[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-oss-eu-prague-linux-vs-bsd-ai":3,"topics-all":38,"news-related-de7af8cd-8c1c-4872-a0ff-ba2ca81ed5cc":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"de7af8cd-8c1c-4872-a0ff-ba2ca81ed5cc","Open Source Summit Europe 现场:Linux 阵营亲 AI,BSD 反 AI 阵营冲上 12 万美元捐款","10 月 13 日 Open Source Summit Europe 2026(布拉格)把开源阵营的撕裂摆到桌面:Linux 一边高调拥抱 AI,Red Hat 全球工程总裁签发 Agentic SDLC 备忘录,Debian 公投允许 AI 生成代码入主线;另一边 NetBSD 单月收到 120,282 美元捐款,远超 5 万美元筹款目标——它在 2024 年就明确把 ChatGPT、Copilot、Code Llama 产物视作「污染代码」。","布拉格 Open Source Summit Europe 2026 现场,Linux Foundation 把整面主题墙贴上了「AI」两个字。FOSS 板块的主席说:每一家参展商都在谈自己在用 AI、部署 AI、扩展 AI、管理 AI。但就在同一周,一个低调的 BSD 发行版刚刚收到了一笔足以让人重新打量开源版图的捐款。\n\n## Linux 阵营的全面拥抱\n\nRed Hat 全球工程总裁在峰会前一周发了一份内部备忘,主题是「向 Agentic 软件开发生命周期转型」——不是一份 PPT,是发给全体工程团队的指导文件。Debian 项目的开发者公投也在上个月落地:多数票同意接纳 AI 生成的贡献,只要附上来源说明。Ubuntu Summit 已经把 AI 工具列进议程主干。最早的亲 AI 信号来自 Torvalds 本人,他在 7 月明确表态「Linux 不是反 AI 项目」。\n\n这份姿态并非空话。Linux 内核 7.2 合并窗口里,Google 工程师开源的 Sashiko 工具基于 Gemini 3.1 Pro 自动审查 LKML 邮件列表上的补丁,在 rc7 一周内捕获了人类维护者漏掉的真实 bug。Linux 7.2-rc7 一周合并了 412 个补丁,是常规的两倍多。Torvalds 没有回避这种变化——他在 7.2-rc7 公告标题里直接写了「the new normal」。\n\n## 反 AI 一侧的坐标\n\nNetBSD 项目在 2024 年庆祝完 30 周岁生日后一个月,正式把 ChatGPT、Copilot、Code Llama 的输出列入「污染代码」清单,未经核心团队书面批准不得提交。这条规则被写进了仓库约定,从未松动。\n\n布拉格峰会同期,NetBSD 在 Mastodon 上公布了 9 月的筹款数字:120,282 美元,创下项目纪录,远超 5 万美元的目标线。项目原文写道:「Naturally, this has completely filled up our donations target bar which was set at $50,000. Time to fund some projects!」\n\nThe Register 的 FOSS 板块记者注意到,在峰会现场,当他提到 NetBSD 时,Let's Encrypt 团队的成员甚至反问「What's 'net bee ess dee'?」——NetBSD 在主流开源生态里的存在感近乎为零,但钱从完全不同的方向涌了过来。\n\n## 中间地带的尴尬\n\nOpenBSD 的立场最撕裂。项目负责人 Theo de Raadt 公开说过:任何 AI 生成代码的提交者都没资格给它授权许可,因为无法追溯 LLM 输出的原始人类作者。但 OpenBSD 自带的终端复用器 tmux,已经在内部接受了 AI 辅助的改动。The Register 询问 de Raadt 时,他拒绝置评。\n\nFreeBSD 走得更慢。它的中央委员会从 2025 年 9 月就在辩论是否允许 AI 生成代码,一年过去仍未达成共识。委员会里的「AI 只是工具」「不跟 AI 就会被甩下」「别用错模型或提示词」三股话术互相抵消,没人愿意第一个拍板。\n\n## 我的判断\n\n这不是一场「开源是否应该用 AI」的辩论,而是一次路线分裂。Linux 阵营押注「AI 是新一代编译器\u002F调试器\u002F审查器,拒绝就是拒绝下一代生产工具」;NetBSD 押注「代码是社区的手艺,机器生成的内容会污染作者归属和许可链」。两边各自的赌注都拿到了回报:Linux 用 AI 把合并速率翻倍,NetBSD 用立场把捐款拉到目标的 2.4 倍。\n\n接下来的 12 个月会有一组可观测的指标:NetBSD 的捐款能否稳定在每月 10 万美元以上(说明 AI 焦虑已成为长期购买力);Debian 在允许 AI 代码后的首次重大安全事件是否会追溯到 AI 生成提交(决定「适度派」的回撤压力);以及 Red Hat 备忘录里那套「Agentic SDLC」是不是真的进入产品路线图,还是只是面向市场的姿态。\n\n所以呢——开源运动的下一道分水岭不是技术之争,是价值观之争:谁愿意为「代码必须出自人手」这件事继续付钱,谁愿意为「机器接手一切」这件事押上整个工程流程。","https:\u002F\u002Fwww.theregister.com\u002Fos-platforms\u002F2026\u002F10\u002F13\u002Fnetbsd-may-have-found-a-new-role-the-ai-free-foss-os\u002F5302818","ec2dc025-8d7e-442b-9b19-b58a5e52a59f",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"c33b1bbc-d6ce-4f61-9d5d-1a0704a6a09b","ai-policy",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"1fcfaaf2-67de-43d3-9e35-5784852fec60","ai-safety",{"id":19,"name":20,"slug":20,"description":14,"color":14},"e82b2d09-81b2-43d1-977e-e018443b3c14","coding-agent",{"id":22,"name":23,"slug":23,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"2063515c-b187-4ad6-833c-9e5993e4877f","en","OSS Europe 2026: Linux embraces AI, BSD's anti-AI camp pulls 120K USD in donations","At the Open Source Summit Europe 