全部 AI 动态

What's new in Grok Bot 00:00 New chat cards 00:31 Speed 00:53 Draft edits 01:08 Email aliases 01:33 X DMs 01:55 Tag @bot 02:31 Connectors 02:48 Voice calls 03:27 Notifications 03:56 Invites and Android Assistant 04:19 Access lists, unread, routines, sign-in This team ships!

https://x.com/i/article/2113912265061593088

https://x.com/i/article/2113633101888847874

Local machines had become a bottleneck for agentic mobile development at @Shopify, so we gave agents their own remote dev environments. They now author 61% of our merged PRs 🧵

@making_cindy 在 iPhone Duo 下的表现。配合远程桌面简直无敌。那个半折叠下的触摸板和键盘，是 AI 在适配时的灵光乍现，我非常喜欢。

Team Bots - Grok Bots for your whole team 1:18 A simple example 1:51 Why they're powerful 2:32 Connecting your tools 3:11 How to make one 3:52 A complex example 5:48 When to make one 7:21 Under the hood 7:55 Recap Currently available to Teams & Enterprise customers - more soon!

https://x.com/i/article/2113363792482480128

Super excited to share some of the work I’ve been doing on supporting 1M token context training on SkyRL for GLM 5.3 Flash, 64x higher than the out of the box Megatron baseline! 🚀 This work was done in collaboration with the amazing folks at @LilaSciences! 🤩 Getting there took key algorithmic and systems optimizations in four main areas: 1) DeepSeek Sparse Attention: run DSA on a sparse kernel; chunk and shard its quadratic indexer. 2) Activation recompute and offloading: recompute activations in backward and offload saved inputs to CPU. 3) Context parallelism: split each sequence across GPUs, with new CP support for KDA and a fix for DSA. 4) Avoiding unnecessary precision upcasts: keep FP32 copies out of memory and off the network. For more details on our implementations and our results, check out our full technical blog here: https://blog.skyrl.ai/blog/1m-context-glm-5-3-flash! This work benefited from the contributions of many people along the way. Big thanks to @pcmoritz, @erictang000, @sumanthrh, @CyrusHakha, @richliaw from @anyscalecompute and Keith Truongcao, Tushar Khot, Vivek Subramanian, Semih Kara from Lila Sciences for all the help and guidance!

Grok Bot can now use Calendar, Reminders, Contacts, and more apps on your phone. Rolling out now on iOS and Android.

there's something really cool about asking your agent to draw a statechart (hypothetical example in this image), then using the /create-formal-verification-skill you made with pstack to find bugs through something like TLA+. you can both understand your code better and how the user flow should work, and fix issues users would otherwise see

It has been interesting to see all the applications of @typesafeai's Jev at @sanity_io. Like our new internal agent harness which allows anyone in the company to quickly spin up a bespoke agent. Jev is used for the router and has worked pretty well so far! We (@rostimelk, @even, and I) wrote up an in-depth post about how it's built so you can do it too. https://www.sanity.io/engineering/how-we-built-our-company-agent-harness

Most people use Grok Bot like a chatbot! But it has its own computer and keeps working while you're away! 8 things you can hand off to it today: 1. Tag @bot under any post on X 2. Handle your X DMs 3. Connect Coinbase, Robinhood or eToro 4. Run daily briefings and reminders 5. Compare prices and ask before buying 6. Open GitHub PRs and ship code 7. Prep your meetings from Gmail, Calendar and Slack 8. Sort and clean your inbox And it asks before it sends, posts or buys anything. Which one would you hand off first?

看到这个街头霸王，80 后的童年记忆回来了！ 确实很感慨，09 年毕业后刚学编程时，自己能写个扫雷、写的 1024 就好开心，写出来自己能玩半天 ... 那时记得还是 Flex 和 Silverlight，也是那时开始对前端特别感兴趣，一直到现在。 现在再回看，这些程序，基本都是 Claude Code / Codex 几分钟就能搞定的小玩意，甚至不咋需要思考！ 突然有个想法，我想把自己小时候和上学时觉得好玩的，现在已经失传的小游戏们，重新做一遍，也让儿子这个 10 后知道知道，他爸小时候有多开心！！ 先从我第一个手机 Moto E770 装的第一个 WAP 页游开始吧，记得叫做：打砖块？

