@tenobrus I recommend starting to read chyna papers V4's is enough to get the picture They are massively scaling up RL, they've learned the tricks, they have the capital, they expect more compute, and they're good. I mean, really good. There's more good Chinese in China than in the US.
honestly was not expecting that one
Some users suspect China is already secretly ahead in scaling reinforcement learning for AI while deliberately releasing open-source versions of US models.
No Digg Deeper questions have been answered for this story yet.
> There's more good Chinese in China than in the US. I suspect Americans do not grasp this point But it's true. Gone are the days when you cleanly scooped the creme de la creme. There's an absurd firehose of elite Chinese grads, they all need work, and you don't see them in SF
@tenobrus I recommend starting to read chyna papers V4's is enough to get the picture They are massively scaling up RL, they've learned the tricks, they have the capital, they expect more compute, and they're good. I mean, really good. There's more good Chinese in China than in the US.
@teortaxesTex @tenobrus I sometimes wonder if China is already secretly ahead, but they purposely release an open-source version of every US frontier model 2-3 months later just to F..k with American labs.
@teortaxesTex The best europoors go to the US, meanwhile the best Chinese stay in China.