4 ms·
I think LLMs could definitely help us stop treating large piles as "lists" (to be completed). You can just "dialogue" with it through Chat. And if the AI has ac
by CrypticShift 4y ago
I think LLMs could definitely help us stop treating large piles as "lists" (to be completed). You can just "dialogue" with it through Chat. And if the AI has access to your recent activities or notes, it can even give you relevant choices. Or you can "navigate" through it in an interactive 2D/3D map that clusters the article/books by (semantic) similarity.
So dialogue and navigation take the place of checking a list.
- mlyle 4y agoThough... It just deepens the problems of curation of all the content everyone consumes through relatively opaque and unpredictable means.
- CrypticShift 4y agoYes, more opaque and unpredictable for sure, but I don’t think it deepens the problems of curation. I believe it is solution. : you trade predictability/transparency for smarter /non-linear curation. You can always choose.
- mlyle 4y agoI just worry about the biases in models becoming our biases, and self-perpetuating.