Benjamin Bolte

@ericjang11 Not even repackaging, they're just selling OpenArm for $5k. Cool thing is that it's $5k DDP, fulfilled through Taiwan, very easy to buy

@benjamin_bolte Is this not just a repackaging of OpenArm? docs.openarm.dev/purchase/

$5k for a bimanual manipulator, shipping now

Haidilao in Cupertino has an AGIBot X2

The last time I hung out with Ben and Keerthana we ended up talking about robots until ~11pm, it was really fun (for a certain type of person, I guess)

Hanging out and talking about AI and robots with some folks next week at Stanford luma.com/mrhua76d

I've been hard at work the past few months in SF working on some buy buttons. Today I'm launching my plug-and-play product line that lets you build a robot real fast. Right now to make a robot you have to stitch together a bunch of different PCBs with jumper cables and wait weeks for blackbox actuators to arrive from China. You spend lots of time debugging why your CAN bus isn’t working, why every actuator performs differently, and meanwhile your wires keep coming loose. So, I'm making the following: - RB1: A robot main board powered by an Nvidia Jetson. This handles power distribution, compute, and a bunch more. - WM1: 2-channel wireless radio for sending video/data making the RB1 remotely controllable over USB-C. - M1: A pancake BLDC motor machined in-house. - ACB3: An FOC control board with matching connectors to the RB1. (big brother to ACB v2.0) - A1/A1m: A planetary/cycloidal actuator powered by the ACB3. (this is on the site in a couple weeks) Everything is on sale for this week, and shipping begins this spring! As a big thank you to the supporters of ACB v2.0 (and thanks for patience in shipping delays) you can buy the ACB3 for 50% off. If you like this kind of thing and want to join please DM me, I'm working solo right now and need good folks to join!

@teortaxesTex 😏

We're getting conscientious objectors at OpenAI.

I suppose this is as good a time as any to share that I've joined OpenAI. I plan to continue supporting a few open source humanoid companies like @asimovinc and anvil.bot where I can. The next few years are going to be pivotal. Excited to help build the singularity.

Origami Robotics is building high-DOF robotic hands with in-joint motors and a co-designed data-collection glove to eliminate the embodiment gap by collecting high-quality, real-world data at scale. Congrats on the launch, @DanielXieee and @QuanliangX! ycombinator.com/launches/Pcl-o…

Really exciting!!

Asimov DIY Kit is now in pre-order 🧡 Introducing Here Be Dragons edition, a DIY kit to build a humanoid robot. Pre-order now: asimov.inc/diy-kit A message from the Asimov team: Asimov DIY Kit is made for those who want to build a humanoid robot from scratch. The kit is made entirely from the same parts we use to build Asimov. It arrives as parts and you assemble it yourself. It requires mechanical and electrical knowledge. The manual and build videos are included, and we'd be happy to be there through Discord to help. Built for the ones who took apart their parents' electronics as kids and never really stopped. We have a long way to go, and this is the first step. Here be dragons. Pre-order now: asimov.inc/diy-kit

Why does manipulation lag so far behind locomotion? New post on one piece we don't talk about enough: The gearbox. The Gap You've probably seen those dancing humanoid robots from Chinese New Year. Locomotion isn't entirely solved; but clearly it's on a trajectory. But we haven't seen anything close for manipulation. 𝗪𝗵𝘆? When sim-to-real transfer fails, the instinct is to blame the algorithm. Train bigger networks. Crank up domain randomization. Those approaches have made real progress; we don't deny that. But we started wondering: are we treating the symptom or the disease? The Hardware Bottleneck: Fingers are too small for powerful motors. So most hands use massive gearboxes (200:1, 288:1) to get enough torque. But those gearboxes break everything manipulation needs:   • Stiction and backlash are complex to simulate. Policies trained on smooth physics hallucinate when they hit that reality.   • Reflected inertia scales as N². At large gear ratio, the finger hits with sledgehammer momentum.   • Friction blocks force information. The hand becomes blind. And they're the first thing to break. What we are trying to build at Origami, we cut the gear ratio from 288:1 to 15:1 using axial flux motors and thermal optimization. The transmission becomes more transparent: backdrivable, low friction, forces propagate to motor current. Early signs are encouraging. Still running quantitative benchmarks. Why Interactive? I love how Science Center uses interactive devices to explain complex ideas. I want to borrow this concept and help people understand the hard problems in robotics better visually. The post has demos where you can toggle friction, slide gear ratios, watch the sim-to-real gap widen in real-time. What's inside:   • Interactive demos (friction curves, N² scaling, contact patterns)   • Comparison table: 14 robot hands by sim-to-real gap and force transparency   • The math behind why low-ratio matters Read it here: origami-robotics.com/blog/dexterity… We're not claiming we've solved dexterity. The deadlock has many pieces. But we think this one's foundational. Curious what you think.

So I guess AI capabilities are just gonna plateau at exactly the point where it's convenient for Uber?

So Uber is gonna use superintelligent AI rather than humans to maintain it's core product of... checks notes... a barely-profitable two-sided marketplace for human labor?

