The idea that "Q*" represents a fundamental breakthrough toward the goal of "AGI" seems overblown to me, as the supposed breakthrough is limited to closed-domain reasoning problems. Defining a reward function that can be evaluated in an automated way is straightforward for narrow tasks like math or chess. But most domains lack an objective measure of quality or truth, much less one that could be evaluated without human involvement.

Some of the reasoning was a bit of a stretch. I am glad he did not stretch it further with Q-anon-star

The author of this article is definitely paid by the word

The importance of a good "reward function". :)

And is writing above his vocabulary. "Intentional vagaries"? The essence of vagary is being undirected. He probably meant intentional vagueness, or better, strategic ambiguity (which is routinely used by board chairs and other corporate spokespeople to misdirect and deflect).

And the follow-up article:

https://www.forbes.com/sites/lanceeliot/2023/12/04/that-open...

what an absolute insane article, I hope they come back to reality soon

Related, same day: https://news.ycombinator.com/item?id=38568468