开启时会模糊预览图，关闭后正常显示
I still think about my manager from my first proper job who was the age then that I am now, his calm humour and almost unconscious care for users and staff (his anecdotes were certainly deliberate, to share values). Lots to learn, years later
Maker of things like Poem/1 (AI hardware), Galactic Compass (iPhone app), founder Acts Not Facts. Blogger. Behind the mountains there are more mountains
AI 能提升下限，但是如果自身水平不足，看不出来生成内容的问题，但如果专业的人去用 AI 生成内容，再去基于 AI 内容二次创作，反而能兼顾创作速度和质量，就像优秀的程序员借助 AI 写代码可以又快又好一样。
以前看过的一个很好的比喻就是把 AI 作品和手写的比做工业制品和手工制品，其实工业制品整体水平已经高过人工制品的平均水平了，大众也不排斥，但好的手工制品能卖好的价钱。 大众不喜欢的是低劣的工业制品，或者工业制品伪装成手工制品。
True!
intercomputer realist @a16zcrypto / editor https://t.co/fGE7XDfsKo / cofounder @FortuneCrypto
borgesian writer? "you know the babel library, the one with infinite books. turns out we need that for ai training (don’t ask why). i’m writing them all"
Reasoning models coming (very) soon. Co-founder @pleiasfr
blessed days when 100 likes were a banger for me
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
Will be crazy if the optimal expert (ie even when we're optimizing for training FLOPs to downstream perf target, not lifetime costs) is <1M parameters. But also, will be logical. Tiny models are surprisingly strong we just need better routing I kind of predicted this btw
blessed days when 100 likes were a banger for me