Latest

Every published item, in order · 29226 items
Nov 10
Chinese contentKylon 把 Claude Code 和 Codex 接入 3,000+ 工具,智能体已处理 500 万笔交易

如果你还觉得 Claude Code 只能写代码,这条值得看看:Kylon 把它和 Codex 接进 3,000 多个工具,让智能体跑营销、销售、运营,已处理 500 万笔交易,约等于 5 万小时人工。想试多智能体协作的团队还有首月免费。

elvis@omarsar0Source
ProductsBuildersSelected05:01
CoreWeave Skills brings Claude Code GPU cluster provisioning

The notable shift is that an agent now owns the whole GPU serving pipeline, including the verification step that checks real utilisation, which most infrastructure templates skip. Developers deploying open models like those served by vLLM can test CoreWeave's stack without writing Kubernetes manifests, making it a practical comparison point against managed offerings from AWS and GCP. Teams evaluating GPU cloud costs and setup overhead should try the preview from the GitHub repo.

CoreWeave@CoreWeaveSource
Chinese contentMeta 发布 Muse Ember:对标 Claude Haiku 5.5,成本约三分之一

Meta 这个 Muse Ember 是专门给 agent 干杂活用的:工具调用、代码子任务、路由抽取这些高频小任务,价格只有 Claude Haiku 的三分之一左右。做 agent 流水线的开发者可以直接算一下迁移成本,省的钱是实打实的。

rohanpaul_ai@rohanpaul_aiSource
Chinese contentClaude Fable 5.2 上线 Vercel AI Gateway,可编排云端沙箱并行任务

Rauch 提前试了 Claude Fable 5.2,说这代模型能并行干远超一台机器的活。他给了个很简单的玩法:Vercel CLI 加 /offload-to-sandbox 技能,智能体自己开和关云端沙箱,不用写复杂的多智能体编排框架。做智能体工作流的人可以直接抄这套方案。

Guillermo Rauch@rauchgSource
Tomorrow 08:00, the daily lands in your inboxTelegram channel or email. Same screening as this page, no ads.
Chinese contentVercel 为其 MCP 加入 Events 事件订阅功能

Rauch 这条推文把 MCP 一个长期痛点讲得很清楚:以前智能体只能轮询等结果,现在 Vercel MCP 直接支持事件订阅,部署成功或失败会主动推 webhook 通知。你自己想搭 MCP 的话,他用 Queues 和 Workflows 给了现成方案,做智能体后端的开发者可以直接抄。

Guillermo Rauch@rauchgSource
PapersBuildersSelected04:32
Stability AI-backed SF Research launches Interactive Visual Gym simulator

Simulator-based evaluation sidesteps the cost and unpredictability of human-subject studies, giving labs a repeatable way to measure assistant quality before deployment. Teams building real-time visual guidance products can use IVG to benchmark models and generate training data cheaply, as the 4B model's gain demonstrates. Researchers in multimodal agents and embodied AI should compare their systems against the 14-assistant baseline it establishes.

Official accountSalesforce AI Research@SFResearchSource
Industry74°04:31
Chinese contentFireworks 遭黑客入侵,客户账单信息与模型权重泄露

Fireworks 被黑客偷了客户账单、模型权重和数据集,年化收入 10 亿美元的它正要按 300 亿美元估值融资,这时候爆出安全事故,投资人尽调肯定要盯紧。用托管推理服务的企业可以核对一下自己的数据有没有暴露,同行也该看看自家风控。

The Information@theinformationSource
ModelsBuildersSelected04:28
Fable 5.2 first impressions: 100+ generations tested against Opus

Benchmark runs of this size are rare in the first days of a release, so the category-level verdicts carry more signal than typical launch impressions. Teams building generative UI, games, or data visualization can use the results to decide whether switching from Opus justifies the cost delta. The video format also shows actual outputs side by side, not just scores.

lmarena.ai@lmarena_aiSource
Chinese contentAnthropic 的 Alex Albert 分享一条模型验收提示词:让新模型通读项目找问题

Anthropic 员工 Alex Albert 的一条实用提示词,让新模型通读你的项目代码库,修掉有把握的问题、其余列出来给你审。比起跑基准分数,用真实项目验收新模型更直观,写代码的人都可以直接抄走用。

