标签: 算力

The data and the anecdata on Jev’s adoption are shocking. Everyone is adopting it. I think it’s a great product, but this is also downstream of the “AI is too expensive/slow” zeitgeist. People are eager to optimize an…

The data and the anecdata on Jev's adoption are shocking. Everyone is adopting it. I think it's a great product, but this is also downstream of the "AI is too expensive/slow" zeitgeist. People are eager to optimize an...

Vercel 创始人 Guillermo Rauch 表示,一个名为 Jev 的 AI 产品正在被“所有人”采用,其真实数据与坊间传闻同样可观;他把这一现象归因于“AI 太贵、太慢”的行业情绪,而非单纯的产品力。

缓存到缓存:LLM 间的直接语义通信(2025)

缓存到缓存:LLM 间的直接语义通信(2025)

Hacker News 上有人提出,如果不同大模型的 KV 缓存可以互相理解,就能让模型之间直接“传缓存”通信,省掉重复预填充和交接延迟。帖子引用了 Ramp Labs 的相关讨论,但目前公开信息显示,这仍是一个概念设想,而非已落地的产品。