Artificial General Inference #AGInf is emerging now. Models do not “understand” like humans, they can generalize across tasks by statistical inference at massive scale by compress data into probability fields and can jump domains through transfer learning.
Due to a double exponential curve, inference quality and coverage scales faster than expected due to compute + data + architecture compounding.
So #AGI is science fiction. #AGInf is the reality: machines that can infer broadly without understanding.
Food for thought; Artificial General Inference #AGInf systems that assign probability ratings to inferred states across unknown or partially known domains, will have much bigger #ForceMultipliers