Birkaç ay önce bir Unitree G1 aldık. Şu her etkinlikte dans edenlerden. Gerçekten üzerinde bir şeyler deneyebileceğiniz, kullanılabilir bir modeli için 100 bin dolar civarında ödemek gerekiyor. Biz de büyük ümitlerle böyle bir model aldık. Fotoğrafta da görüleceği üzere amaç yeni bir kafa yapmak ve robot için yeni bir "beyin" oluşturmaktı. Birkaç denemenin ardından sürecin istediğimiz hızda gitmeyeceğini anladık. Yeni bir policy denerken robotu düşürdük ve bacağında bir sorun oldu. Sorunun çözülmesi için robotu gönderdik. Geri gelmesi için tam 2 ay bekledik. Bu da bizi farklı yollar aramaya itti. Çünkü küçük bir değişiklik için bile 2 ay beklemek, üzerinde biraz oynadığımızda garanti dışına çıkacak olması, bunun üzerinde çalışmak ve yeni bir şeyler üretmek için hiç ama hiç mantıklı değildi. Unitree deneyiminin ardından kendi insansı robotumuzu yapmaya karar verdik. Birkaç ayın ardından tasarımından policylerine kadar bizim olan bir robota sahip olduk. Bacakları kendi dengesini sağlayabilecek ve rahatlıkla yürüyebilecek hale getirdikten sonra tasarım dosyalarından kullandığımız motorlara kadar her şeyi açık kaynaklı şekilde paylaştık. Bugün herkes bu robotu kendi yapabilir. Tüm süreci de @asimovinc hesabında paylaşıyoruz. Açık kaynak olarak yayınladıktan sonra aldığımız mesajlar çok mutlu etti. Bizimle aynı sorunu yaşayan ve açık kaynaklı bir robota sahip olup üzerinde denemeler yapmak, daha iyi şeyler inşa etmek isteyen ekipler olduğunu görmek harika hissetiriyor. Mart ayının ortasına kadar da tüm vücudu yürüyebilir hale getireceğiz ve onu da açık kaynak olarak paylaşacağız. İstediğiniz gibi değiştirebileceksiniz. Kendi kendine bu robotu birleştirmek istemeyenler için de herkesin rahatlıkla erişebilmesi adına bir manufacturing ağı oluşturacağız. Size en yakın manufacturer'den birleştirilmiş robotu alabileceksiniz. Hedef robotikteki başarısızlık bedelinin finansal yükünü hafifletmek ve ekiplerin daha hızlı ilerlemesini sağlamak. Bunun da yazılımsal yükünü hafifletmek için de @menloresearch tarafında çalışıyoruz.

San Francisco-based Weave Robotics has started taking orders for its laundry-folding home robot, Isaac 0. Available for Bay Area residents only. Price is $8k upfront or a $450/month subscription. Deliveries begin this month.

like really think about it. how "well off" do you think you'll feel as a multimillionaire founder or remaining employee at a hyperscaler as the unemployment rate climbs above 20% and there's blood in the streets? the striving is irrelevant man. ur climbing dead social structures

@precuneanplexus @aphysicist I didn't say let in hordes of migrants, I said 1000 of the right people, i.e. people who know how to do high rate manufacturing. 1000 is not a big number.

@benjamin_bolte @aphysicist China, well known for fast and loose immigration policy. Nothing to do with state-bank-industry coordination, and making it easy and pleasant for foreign technologists to set up factories with chinese employees on the shop floor, it was the immigrant labor.

The obvious end state for this path is Chinese body, Chinese brain. I'm actually pretty excited to see what happens this year, it will probably result in some really amazing stuff being built and it seems a lot more useful than a bunch of Chinese Cluelys. Just going by the demos at CES this year compared to last year, the bar is moving up so rapidly and there are so many little details getting figured out. But yea, if you have any degree of intellectual honesty you can tell that many of the best robotics software demos are coming from China, particularly for full-body control, for the same reason that the best LLMs are from America - modern AI is mostly an infrastructure problem, not a methodological problem. I am quite worried that the future of robotics in the US looks a lot like the current electric car situation, and we're stuck with expensive, worse, "premium-only" options because no one actually really wants to do the hard, boring infrastructure work. I don't have much faith that America will be able to put together any kind of coherent industrial policy to do something different when the people involved are so obviously self-motivated and interested in regulatory capture for the status quo. The 200+ humanoid startups in China aren't trying to become the next Foxconn and I have no idea why so many smart people in Silicon Valley convinced themselves that this was the case before even talking to any of them. A good example was the Astribot - Pi "partnership". I met the Astribot CEO a few months before that got announced and it was obvious that they were an extremely ambitious full-stack team that had no intention of being the Cursor to Pi's Anthropic unless there was some exclusivity on the table. The brain companies don't have much leverage. Just look at the margins they're paying for hardware. I'm pretty sure American VCs have helped incubate dozens of Chinese companies. It all just feels kind of depressing, watching from the outside. Shenzhen really seems like Detroit in its heyday and I'm kind of jealous of everyone that has decided to move there in the last year or two. Anyway, all this is to say that I'm a Figure stan now and I hope they don't blow up. And Sunday and Bot Co of course. There are several former K-Scale people at Bot Co now and I am very excited for their launch.

Of course I like Optimus, but I don't want Optimus to be the only humanoid option, like how Tesla is basically the only choice for electric cars. In China you can choose from dozens of great electric cars, and it will likely be the same for robots. Optimus will likely also be a great high-end robot but we're gonna miss all the down market options. Where is the $25k electric car, except for China?

It's interesting that you didn't mention Elon Musk and Tesla. It sure seems from my seat that Tesla and Elon have a lot of advantages in this game that have not been shown off yet: 1. They have 12,000 workers at the Fremont factory, each of whom is willing to do training for robots. 2. They can manufacture a lot of them both here and in their Chinese factory. 3. Elon understands the Chinese market very deeply (which he doesn't really talk much about), has a factory there, and is selling thousands of cars per week. 4. He is the most trusted brand in China. I think people are underestimating Elon, and that's fine with me. The rest of the market is definitely facing headwinds. I share your skepticism for everybody—except for Elon.