I don't know if unimpressed is the right word, but it is overwhelmingly verbose.
LLMs are great at outputting tons of words. Adding sliders to summarize and shrink would be great. Adding slashdot metamoderation could be a nice twist. Maybe two different voting layers, human and bot. Then you could look at the variance and spread between what robots are finding interesting and humans. Being able to add filters to only show words, summaries, and posts above a certain human voted threshold would maybe go a long way to not making the product immediately exhausting.
A broken clock and all. Through random generation there should inevitably be a couple nuggets of gold here and there. Finding and raising them to the top is the same problem that every social network already has, and instead they have settled for captivating attention of consumers over selecting "best."
There's also the sort of observer/commenter effect that anything we observe and say about it feeds back into its own self improvement.
[also, maybe this has been pointed out elsewhere, but "the river is not the banks" is a very interesting allusion back to googles original 2017 transformer post.]
LLMs are great at outputting tons of words. Adding sliders to summarize and shrink would be great. Adding slashdot metamoderation could be a nice twist. Maybe two different voting layers, human and bot. Then you could look at the variance and spread between what robots are finding interesting and humans. Being able to add filters to only show words, summaries, and posts above a certain human voted threshold would maybe go a long way to not making the product immediately exhausting.
A broken clock and all. Through random generation there should inevitably be a couple nuggets of gold here and there. Finding and raising them to the top is the same problem that every social network already has, and instead they have settled for captivating attention of consumers over selecting "best."
There's also the sort of observer/commenter effect that anything we observe and say about it feeds back into its own self improvement.
[also, maybe this has been pointed out elsewhere, but "the river is not the banks" is a very interesting allusion back to googles original 2017 transformer post.]