>In fact there's an almost culty refusal to take embodiment seriously as a bedrock factor in consciousness.

We're researching artificial intelligence, not artificial consciousness. People who are building AI systems largely don't care about the question of consciousness, because there's no evidence to suggest that it's in any way necessary to do the kind of things we want to do with AI.

The people who are very concerned about issues like consciousness and embodiment are in no way being prevented from running their own experiments, but they aren't running many experiments and those experiments they do run largely aren't particularly successful or interesting. The experiments of people who are just training models to do stuff are generally very successful and very interesting.