大概是本月10号开始,Antigravity各种模型都会因繁忙类问题中断, 无论是PRO/Flash亦或是第三方模型。 Error: HTTP 503 Service Unavailable Sherlog: Headers: {"Alt-Svc":["h3=\":443\"; ma=2592000,h3-29=\":443\"; ma=2592000"],"Content-Length":["421"],"Content-Type":["text/event-stream"],"Date":["Sat, 18 Apr 2026 15:40:19 GMT"],"Server":["ESF"],"Server-Timing":["gfet4t7; dur=70"],"Vary":["Origin","X-Origin","Referer"],"X-Cloudaicompanion-Trace-Id":["daaacffb07a0b24"],"X-Content-Type-Options":["nosniff"],"X-Frame-Options":["SAMEORIGIN"],"X-Xss-Protection":["0"]} { "error": { "code": 503, "details": [ { "@type": "type.googleapis.com/google.rpc.ErrorInfo", "domain": "cloudcode-pa.googleapis.com", "metadata": { "model": "gemini-3-flash-agent" }, "reason": "MODEL_CAPACITY_EXHAUSTED" } ], "message": "No capacity available for model gemini-3-flash-agent on the server", "status": "UNAVAILABLE" } } 重试几次后可能会磨磨蹭蹭思考一下,然后再次罢工 使用体验极度糟糕。 排查过一些可能的原因,但没有解决: 本人自购新加坡区AI PRO计划。课金取得,非学生免费。 只在单台电脑上登录过,无家庭共享。 网络节点已换过,没区别。 已尝试取消授权并重新登录,无效。 已尝试重新初始化 >Antigravity: Reset onboarding ,无效。 目前有点怀疑是不是有以前用过的节点IP被拉黑,然后账号已经跟着被风险了。 Antigravity出这种恶心问题之后,我也尝试换Gemini Cli,但是遇到了 The caller does not have permission ,网上说是比较常见的问题,所以放弃了。 请问有佬友遇到相同问题吗?谢谢。 3 个帖子 - 3 位参与者 阅读完整话题
Article URL: https://github.com/drasimwagan/mdv Comments URL: https://news.ycombinator.com/item?id=47816629 Points: 2 # Comments: 0
stream disconnected before completion: Transport error: timeout 问一下 最近频繁出现这个问题 自己接的cpa 是vps的问题还是我的网络问题啊 之前好像基本没有这个问题的 这两天好严重 求佬友解答 1 个帖子 - 1 位参与者 阅读完整话题
Article URL: https://solyto.app Comments URL: https://news.ycombinator.com/item?id=47816514 Points: 8 # Comments: 6
Article URL: https://github.com/Nour833/StegoForge Comments URL: https://news.ycombinator.com/item?id=47816491 Points: 1 # Comments: 0
非开发出身一直对Obsidian敬而远之,但最近还是入了Obsidian的坑,先说结论就是舒服极了。Vibe coding 都能上手的人Obsidian 几乎也算是0门槛了。 Notion转Obsidian原因有几个: 1、Notion 现在功能越做越臃肿了,仿佛看到了当年使用印象笔记的影子,这个是我不喜欢的 2、教育账户多少感觉存在风险,笔记日益多了,那天挂了就惨了。 3、AI 好用但也贵,不舍得花钱,也不愿意组别人空间 4、 不得不承认很喜欢notion 的强大数据库多维表格功能,唯一还创建了一个套第二大脑笔记系统,非常舒服,但迁移成本很高。用数据库创建的笔记后面迁移到Obsidian 出现大量格式问题。 5、CC大行其道后,感觉.MD格式文件才能顺应AI潮流,我迁移差不多了就出现Karpathy爆款记录笔记方法,也证明了我的判断。 目前没事就折腾obsidian 的格式主题,怎么弄好看。虽然时间花在这些东西上很不值得,因该去记录,但让自己舒服才符合自己使用习惯。 推荐几个好用的插件吧: #主题 : Minimal 强推! 配合Minimal Theme Settings; #AI: Claudian 强推! Obsidian 中直接操作笔记; 导航管理: notebook navigator 强推! 比Obsidian 原生导航好用一万倍的神奇,如果让我只推荐一个插件那必是它了; 附件管理: Custom Attachment Location 强推! obsidian 笔记文件中包含的图片等格式附件有个专门文件夹管理,这个工具就是随笔记一起管理附件好用 多端同步: Remotely Save 强推! 配合云端工具,实现多端同步 4 个帖子 - 4 位参与者 阅读完整话题
Comments URL: https://news.ycombinator.com/item?id=47816451 Points: 1 # Comments: 0
GAI is a flexible Go library for building agent-style applications on top of LLMs Comments URL: https://news.ycombinator.com/item?id=47816285 Points: 1 # Comments: 0
Article URL: https://github.com/raullenchai/Rapid-MLX Comments URL: https://news.ycombinator.com/item?id=47816238 Points: 1 # Comments: 0
Article URL: https://hack-game-pi.vercel.app/ Comments URL: https://news.ycombinator.com/item?id=47816065 Points: 1 # Comments: 1
还是reconnecting,好像还是去识别的我之前用的公益站添加的api,这应该怎么办呀 2 个帖子 - 2 位参与者 阅读完整话题
我这么大一个 不支持codex,只支持/v1/chat/completions请求 没人看见吗? 后台全是报错。。。。。。 (刚好当鉴别机器人了。。。。。。) 17 个帖子 - 12 位参与者 阅读完整话题
No self-reporting. Only code. Built a platform that tries to infer what you know from your actual work. Comments URL: https://news.ycombinator.com/item?id=47815869 Points: 1 # Comments: 0
https://www.nature.com/articles/s41586-026-10319-8 1 个帖子 - 1 位参与者 阅读完整话题
