I asked researchers: Will scaling reasoning models like o3 and R1 unlock superhuman reasoning?

Read this yesterday off twitter. Since o3-mini just came out into developer preview, does anyone else think this discussion from researchers feels like a reach or an overreaction? Comments