Robot Hands for Modern AI and Real Work

contingencies

This seems to be a lot of noise about no commercially applicable outcome. Industrial robots use either pneumatic or magnetic grippers for a reason. Special end effectors are designed when necessary. Humanoids may have been faddish last year, but I think we all realize they're poorly matched in terms of complexity, cost, computational overhead and energy budget to the apparently demonstration-targeted sample tasks of manufacturing (especially the idiotic approach of having a humanoid use human hand tools!) and painting (which derives its value not from mechanism but from creativity, technique and humanistic context). Zooming out a little: while quaint historically and an interesting place, Boston as a city seems to be out on a psycho-social/commercial limb ... very far from commercial relevance on so many axes. This sentiment is echoed by experienced locals. I wonder if Boston Dynamics which apparently took much of its historic funding from the military industrial complex will survive the rise of Anduril: personally, I feel it may have already passed its viral video apex. If humanoids had any value in conflict we'd be seeing them in Ukraine. Their conspicuous absence represents nothing less than the absence of efficacy for any continuation of Boston Dynamics' traditional funding story.

speedbird

Forget the pinky, add a second opposable thumb