new_api_panic: Panic detected, error: runtime error: invalid memory address or nil pointer dereference. Please submit a issue here: GitHub - QuantumNous/new-api: A unified AI model hub for aggregation & distribution. It supports cross-converting various LLMs into OpenAI-compatible, Claude-compatible, or Gemini-compatible formats. A centralized gateway for personal and enterprise model management. 🍥 · GitHub | Upstream: {“error”:{“message”:“Panic detected, error: runtime error: invalid memory address or nil pointer dereference. Please submit a issue here: https://github.com/Calcium-Ion/new-api",“type”:"new_api_panic ”}} 2 个帖子 - 2 位参与者 阅读完整话题
概述 感谢各位上次发帖很多佬的交流, 上次是豪华配置,这次测了弱一点的配置 首先期望不要太高,其实这个水平的模型OpenCode还有厉害一点的Minimax M2.5免费用( 虽然刚才出现的Bug Minimax也没修好 能玩,但是上下文看自己的操作,如果有核显则可以拉到100K上下文,没有的话可能20-50K上下文了。( 所以我特别喜欢有核显的电脑 )显存比较紧急的话可以划分1层给CPU,可以拉高20K上下文应该 体感可以编程,没有什么问题 写了个时钟,还有个贪食蛇 相关说明 Apex量化的I-MINI GGUF表现真的很亮眼,损失感觉很小? Qwen3.6 35B A3B的上下文真的好便宜 模型在这里,I-MINI版本就13.3G(这里不加载视觉模块了,显存不够): Qwen3.6-35B-A3B-APEX-GGUF · 模型库 如果有Intel 358H, 338H 32G+1T, 或者AMD 890M, 780M 的用户也可以试试看,内存大可以选I-Compact的17G版本 部署环境 硬件 CPU 12450H 显卡 RTX 5060 Ti 16G 内存 单根 16G DDR4 3200 注意:显卡上没有接任何输出,BIOS设置的核显优先,界面渲染都交给了核显,如果开个渲染个界面可能就剩下13-14G显存,上下文只能开比较少或者拿一层给CPU,decode速度会降低25% 软件 后端 LM Studio 部署模型:Qwen3.6-35B-A3B-APEX-I-Mini.gguf Decode速度: 80tps 层数:全部放在GPU上 上下文:100K 关闭MMAP, 不保持模型在内存中 打开快速注意力,K缓存 V缓存量化均为Q8, Q4好像有BUG → 会导致Prefill非常慢 建议: 用来编程时,如果第一步没能做好,建议直接从第一步重开多试一次,应该会比修bug要好点,改代码bug能力没有写代码能力强的感觉 本地还能玩玩Heretic(虽然这个模型好像没什么感觉,RP不是很好,总之玩玩也不赖) 这个量化确实损失感觉没多少的样子,因为同样概率发生的bug我跑Q6量化的版本也有概率发生 对于天才编程佬们来说,模型的能力还是远远不够的,这篇文章没什么帮助,虽然如此,但是还是想要分享一下 如果发生长时间卡住,可能是模型跑出循环思考bug了,可以中断一下重新跑 题外话 话说L站没有本地部署模型的标签吗(逃 附加截图 2 个帖子 - 2 位参与者 阅读完整话题
TestingCatalog – 17 Apr 26 Exclusive: Early look at Grok Computer and Grok Build xAI is set to launch Grok Build and Grok CLI next week, with Grok 4.3 Early Access already live for Grok Heavy subscribers on web and mobile. Grok Computer is likely planned as an Electron-based desktop app. 2 个帖子 - 2 位参与者 阅读完整话题
I built LogsGo as a learning project to explore log ingestion, querying, and storage tradeoffs. It’s a small Go-based system where logs come in over gRPC, land in memory first, then flush into local storage and optionally S3-compatible object storage. I also added a simple query language plus a small UI to inspect log occurrences over time. This wasn’t built because I think the world needed “another logging system” or because I’m an expert here. I mostly wanted to learn by building something end to end: ingestion paths, storage layering, querying, retention, auth/TLS, and some UI work. Repo: https://github.com/Saumya40-codes/LogsGO I’d genuinely appreciate feedback, including “this design is wrong for X reason” type feedback. If parts of it feel overengineered / naive / badly thought through, that’s useful for me too. Comments URL: https://news.ycombinator.com/item?id=47815402 Points: 1 # Comments: 0
Article URL: https://github.com/Higangssh/pvm Comments URL: https://news.ycombinator.com/item?id=47815391 Points: 4 # Comments: 0
150 applications. One offer. Each application took 5+ manual steps. Separate tools, separate tabs, separate sites — none of them talking to each other. Generic output. Over an hour per application. Paste a job description — or pull it from any job site with the Chrome extension — and five AI agents run an orchestrated pipeline in under 30 seconds: analyzing the role, scoring your fit, researching the company, writing a targeted cover letter, and tailoring your resume to the role. Sequential where it needs to be, parallel where it can be, each agent's output feeding the next. Also includes a dashboard to track every application. And tools for everything around it: interview prep with mock sessions, salary negotiation, job comparison, follow-ups, thank you notes, and references. Runs on your machine. No subscriptions, no data stored on our servers — just your own Gemini API key connecting directly to Google. Comments URL: https://news.ycombinator.com/item?id=47815326 Points: 1 # Comments: 0