LLMs have generally sucked the oxygen out of the room, but robotics seems to have been particularly impacted

Been meeting with robotics experts over the last few days, and the mood is gloomy to say the least

People are realizing that Moravec's Paradox[1] means we have a choice between

1/ spending a fortune to automate something that ~any human can do for $10-20 / hr ($2-$5 with teleoperations)

en.wikipedia.org/wiki/Moravec%27s_paradox

2/ spending comparatively less time automating a very high-value activity, which few people can do for $50-$200 / hr

Worse, it feels like very "basic" problems (like grasping) are still unsolved, and the hardware is simply still too expensive and too low reliability

Add to that the fact that hardware is hard (everything that can go wrong goes wrong, and the iteration loops are nightmarishly long), and it's no wonder to see most players reduce investment

Amazon has closed up shop

OpenAI pulled out

MS and Goog significantly reduced