Neekey Ni
757 posts

Neekey Ni
@NeekeyNi
CTO / Co Founder @cotreat I’m doing Engineering, design, product and marketing














GPT-4 does a lot more than you think, and none of the features are documented (it isn’t even clear how much the creators of these LLMs know about their capabilities) Just ask it to do stuff, and when it says no, ask it if it can figure out any way to do it creatively using code.




🚨Big things are happening in humanoid robotics!🚨 As we saw with drones and quadruped robots, it can take a mere decade for bleeding edge R&D areas to become "solved" platforms for commercial consumer use cases. Once open-ended research questions around reliability, dexterity, and cost have all been answered in a resounding fashion. Today, you can take your pick of low-cost stable consumer options that may have been impossible to imagine just 10 years ago. And now, the writing is on the wall: the next few years will be transformative in bringing general-purpose humanoids from small scale proof-of-concept demos (ie. Atlas in the 2010s) to truly mature products available at scale. 2 notable recent announcements: - Unitree H1 (m.unitree.com/en/h1/). @UnitreeRobotics did for quadrupeds what DJI did for drones. Absolutely believable that Unitree can run it back and disrupt a new form of robots. - AgiBot RAISE-A1 (agibot.com). In case you aren't familiar with the founder 稚晖君 (Peng Zhihui), he is a legit 100x Engineer that recently left Huawei's prestigious "Youth Genius" program to start his own company, AgiBot. I've been following his amazing YouTube (@user-ow7ej5ss7j" target="_blank" rel="nofollow noopener">youtube.com/@user-ow7ej5ss…
) for a while now, and the proof is in the pudding: it seems that Peng is now only able to hack on full-stack robotics side projects, but also ship a real production robot. From inception to an on-stage demo of self-powered bipedal locomotion without support, it took less than 6 months. These recent entrants join other humanoid players that I'm also particularly excited by: - 1X EVE and NEO (1x.tech) @1x_tech is well-positioned with a large head-start in building out an already impressive hardware platform and the buy-in of OpenAI. The long-term AI vision is led by my friend @ericjang11, whose prescient insight around generative modeling, imitation learning, and language-conditioning has influenced many research efforts to scale up smart robots, including those of my own team at Google DeepMind. - Tesla Optimus (tesla.com/AI) @Tesla_AI impressed robotics experts with how quickly they went from marketing pitch to live hardware demo in less than a year. Once they reach some technological maturity, there are clear synergies with their existing supply chain and distribution know-how. The manipulation R&D efforts are in good hands (haha) with @julianibarz, my old manager responsible for creating the famous "Google Arm Farm"! - Figure 01 (figure.ai) @Figure_robot is one of the best-funded players led by a superstar founder @adcock_brett, who has a proven track record of shipping hard tech products (prev. @flyarcher). Long-term AI vision led by @corelynch, my old colleague whose keen research insight created a new subfield of learning from unstructured robot play data. - Clone (clonerobotics.com) @clonerobotics is taking a unique approach with bio-memetic musculoskeletal humanoids. Big bet to swing for the fences, and led by the extremely sharp @dhanushisrad. When everyone else zigs, Clone zags. There are a ton more cool humanoid efforts but these are the ones that I'm particularly excited by, thanks to their focus on building the Robot Brain and not just the Robot Body! Embodied AI is one of the most exciting challenges of our lifetimes, and I'm exceedingly optimistic about building it.








