马斯克表示 Grok 4.8 的预训练已完成，目前进入 RL 阶段，评测指标开始上升，本周晚些时候将向 Bot 用户软发布。他称团队弃用 JAX/SGLang，用 C/C++ 从零重写了整个 RL 和推理栈，并清理了 SFT 数据；Grok 4.8 与 4.7 将没有相似之处。东锡 NLP 转发该内容并评论称，马斯克对训练技术细节的描述很专业。
马斯克披露 Grok 4.8 的训练与 RL 进展，读者可了解其底层技术栈重写和本周软发布安排。
Grok 4.8 就在眼前。
SpaceX AI 的 RL 和 inference infra 全部重写。
马斯克对训练技术细节描述的太准确了，像是一名专业的模型训练人员说的话。在这样的老板手下干活，啥鱼也别想摸了咱们这边就是说。
【引用 @elonmusk】We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++, which took some extra time.
Training (aka pre-training) is done. Lot of cleanup was needed on the SFT data to get that mostly right. Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes.
Grok 4.8 will bear no resemblance to 4.7 (thankfully). It is not yet incredible, but, this time, the foundations are built right, so I expect it to evolve to be something special.
We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++, which took some extra time. Training (aka pre-training) is done. Lot of cleanup was needed on the SFT data to get that mostly right. Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes. Grok 4.8 will bear no resemblance to 4.7 (thankfully). It is not yet incredible, but, this time, the foundations are built right, so I expect it to evolve to be something special.在 X 查看被引用的帖子
来源：Dongxi 东锡 NLP · x.com