of signal out of human data, and let robots feel the force! 🌐 Website: forceband-emg.github.io 📄 Paper: arxiv.org/abs/2606.26093 💻 Code: github.com/Bottle101/Forc… 🎥 Video: youtube.com/watch?v=Otw6uX… 🧵 1/n
YouTube video
YouTube
English
11
45
203
49.5K
Sphoenix
Sphoenix@SphoenixAI·
@DoWCIODavies Yay! So excited for this program. I’ve been checking every day.
English
0
0
3
82
DoW CIO Kirsten Davies
DoW CIO Kirsten Davies@DoWCIODavies·
I’m excited to announce we will soon launch the pilot of the Cyber Registered Apprenticeship Program! If you are interested in applying, please set up an account with USAJOBS, get your USAJOBS resume ready now, the initial application window will not stay open for long. ⏱️ Don’t wait a day more! Monitor the following channels for the most up to date information: 🌐DoW Cyber Apprenticeship web page – the link to the apprenticeship opportunity will be posted here when it is live. 🦅USAJobs - Set an alert for cyber apprenticeship. 📧If you have already emailed us, we will notify you by email when the announcement opens 👀Follow us on our socials - we will post the link to the Apprenticeship Opportunity once it is live! Thank you so much for your interest in this new program! 🇺🇸Don’t miss this incredible opportunity to join the DoW’s cyber workforce! For more information check out: dodcio.defense.gov/Cyber-Workforc…
DoW CIO Kirsten Davies tweet media
English
4
2
13
920
Sphoenix retweetledi
Yoonseok Pyo
Yoonseok Pyo@passionvirus·
Launched the official webpage for 'AI Sapiens', @ROBOTIS' open-source humanoid! Explore resources on hardware, software, and RL-based motion generation. Some sections are "Coming soon" and will release in phases. Check it out: docs.robotis.com/docs/systems/a…
Yoonseok Pyo tweet mediaYoonseok Pyo tweet mediaYoonseok Pyo tweet media
English
5
29
104
10.6K
Sphoenix
Sphoenix@SphoenixAI·
@ryancjulian I want to frame this. I’ve been building an automated TPM because 👀
English
0
0
0
79