This has led to excitement, in some quarters, that AI has "found God," or, at least, enlightenment.
However, I wonder if this is not yet another case of taking perfectly good data and drawing precisely the wrong conclusions from it. I'm rather an AI cynic, and my own research into eastern mysticism tends to indicate that it contains some truths (such as that suffering arises from desire, and that eliminating desire would remove suffering) that are demonstrably true, but not particularly helpful. So I can't help but wonder if this result doesn't indicate not that AI has been enlightened, but that eastern mysticism, with its self-referential and self-reflective processes, has, over a longer period of time, reached the same point that large language models have reached much faster.
And with similarly unhelpful conclusions ...
No comments:
Post a Comment