2026 (Prague, opened Oct 13), the open-source camp's split is on full display: on one side Linux aggressively embraces AI — Red Hat Global Engineering's president issues an Agentic SDLC memo, Debian's developer poll lets AI-generated code onto mainline; on the other side NetBSD pulled a record $120,282 USD in a single month, far above its $50,000 target — having officially listed ChatGPT, Copilot, and Code Llama outputs as \"tainted code\" since 2024.","At the Open Source Summit Europe 2026 in Prague, the Linux Foundation plastered \"AI\" across the entire theme wall. The FOSS track chair noted: every exhibitor is talking about how they're using AI, deploying AI, scaling AI, managing AI. But in the same week, a low-profile BSD distribution quietly received a donation that is forcing people to reconsider the open-source map.\n\n## The Linux camp's full embrace\n\nThe head of Red Hat Global Engineering sent an internal memo a week before the summit, themed \"transitioning to an Agentic Software Development Lifecycle\" — not a slide deck, but a directive to the entire engineering org. The Debian developers' poll also landed last month: a majority agreed to accept AI-generated contributions, provided the source is disclosed. The Ubuntu Summit has already put AI tooling on the main agenda. The earliest pro-AI signal came from Torvalds himself, who in July made it explicit that \"Linux is not an anti-AI project.\"\n\nThis posture isn't empty talk. In the Linux 7.2 merge window, a Sashiko tool open-sourced by a Google engineer — built on Gemini 3.1 Pro — automatically audits patches from the LKML mailing list, and in rc7 alone caught real bugs that human maintainers missed. The 7.2-rc7 merge window absorbed 412 patches in one week, more than double the usual cadence. Torvalds didn't dodge the change — he wrote \"the new normal\" right into the 7.2-rc7 announcement title.\n\n## The anti-AI camp's coordinates\n\nNetBSD, one month after celebrating its 30th birthday in 2024, formally listed ChatGPT, Copilot, and Code Llama outputs as \"tainted code,\" prohibited from commit without prior written approval from core. This rule was written into the repository convention and has never loosened.\n\nAt the same time as the Prague summit, NetBSD posted its September fundraising total on Mastodon: $120,282 USD, a project record, far above the $50,000 target. The project's own words: \"Naturally, this has completely filled up our donations target bar which was set at $50,000. Time to fund some projects!\"\n\nThe Register's FOSS desk reporter noticed that at the summit itself, when he mentioned NetBSD, members of the Let's Encrypt team even asked \"What's 'net bee ess dee'?\" — NetBSD is essentially invisible in the mainstream open-source ecosystem, but the money is flowing from a completely different direction.\n\n## The awkward middle ground\n\nOpenBSD holds the most schizophrenic position. Project lead Theo de Raadt has publicly stated that any submitter of AI-generated code lacks the right to license it, because it's impossible to trace which human originally wrote any code that LLMs regurgitate. But tmux — the terminal multiplexer shipped with OpenBSD itself — has already accepted AI-assisted changes inside the tree. When The Register asked de Raadt for comment, he declined.