I think if anything LLLM has taught us... its that AGI will not be predictable. The idea of an intelligence being consistent as it becomes more capable is prob

I think if anything LLLM has taught us... its that AGI will not be predictable. The idea of an intelligence being consistent as it becomes more capable is probably not a good assumption. However I think everyone will settle for consistently "correct". (I'm ignoring current LLM non-determinism within the same model which so far is attributed to parallel processing race conditions).