Alex Albert@alexalbert__Source
Chinese contentClaude Fable 5.2 上架 Arena.ai,Battle Mode 与 Agent Mode 可直接试用

Anthropic 新模型 Claude Fable 5.2 现在在 Arena.ai 上就能试,还开了 Agent Mode。有人拿它跑了 100 多次生成,跟 Opus 比 3D 场景、小游戏和 UI 生成,哪边强、哪边贵都讲了。想选模型写代码或做页面的开发者可以看看再决定。

lmarena.ai@lmarena_aiSource
Chinese contentAnthropic 发布 Claude Fable 5.2 和 Mythos 5.2:能力超 Opus 5.5,API 降价 20%

Anthropic 这次一次发了 Fable 5.2 和 Mythos 5.2 两个高端模型,编程评测分数超过 GPT-6 Astra 和 Gemini 4 Carbon,API 还降价两成。Pro 用户这一周能免费用 Fable,写代码的可以趁机试试它到底比 Opus 5.5 强多少。

宝玉@doteySource
Chinese contentClaude Fable 5.2 上线 Arena,可测试智能体任务表现

想直观感受新模型实力的话,现在可以直接在 Agent Arena 里拿真实任务测 Claude Fable 5.2,它能调搜索、文件系统和终端干活。得分出来后可以和 Arena 上其他模型横向对比,做智能体开发的开发者值得先跑几个任务试试。

lmarena.ai@lmarena_aiSource
How-toBuildersOfficialSelected04:20
OpenAI dot workflows for testing iOS apps across Macs

The subtle point is that dot's value here comes from owning a second machine, not raw model quality — delegating tests to separate hardware avoids the cursor and focus contention that plagues single-machine agent use. iOS developers juggling multiple SDK versions and QA leads comparing regressions should try the cross-machine repro workflow first. Compared with manual device farm setups, this needs only a spare Mac and an OpenAI account.

Primary source@OpenAIDevs@OpenAIDevsSource
PapersBuildersSelected04:18
Microsoft BOXBOX lets agents hot-reload their own harness mid-run

The non-obvious shift here is treating the agent harness as a live, mutable component rather than a fixed artifact frozen before deployment. Teams running multi-hour agent workloads — long coding jobs, research pipelines, infrastructure automation — can apply this pattern to let the agent adapt its own scaffolding without losing session state. It is most relevant to agent framework builders who currently restart or re-prompt when harness limits surface mid-run.

DAIR.AI@dair_aiSource
ProductsBuildersSelected04:17
Claude Fable 5.2 lands in Perplexity Computer and Search

The per-task cost figure of US$14.51 with a 0.851 score gives buyers a benchmark for comparing agentic model economics across providers, not just raw accuracy. Teams evaluating AI agents for research and browser automation can use Perplexity Computer as a hands-on testbed. Access is split by subscription tier, so Pro users get Computer only while Search requires Max — a pricing detail that matters for budget planning.

Official accountPerplexity@perplexity_aiSource
Chinese contentYC 创业公司 Archgen 用 3D 堆叠晶体管设计芯片,宣称降低 40% 功耗

这家叫 Archgen 的 YC 公司把晶体管往楼上盖,而不是继续往小里做,同一份活最多省 40% 电。他们的 AI 温度预测模型比传统软件快 5000 倍,工程师画芯片时就能当场看到哪里会过热。做芯片设计或关注后摩尔时代路线的人可以看看。

rohanpaul_ai@rohanpaul_aiSource
Chinese contentGenspark Media Studio 支持一句话生成完整影片

Genspark 新上了 Media Studio,一句话就能跑通从故事大纲到分镜再到剪辑的整条流程,每个环节你都能审批修改。比起 Sora 那类直接出片的工具,它更像给你配了个能商量着来的剧组。做短视频或品牌内容的人可以试试。

Genspark@genspark_aiSource
Chinese contentNuance Labs 推出 Emma:基于 Nuplex 1.0 全双工模型的双向实时 AI 视频人设

Nuance Labs 搞了个能边听边说的视频角色 Emma,回复时机 89% 对齐真人。有意思的是她的表情来自一位真实演员的录制数据,眨眼、眼神这些小动作都是一个人的。做数字人或虚拟主播的可以看看思路。

rohanpaul_ai@rohanpaul_aiSource

Only the latest 2000 items are listed; for older ones see the daily archive