I feel like the true challenge in making humanoid robotics may not actually be the technology, but the economic viability of the proposed business model.

The robot can do more of course, but at least in my understanding there's something missing in the business model. Super great design in terms of making the robot safe and approachable though, I really do respect the work they've done overall.

One exception I'm aware of is 1X. They're working on a daily assistive robot, to fold laundry and wash dishes I think? This one also doesn't make sense me - who would spend $20,000 on a robot when a dishwasher can get the job done for a few hundred?

Personally, I like Boston Dynamics' robot Handle in this use case: www.youtube.com/watch?v=-7xv... Why limit yourself to the human form-factor?

Companies like Figure, Tesla, Agility Robotics etc. have I suppose correctly ascertained that the manufacturing space already spends a lot on setup costs for e.g. pick-and-place machinery. Since factories are designed for humans to work there, is the most effective shape for a machine human-shaped?

🤔

What's the current status in terms of explainability for RL-based solutions? Do traditional approaches still offer a greater advantage in highly regulated environments?