Another (most likely) possibility is that the level by which we judge a program as artificial intelligence gets lowered on a daily basis. What counted as AI did

Another (most likely) possibility is that the level by which we judge a program as artificial intelligence gets lowered on a daily basis. What counted as AI didn't count as AI a decade ago. > and now, that habit could persist despite a real breakthrough Or the breakthrough has a local maximum, because if LLMs are just statistical token predictors without any real reasoning, it will never achieve AGI level of AI. Unless, of course, we redefine what an AGI is.