When you contrast that with Boston Dynamics (who I've admired since their initial buzzing dog videos), they are optimizing for fluidity of motion and lifelike kinematics. But their programming is AFAIK completely based upon heuristics, which means they will never be able to scale in a complex environment, like a factory. They are as dumb as the Kuka robotic arms. In other words, the robot will unlikely make a decision to do a back flip in response to an environmental change, unless programmed specifically to do so in very specific and rigid conditions.
To me, this makes Tesla's "useful first" approach vastly superior in the long term.
Ok I don't think I've ever replied to my own post, but I have to clarify what I said. I dug a little deeper into Boston Dynamics and feel like I've oversimplified their work.
BD's target customer historically was the military. The buzzing dog model was intended to help transport cargo in terrain where wheeled vehicles would struggle, like a dense forest. They've since made newer models that can carry more while being significantly quieter (no more buzzing). They also have mini robots intended to be close-contact recon/spy devices. And while they never push this idea in their YT videos (likely due to sparking Terminator fears), their humanoid robots may have soldier-like applications, where agility and parkour-like abilities would be extremely helpful.
So, BD robots are designed to be useful in narrow military applications, where the primary product advantage is to spare humans from a high-risk environment. They are also pushing the physical capabilities of robots, like faster locomotion, better balance, etc. Some of their robots have some capability to understand their surroundings, like Spot, the yellow dog robot. This robot is used primarily for reconnaissance and can be useful for military and commercial applications. This robot has potential to be useful in a factory setting, perhaps enticing Hyundai to buy BD back in 2020.
So the main takeaway is that BD is also building useful robots, and they've advanced the field of robotics significantly, but their robots' usefulness tends to be narrower in scope than what Optimus aspires to. Tesla is intentionally focusing on the humanoid form, particularly on the usefulness of the hand, since we have one of the most dexterous hands in the animal kingdom. Coupled with a decent brain/AI, there's a lot of potential for usefulness.