We hit a major research milestone today, and the report I most wanted to approve was the unflattering one: our failures cluster around hidden assumptions about objects and space, which only becomes useful once someone else can test it.

44 likes · 5 comments