LLM Prompt Caching Versus Semantic Caching Caching LLM costs is two different problems: reuse the prefix with prompt caching, reuse the answer with semantic caching, and never confuse one for the other https://open.substack.com/pub/bytebytego/p/llm-prompt-caching-versus-semantic?r=57j98&utm_campaign=post-expanded-share&utm_medium=web

CS 329Z: Engineering AI Agents Stanford / Fall 2026 @stanfordnlp https://cs329z.stanford.edu/ 课程定位：从"模型"到"系统"的工程学 覆盖：简单 LLM 流水线 → 复合 AI 系统 → 自主 Agent。三位讲师的背景也高度互补： @Diyi_Yang（斯坦福 NLP 教授，人机交互与社会计算方向） @michaelryan207（DSPy 核心贡献者，自动评估 AutoMetrics 作者） @jyangballin（SWE-agent / SWE-bench / SWE-smith 作者，软件工程 Agent 领域最重要的研究者之一） # 课程主线：三大工程挑战 贯穿全课的三个核心问题——分解（decomposition）、数据（data）、评估（evaluation）。11 周的内容基本围绕这三条线展开，可以分为五个模块： 模块 1：构建基元（Week 2–3） · LLM 作为构建材料：API/SDK（litellm）、结构化输出、约束生成、解码策略、test-time compute、上下文工程、模型选型与成本/延迟权衡 · RAG：embedding、向量库、分块策略、混合检索、cross-encoder 与 ColBERT 后期交互 · 工具调用：函数调用 API、MCP（Model Context Protocol）、工具设计、代码沙箱、错误处理与重试 模块 2：框架与设计模式（Week 3–5） · 框架层：DSPy（signature / module / optimizer）、LangChain/LangGraph、LlamaIndex，重点是"框架抽象了什么 vs. 你手写了什么" · 设计模式：workflow vs. agent 的分类学，五种可组合 workflow 模式，ReAct / plan-and-execute / reflection，"scaffold（脚手架）本身就是设计决策" · 记忆架构：短期/长期记忆、记忆作为工具动作、文件系统作为外化记忆、跨 Agent 记忆（MemGPT、Mem0、Generative Agents） · 多 Agent 系统：编排模式、handoff 与状态传递，以及一个很有态度的对照阅读——既读 AutoGen，也读《Why Do Multi-Agent LLM Systems Fail?》和 Neubig 的《Don't Sleep on Single-agent Systems》 模块 3：优化（Week 5） · 从提示词到微调的全景：GEPA、MIPROv2、OPRO、TextGrad（提示优化）；LoRA/QLoRA、蒸馏、RLHF/DPO（权重优化）；test-time scaling（推理算力） · 核心问题是决策框架：什么时候优化 prompt、什么时候优化 weights、什么时候堆推理算力 模块 4：数据与评估（Week 6–8）——最有分量的部分 · 数据：trace、demonstration、feedback 三类数据；训练数据 vs. 评估数据；数据飞轮；合成数据；从 Agent 轨迹构建数据集（SWE-smith） · 评估基础：为什么 eval 难；4 元组框架（request / environment / stopping criteria / scorer）；好 benchmark 的性质；tinyBenchmarks · 评估基础设施：三类 grader、LLM-as-judge 的 prompt 设计与已知偏差、pairwise vs. pointwise、非确定性指标 pass@k vs. pass^k、harness 设计 模块 5：安全与前沿（Week 8–11） · 安全：工具访问的隐私风险、prompt injection（含间接注入）、红队、沙箱与权限模型、输出护栏、human-in-the-loop · Coding Agent：SWE-agent、Claude Code、OpenHands 的端到端架构对比 · 主动式 Agent：从 reactive 到 proactive，General User Models（GUM）、Next Action Prediction，以及"Agent 何时应主动、何时应等待"的 mixed-initiative 问题 · 开放问题：多模态/web/计算机使用 Agent、科学 Agent、长时运行架构、生产可观测性（tracing、monitoring、成本管理） # 作业设计：一手建、一手评 HW1：从零构建 Agent 系统（10%） 给定论文库，构建能检索并推理回答科学问题的 Agent。 · Part A：只用 litellm 手写 RAG + 工具调用 + ReAct 式 Agent 循环 · Part B：用 DSPy 重建关键组件，并反思框架抽象了什么 HW2：评估一个 Agent（10%） 给定一个预构建 Agent，设计完整评估套件：代码型 grader、至少一个 LLM-as-judge、用 4 元组框架构建 benchmark 任务、错误分析。