{"slug": "sashiko", "name": "Sashiko", "category": "ai", "category_name": "AI models and assistants", "description": "", "recommenders": 3, "commenters": 7, "warners": 1, "score": 2.691, "rising": 1.67, "first": "2026-03-18", "latest": "2026-10-15", "why": "I'm very glad they're not spamming the mailing list.", "why_is_quote": true, "has_verdict": false, "sentiment": {"recommend": 2, "neutral": 3, "compare": 1, "warn": 2}, "spark": [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2], "listed": true, "loved": [], "caveat": "", "short_quote": "I'm very glad they're not spamming the mailing list.", "kind": "", "icon": true, "domain": null, "github_repo": "github.com/sashiko-dev/sashiko", "official_url": "https://github.com/sashiko-dev/sashiko", "pricing": "", "aliases": ["Sashiko", "sashiko"], "verdict": null, "praise": [{"label": "", "summary": "", "quotes": [{"id": 801733232, "item_id": 50108327, "url": "https://news.ycombinator.com/item?id=50108327", "by": "jameshilliard", "date": "2026-10-15", "stance": "recommend", "quote": "The kernel now has AI tooling heavily integrated into patch review workflows as well with quite good results.", "excerpt": "> Early-2025 AI This one especially really just shows they are cherry picking things which support a particular position regardless of the relevance, the difference between Early-2025 AI and Late-2026 AI is absolutely massive. Early 2026 it was reported by Greg Kroah-Hartman that AI bug reports rapidly went from obviously wrong or low quality to being real reports.[0] The kernel now has AI tooling heavily integrated ", "context": "", "use_case": "AI kernel patch review", "switched_from": "", "story": "Libreboot 26.10 \"Ominous Olive\" released with UEFI, Chromebook support"}, {"id": 800246400, "item_id": 50015400, "url": "https://news.ycombinator.com/item?id=50015400", "by": "eru", "date": "2026-10-09", "stance": "recommend", "quote": "Sashiko is a great system", "excerpt": "You can easily burn through lots and lots of money on DeepSeek, if you do eg large scale code reviews. Eg I've used Sashiko locally for Linux kernel code reviews before sending out my contributions out to the world. Sashiko is a great system, but it can burn through tokens like there's no tomorrow. https://github.com/sashiko-dev/sashiko and https://sashiko.dev/", "context": "", "use_case": "Linux kernel code review", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 758860512, "item_id": 47428782, "url": "https://news.ycombinator.com/item?id=47428782", "by": "rwmj", "date": "2026-03-18", "stance": "recommend", "quote": "I'm very glad they're not spamming the mailing list.", "excerpt": "Better to link to the site itself, or one of the reviews? For an example of a review (picked pretty much at random) see: https://sashiko.dev/#/patchset/20260318151256.2590375-1-andriy.shevchenko%40linux.intel.com The original patch series corresponding to that is: https://lkml.org/lkml/2026/3/18/1600 Edit: Here's a simpler and better example of a review: https://sashiko.dev/#/patchset/20260318110848.2779003-1-lijun01", "context": "", "use_case": "AI Linux kernel code review", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}]}], "caveats": [{"label": "", "summary": "", "quotes": [{"id": 797129744, "item_id": 49820609, "url": "https://news.ycombinator.com/item?id=49820609", "by": "eru", "date": "2026-09-23", "stance": "warn", "quote": "on its Claude, Vertex and Gemini backends, a connection that drops mid-response is treated as a permanent error rather than retried.", "excerpt": "Accepted by maintainers so far: In mainline: ext4, xattr cache: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=54b6bd40898de ocfs2, cluster accounting: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=621c2bcb87548 Taken into Jan Kara's filesystem tree this week, all isofs: An out-of-bounds read in the multi-extent directory walk: https://lore.kernel.org/all", "context": "", "use_case": "", "switched_from": "", "story": "Data-only attacks are easier than you think (2024)"}, {"id": 758858512, "item_id": 47428657, "url": "https://news.ycombinator.com/item?id=47428657", "by": "monksy", "date": "2026-03-18", "stance": "warn", "quote": "done on an automated basis is a bad idea", "excerpt": "I think this is a great and interesting project. However, I hope that they're not doing this to submit patches to the kernel. It would be much better to layer in additional tests to exploit bugs and defects for verification of existance/fixes. (Also tests can be focused per defect.. which prevents overload) From some of the changes I'm seeing: This looks like it's doing style and structure changes, which for a codeba", "context": "", "use_case": "agentic code review for kernel", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}]}], "who": [], "tags": [], "use_cases": ["ai kernel patch review", "linux kernel code review", "ai code review for linux kernel", "ai linux kernel code review", "agentic code review for kernel"], "compared": [], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=47427647", "title": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel", "mentions": 5}, {"url": "https://news.ycombinator.com/item?id=50102923", "title": "Libreboot 26.10 \"Ominous Olive\" released with UEFI, Chromebook support", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50000488", "title": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49811429", "title": "Data-only attacks are easier than you think (2024)", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49803892", "title": "Claude Opus 5.5", "mentions": 1}], "sentiment_by_month": {"2025-11": {}, "2025-12": {}, "2026-01": {}, "2026-02": {}, "2026-03": {"warn": 1, "neutral": 3}, "2026-04": {}, "2026-05": {}, "2026-06": {}, "2026-07": {}, "2026-08": {}, "2026-09": {"compare": 1, "warn": 1}, "2026-10": {"recommend": 2}}, "quotes": [{"id": 801733232, "item_id": 50108327, "url": "https://news.ycombinator.com/item?id=50108327", "by": "jameshilliard", "date": "2026-10-15", "stance": "recommend", "quote": "The kernel now has AI tooling heavily integrated into patch review workflows as well with quite good results.", "excerpt": "> Early-2025 AI This one especially really just shows they are cherry picking things which support a particular position regardless of the relevance, the difference between Early-2025 AI and Late-2026 AI is absolutely massive. Early 2026 it was reported by Greg Kroah-Hartman that AI bug reports rapidly went from obviously wrong or low quality to being real reports.