swyx@swyx人类输入/输出成本高昂，因此在听了 @mattpocockuk 和 @trq212 的建议后，我做了一个 /align-me 修改，支持批量提问而非逐轮问答。这与推测解码的思路一致——通过提前预判 2-10 步来加速。 对设计探索类工作效果极佳。
human i/o is costly, so after listening to @mattpocockuk and @trq212 i made an /align-me modification which allows for batches of questions instead of round-by-round. same intuition as spec decoding, you speed up by looking ahead 2-10 steps
works INCREDIBLY for design explorations
来源：swyx · x.com
人类输入/输出成本高昂，因此在听了 @mattpocockuk 和 @trq212 的建议后，我做了一个 /align-me 修改，支持批量提问而非逐轮问答。这与推测解码的思路一致——通过提前预判 2-10 步来加速。 对设计探索类工作效果极佳。
human i/o is costly, so after listening to @mattpocockuk and @trq212 i made an /align-me modification which allows for batches of questions instead of round-by-round. same intuition as spec decoding, you speed up by looking ahead 2-10 steps
works INCREDIBLY for design explorations
来源：swyx· x.com