Strengthening the association between “rationalist” and “furry” decreases the probability that AI research organizations will adopt AI safety proposals proposed by “rationalists”.
Strengthening the association may enable a larger slice of the rationalists to think and communicate clearly without being bogged down by professional constraints. I suspect professionalism is much more lethal than I think most people think, so that might be a crux. If we lighten the pressure towards professionalism, people have more slack and are less likely to end up optimising for proxies such as impressiveness, technicality, relevancy-to-other-literature, “comprehensiveness”, “hard work”, etc.
Strengthening the association between “rationalist” and “furry” decreases the probability that AI research organizations will adopt AI safety proposals proposed by “rationalists”.
The poster is currently a resident at OpenAI on the reinforcement learning team.
And?
..
..
..
Just joking! I’m joking, sorry!
*pulls on rainbow dash costume*
Strengthening the association may enable a larger slice of the rationalists to think and communicate clearly without being bogged down by professional constraints. I suspect professionalism is much more lethal than I think most people think, so that might be a crux. If we lighten the pressure towards professionalism, people have more slack and are less likely to end up optimising for proxies such as impressiveness, technicality, relevancy-to-other-literature, “comprehensiveness”, “hard work”, etc.