Chubby♨️ · @kimmonismus · X·2026-09-22 06:48·2小时前
AI 导读

据 The Information 报道,一名 OpenAI 员工称 OpenAI 内部已基本实现新实验模型训练流程的自动化。引用推文提到,OpenAI 新训练的模型解决了 100 个长期未解的数学难题,连其研究人员都感到意外;同期 MiMo 发布新模型,Grok 4.7 表现不及预期。

Chubby♨️@kimmonismus
41AI 编辑部评分,满分 100
2026-09-22 06:48· 2小时前
AI 导读

据 The Information 报道,一名 OpenAI 员工称 OpenAI 内部已基本实现新实验模型训练流程的自动化。引用推文提到,OpenAI 新训练的模型解决了 100 个长期未解的数学难题,连其研究人员都感到意外;同期 MiMo 发布新模型,Grok 4.7 表现不及预期。

„Internally, OpenAI has largely automated the process of training new experimental models, the OpenAI employee said.“

Via the information

We did it friends. Recursive self improvement is here.

Chubby♨️WTF I was on a plane for 10 hours. Within 10 hours OpenAI released a blogpost that their newly trained model solved 100(!!) open long standing math problems tha...

来源:Chubby♨️· x.com