|
in my testing, deepseek crushes western models on ancient chinesephilosophy and literature, while also having a much stronger commandof english than my first-hand chinese sources. it feels like communingwith literary/historical/philosophical knowledge across generations thati never had access to before. it's quite emotionally moving. it makessense too, western labs dont care about training on chinese data (butchinese labs train on both).remember that china has several thousandyears more of literary history than the west does (because we lost themajority of our roman/greek/egyptian literature and china preservedtheirs). basically; our Al models are missing the literary foundations ofwestern thought, but the chinese models have theirs intact. this couldbe both a classical "data advantage" and a less-obvious advantage forspiritual and philosophical self-actualization. |
|