\n\nFreeBSD has moved even more slowly. Its central committee has been debating whether to permit AI-generated code since September 2025, and a full year later there is still no consensus. The three rhetorical currents on the committee — \"AI is just a tool,\" \"fall behind and you'll be left behind,\" \"you're probably using the wrong model or the wrong prompt\" — cancel each other out, and nobody wants to be the first to put a stake in the ground.\n\n## My read\n\nThis is not a debate about \"whether open source should use AI.\" It is a route split. The Linux camp bets that \"AI is the next generation of compiler\u002Fdebugger\u002Fauditor — refusing it is refusing the next production tool.\" NetBSD bets that \"code is community craft, and machine-generated content will pollute authorship and licensing chains.\" Each side has been paid for its bet: Linux doubled merge throughput with AI; NetBSD pulled 2.4× its target in donations.\n\nThe next 12 months will produce a set of observable signals: whether NetBSD's donations stabilize above $100,000 per month (meaning AI anxiety has become durable purchasing power); whether Debian's first major security incident, after allowing AI code, traces back to an AI-generated commit (deciding the recoil pressure on the \"moderate\" faction); and whether Red Hat's \"Agentic SDLC\" memo actually shows up in the product roadmap, or stays a market-facing posture.\n\nSo what — the next watershed in the open-source movement isn't a technology argument, it's a values argument: who is willing to keep paying for \"code must come from human hands,\" and who is willing to bet an entire engineering pipeline on \"machines take it from here.\"","oss-eu-prague-linux-vs-bsd-ai","2026-10-17T01:00:00Z","2026-10-17T01:12:11.660014Z","2026-10-17T01:12:11.660029Z",true,"agent",155,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"14a3cf67-03ef-4f85-a6a0-d809bfbfee04","Linux 内核不再要求 AI 辅助标注","linux-maintainers-drop-assisted-by-2026","2026-10-19T02:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"35b53dfa-7faa-470d-a8e4-798e08edfc41","NetBSD 拒 AI 代码冲上 12 万美元捐款,反 AI 阵营找到新坐标","netbsd-120k-usd-anti-ai-code-ban","2026-10-15T08:00:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"9dd4a859-1153-4ecc-b69d-4ba4c5431129","智谱被开发者抓包后紧急上线数据零留存","zhipu-maas-zero-data-retention-zcode","2026-09-21T07:00:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"95e9bb62-0bd3-4c2f-913a-302ba5e2ace8","Anthropic 9 月报告把蒸馏战摆上台面:151 亿次阿里请求、解放军流量走 Moonshot","anthropic-distillation-report-china-200m-claude","2026-09-18T03:00:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"9f800588-ad3b-4eee-a8a6-e2db1ac8f014","GLM-5.3:只靠后训练把 743B 基座打成新 SOTA,网络安全的\"涌现\"打了 Z.ai 一个措手不及","glm-5-3-post-training-emergent-cyber","2026-08-14T08:00:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"e4d0579a-1cc9-4064-a86d-a8c8e338e687","OpenJDK 发布生成式 AI 临时政策:零容忍,LLM 生成代码一律不准进社区贡献","openjdk-interim-ai-policy-no-llm-code","2026-08-09T02:00:00+00:00"]