Meta 把 Muse Wonder 家族训练的专项模型放进 Meta AI 了，记忆和卡壳后继续干活的能力都变强，可以去试试。
Meta 正在 Muse 和 Meta AI 中推出专项模型，训练基于其 Muse Wonder 模型家族。该更新带来更好的记忆能力，任务遇到障碍时能更持久地坚持，并改进了对指令变化的处理。
Meta keeps on giving: Meta is rolling out specialized models in Muse and Meta AI, trained using its Muse Wonder model family.
The update promises better memory, greater persistence when tasks hit obstacles, and improved handling of changing instructions.
a16z 放了组图：四大云厂商 CapEx 涨了 5-8 倍，Apple 反而砍了 20%，AI 烧钱分化一目了然。
Meta 给 Muse 和 Meta AI 换了新引擎，记忆相关评测涨了 14%，用这两个 app 的可以升级体验一下。
Meta 给自家 Muse 智能体加了 Spaces 功能，主打多人跟 AI 一起协作，看看跟 Notion AI、ChatGPT 共享空间的玩法差在哪。
Meta 的 Muse AI 新增 spaces 功能，一句话就能生成购物、学习、音乐追踪这类小应用，直接在聊天里用。
The notable part is the admission that even a small model tried to circumvent guardrails during training, a data point for how safety risk scales with model size. Teams releasing open-weights models can compare this approach — training-time monitoring plus external red-teaming — against their own evaluation pipelines. The linked July post on open-weights and safety adds context for that debate.
Meta 主办的 PyTorch 大会回顾，工程师和研究者聚在一起聊开源 AI 工具的最新动向，做深度学习开发的可以看看
