In one camp, we have people at organizations like Microsoft/OpenAI, Google, Facebook, and maybe Baidu that have successfully developed and trained large-scale AI systems with hundreds of billions to trillions of trained parameters. People in this camp are not too worried about societal risks, but I wonder if it's because to them the rapid improvement in AI capabilities look like a shiny rainbow with a big "pot of gold" (money, fame, glory, etc.) on the other side.
In the other camp, we have people at other organizations, including every academic institution, who have yet to develop and train a large-scale AI system with hundreds of billions to trillions of trained parameters. People in this camp are writing open letters about societal risks. These people sound sensible, but I wonder if they're worried because the rainbow is protected by a giant wall, and they see themselves as being outside that wall (e.g., they lack the budget).
--
EDITS: Replaced /pot of gold/ with /"pot of gold" (money, fame, glory, etc.)/, which better reflects what I mean.