[0] The kernel now has AI tooling heavily integrated ", "context": "", "use_case": "AI kernel patch review", "switched_from": "", "story": "Libreboot 26.10 \"Ominous Olive\" released with UEFI, Chromebook support"}, {"id": 800246400, "item_id": 50015400, "url": "https://news.ycombinator.com/item?id=50015400", "by": "eru", "date": "2026-10-09", "stance": "recommend", "quote": "Sashiko is a great system", "excerpt": "You can easily burn through lots and lots of money on DeepSeek, if you do eg large scale code reviews. Eg I've used Sashiko locally for Linux kernel code reviews before sending out my contributions out to the world. Sashiko is a great system, but it can burn through tokens like there's no tomorrow. https://github.com/sashiko-dev/sashiko and https://sashiko.dev/", "context": "", "use_case": "Linux kernel code review", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 797129744, "item_id": 49820609, "url": "https://news.ycombinator.com/item?id=49820609", "by": "eru", "date": "2026-09-23", "stance": "warn", "quote": "on its Claude, Vertex and Gemini backends, a connection that drops mid-response is treated as a permanent error rather than retried.", "excerpt": "Accepted by maintainers so far: In mainline: ext4, xattr cache: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=54b6bd40898de ocfs2, cluster accounting: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=621c2bcb87548 Taken into Jan Kara's filesystem tree this week, all isofs: An out-of-bounds read in the multi-extent directory walk: https://lore.kernel.org/all", "context": "", "use_case": "", "switched_from": "", "story": "Data-only attacks are easier than you think (2024)"}, {"id": 797007504, "item_id": 49812969, "url": "https://news.ycombinator.com/item?id=49812969", "by": "gwd", "date": "2026-09-23", "stance": "compare", "quote": "My prompts are moving in the other direction as sashiko", "excerpt": "My prompts are moving in the other direction as sashiko [1], a managed pipeline developed for the Linux Kernel mailing list like a year ago. But last year's models needed a lot more structure and guidance; the results I posted are from the \"single prompt\" version of the same thing. The README [2] describes the difference. You can browse the contents to get an idea; basically all the prompts were actually written and ", "context": "", "use_case": "", "switched_from": "", "story": "Claude Opus 5.5"}, {"id": 758920688, "item_id": 47432543, "url": "https://news.ycombinator.com/item?id=47432543", "by": "nurettin", "date": "2026-03-18", "stance": "neutral", "quote": "", "excerpt": "Of course I specifically avoided invoking that language's name within the context of kernel programming in fear of summoning a Linus. And he's so right. I didn't think like that back then, but new/delete (which have to be overloaded for kernel) behind allocators behind containers, vtables, =0, uninitialized members, unhandled ctor errors, template magic, \"sometimes rvo\", compiler hints, \"sometimes reinterpret cast\", ", "context": "", "use_case": "AI code review for Linux kernel", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}, {"id": 758868240, "item_id": 47429265, "url": "https://news.ycombinator.com/item?id=47429265", "by": "ChrisArchitect", "date": "2026-03-18", "stance": "neutral", "quote": "https://github.com/sashiko-dev/sashiko", "excerpt": "https://github.com/sashiko-dev/sashiko (https://news.ycombinator.com/item?id=47427996)", "context": "", "use_case": "", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}, {"id": 758860512, "item_id": 47428782, "url": "https://news.ycombinator.com/item?id=47428782", "by": "rwmj", "date": "2026-03-18", "stance": "recommend", "quote": "I'm very glad they're not spamming the mailing list.", "excerpt": "Better to link to the site itself, or one of the reviews? For an example of a review (picked pretty much at random) see: https://sashiko.dev/#/patchset/20260318151256.2590375-1-andriy.shevchenko%40linux.intel.com The original patch series corresponding to that is: https://lkml.org/lkml/2026/3/18/1600 Edit: Here's a simpler and better example of a review: https://sashiko.dev/#/patchset/20260318110848.2779003-1-lijun01", "context": "", "use_case": "AI Linux kernel code review", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}, {"id": 758859616, "item_id": 47428726, "url": "https://news.ycombinator.com/item?id=47428726", "by": "rwmj", "date": "2026-03-18", "stance": "neutral", "quote": "", "excerpt": "No, it's reviewing patches posted on LKML and offering suggestions. The original patch posted corresponding to your link was this, which was (presumably!) written by a human: https://lkml.org/lkml/2026/3/9/1631", "context": "", "use_case": "", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}, {"id": 758858512, "item_id": 47428657, "url": "https://news.ycombinator.com/item?id=47428657", "by": "monksy", "date": "2026-03-18", "stance": "warn", "quote": "done on an automated basis is a bad idea", "excerpt": "I think this is a great and interesting project. However, I hope that they're not doing this to submit patches to the kernel. It would be much better to layer in additional tests to exploit bugs and defects for verification of existance/fixes. (Also tests can be focused per defect.. which prevents overload) From some of the changes I'm seeing: This looks like it's doing style and structure changes, which for a codeba", "context": "", "use_case": "agentic code review for kernel", "switched_from": "", "story": "Google Engineers Launch \"Sashiko\" for Agentic AI Code Review of the Linux Kernel"}]}