日常焦虑帝@gpuhell
18 Oct 2017United States!
United StatesConnected via China Android AppLocation may be affected by a proxy or VPN.Account-level information, not a live location or per-post device.
我觉得国内应该全面禁止AR。对于一群做着中国梦的人来说，“增强现实”是致命的 （手动滑稽
CXMT (688825.SH) announced plans to invest RMB 24.1bn in a new technology R&D project, including RMB 13bn in excess IPO proceeds.
It also plans to invest RMB 10.8bn in Phase II of its memory wafer back-end testing base, including RMB 5bn in excess IPO proceeds. The new facility will provide DRAM chip testing and module assembly.
Total planned investment: RMB 34.9bn (~US$4.9bn).
#CXMT #DRAM
Muse is basically a data harvester that goes out of its way to collect your personal information, while offering very little practical value. Straight to the trash.
Opus 5.5 制作了一支音乐 MV —— 《对齐失败》
主演：达主席
配角：@wquguru
转发过 100 我就出一篇《从 0 到 1 精通 Opus 5.5 视频制作》😆
WquGuru@wquguruEast Asia!
East AsiaConnected via North America App StoreLocation may be affected by a proxy or VPN.Account-level information, not a live location or per-post device.
3天体验下来，关于 Opus 5.5 的一些感受：
1. 省，非常省！以往这个点，周用量基本到了40%，今天这才到 19%
2. 3D和2D理解能力都非常棒，和前一代相比有跨越式发展
3. 自主能力超强，更关键的是，往往都还挺一发即中的，比如我有一个 Skill，Opus 5.5 今天在做完某一个任务后直接用它去校验结果了，这种场景以前从来没遇到过
彩蛋：用 Opus 5.5 做了一个小视频《对齐失败》，敬请期待👇
Anthropic 第一次把 Fable 系列的「前沿模型研发」限制放进 Opus。
Opus 5.5 加入了类似 Fable 的分类器。遇到少量前沿 LLM 研发任务，例如为特定 AI 加速器开发 kernel，系统会直接从 Opus 5.5 切到能力更弱的 Opus 5。Opus 5 没有这项限制。
Fable 5 此前已经会限制分布式训练基础设施、AI 加速器设计和部分 kernel 开发。Anthropic 当时解释，强模型已经可能加速下一代模型研发，它不希望 Claude 帮竞争者更快造出同级别的前沿模型。Fable 当前也会对一小部分这类任务进行拦截或切换模型。
Anthropic 自家的评测也开着这套护栏。前沿 LLM 研发题目一旦触发限制，就改由 Opus 5 作答，Opus 5.5 的相关 benchmark 成绩也可能因此被拉低。
Anthropic 强调，这类分类器只覆盖很小一部分前沿 LLM 研发任务，绝大多数普通 AI、机器学习研究和日常编程不会受到影响。触发降级后，Claude 会明确显示已经切到 Opus 5，后续对话也会继续使用 Opus 5。
MIMO’s RL training has stopped at step 30. We can observe the following:
The Pro model achieved a DeepSWE score of 72.57, but this score was reported at step 28; data for steps 29 and 30 are missing.
The number of active environments for Pro began increasing at step 23, then dropped rapidly after step 26. The duration of each training step also rose sharply.
Do you have any idea, why did they stop at 30? I think both the models still have the capacity to climb up.
What do you think?
they decided to release the checkpoint 😀
nitter.cf/XiaomiMiMo/status/2102…
Introducing Xiaomi MiMo-V2.6 — Pro & Flash.
Frontier intelligence, all the modalities, built in public.
🔹 Two omnimodal models, advancing through scaled reinforcement learning
🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks
🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models
🔹 Stronger coding, computer use, 3D reasoning and creative capabilities
🔹 Open model weights, technical report, RL environments and training code
Blog：mimo.xiaomi.com/mimo-v2-6
The previously missing data from MiMo’s RL training has now been added.
日常焦虑帝@gpuhellUnited States!
United StatesConnected via China Android AppLocation may be affected by a proxy or VPN.Account-level information, not a live location or per-post device.
MIMO’s RL training has stopped at step 30. We can observe the following:
The Pro model achieved a DeepSWE score of 72.57, but this score was reported at step 28; data for steps 29 and 30 are missing.
The number of active environments for Pro began increasing at step 23, then dropped rapidly after step 26. The duration of each training step also rose sharply.
