<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>llmnews.dev</title>
    <link>https://llmnews.dev/</link>
    <description>Models, local LLMs, agents, tools, and research — in one feed.</description>
    <language>ru</language>
    <lastBuildDate>Mon, 14 Sep 2026 20:45:01 GMT</lastBuildDate>
    <atom:link href="https://llmnews.dev/feed.xml" rel="self" type="application/rss+xml" />
    <item><title>Ask HN: Is your AI LLM based web service secure from prompt injection?</title><link>https://news.ycombinator.com/item?id=49703186</link><guid isPermaLink="false">4166393af5a59c46b2a52e8f6c4a3bb2f6fab464172d805bf3cfaf86c083f7d5</guid><description></description><pubDate>Mon, 14 Sep 2026 20:08:25 GMT</pubDate><category>security</category><source url="https://news.ycombinator.com/item?id=49703186">Hacker News</source></item><item><title>Making a vintage LLM from scratch; Take #2</title><link>https://crlf.link/log/entries/260911-1</link><guid isPermaLink="false">f1807deb5fffc4f3cbd0eca424354670d1e5ddeeda0821ee7bed020285f5e6d4</guid><description></description><pubDate>Mon, 14 Sep 2026 19:52:34 GMT</pubDate><category>industry</category><source url="https://crlf.link/log/entries/260911-1">Hacker News</source></item><item><title>b10969</title><link>https://github.com/ggml-org/llama.cpp/releases/tag/b10969</link><guid isPermaLink="false">62bfed8e2bcae889bb33f434c2cdc7ce63fe8f469b6520919e0ec409df623e67</guid><description>ci : add ubuntu-cuda builds to release (#28186) release : add ubuntu-cuda build job (12.8/13.3, x64+arm64) Add GCC 14 for CUDA arm64 builds in CI Eplicit bash Install git for CCCL fetch Install git before we clone/checkout Match CI names for WIndows Whitelist llama.cpp repo to git Use $GITHUBWORKSPACE Also ship dependent libs on Ubuntu Need NCCL additionally</description><pubDate>Mon, 14 Sep 2026 19:50:18 GMT</pubDate><category>hardware</category><source url="https://github.com/ggml-org/llama.cpp/releases/tag/b10969">GitHub · ggml-org/llama.cpp</source></item><item><title>Show HN: Build2me – Prove2Me for software agent swarms build code</title><link>https://github.com/shitianfang/build2me</link><guid isPermaLink="false">233beee97dcdd76a74432ee0e52228f10e717aab0464ef0dad9bba3f5b95c340</guid><description></description><pubDate>Mon, 14 Sep 2026 19:49:56 GMT</pubDate><category>agents</category><source url="https://github.com/shitianfang/build2me">Hacker News</source></item><item><title>Show HN: Slowave – local adaptive memory for coding agents</title><link>https://github.com/slowave-ai/slowave</link><guid isPermaLink="false">efc65f9f557bfe8a4bbde09a824749343fc30cecd4a2eec60ac1df448c496b04</guid><description></description><pubDate>Mon, 14 Sep 2026 19:49:12 GMT</pubDate><category>agents</category><source url="https://github.com/slowave-ai/slowave">Hacker News</source></item><item><title>Agentic AI cringe wars [video]</title><link>https://www.youtube.com/watch?v=U-Rqv9dOB1U</link><guid isPermaLink="false">f0cc214228053ad65528dbdebf757980088c72b2eb3c1340a49857bd777483ac</guid><description></description><pubDate>Mon, 14 Sep 2026 19:49:03 GMT</pubDate><category>agents</category><source url="https://www.youtube.com/watch?v=U-Rqv9dOB1U">Hacker News</source></item><item><title>DeepSeek down for API/app and web</title><link>https://downdetector.co.uk/status/deepseek</link><guid isPermaLink="false">684b833884b201d535dfeb298d3642fcf54e84311a844e2c7224a29cd68c21ed</guid><description></description><pubDate>Mon, 14 Sep 2026 19:48:31 GMT</pubDate><category>tools</category><source url="https://downdetector.co.uk/status/deepseek">Hacker News</source></item><item><title>Agentic Coding Strains CI: Scaling Test Impact Analysis at Anthropic</title><link>https://claude.com/blog/agentic-coding-is-straining-ci-heres-how-we-scaled-test-impact-analysis-at-anthropic</link><guid isPermaLink="false">86abcc6f17fc520e81400d2a923ff7fbed92713023da912470e66ad10a0f8871</guid><description></description><pubDate>Mon, 14 Sep 2026 19:45:27 GMT</pubDate><category>agents</category><source url="https://claude.com/blog/agentic-coding-is-straining-ci-heres-how-we-scaled-test-impact-analysis-at-anthropic">Hacker News</source></item><item><title>DeepSeek Is Down</title><link>https://status.deepseek.com</link><guid isPermaLink="false">96ebfff458b6b94e1884931738e0ebf122db46ea844ec1fa1e62964ddc64dff0</guid><description></description><pubDate>Mon, 14 Sep 2026 19:20:11 GMT</pubDate><category>models</category><source url="https://status.deepseek.com">Hacker News</source></item><item><title>Radio – a chatroom for multiplayer agent work</title><link>https://radio.plasma.ai</link><guid isPermaLink="false">b458441ad47ccfd9236c7112010016193a5cf25da10e1b0a4eb7223d68d6b13a</guid><description></description><pubDate>Mon, 14 Sep 2026 19:09:29 GMT</pubDate><category>agents</category><source url="https://radio.plasma.ai">Hacker News</source></item><item><title>A functional taxonomy for LLM inference in agentic</title><link>https://jeffauriemma.leaflet.pub/3mv6jnffo6k24</link><guid isPermaLink="false">ee311dfcc302e2b083cc1b7893ac60879f747a00b1774e7d1cafd5335dd6cc26</guid><description></description><pubDate>Mon, 14 Sep 2026 18:59:21 GMT</pubDate><category>local-llm</category><source url="https://jeffauriemma.leaflet.pub/3mv6jnffo6k24">Hacker News</source></item><item><title>Anthropic Disrupts Bioweapons Efforts, Russian Hacking, Chinese Claude Misuse</title><link>https://www.usnews.com/news/world/articles/2026-09-10/anthropic-disrupts-russian-chinese-ai-campaigns-targeting-its-claude-models</link><guid isPermaLink="false">5d7b9db2865674a00851ebe63fedba11a9023bc11b0c29d68cfec7953ac24ca2</guid><description></description><pubDate>Mon, 14 Sep 2026 18:56:25 GMT</pubDate><category>models</category><source url="https://www.usnews.com/news/world/articles/2026-09-10/anthropic-disrupts-russian-chinese-ai-campaigns-targeting-its-claude-models">Hacker News</source></item><item><title>OpenAI Buys Startup Developing Smartphone Camera</title><link>https://www.wsj.com/tech/openai-buys-startup-developing-smartphone-camera-63590370</link><guid isPermaLink="false">6df0c7bd2f8c78a525677816a44559144be48122c5cf29bfa506cca4de08bc08</guid><description></description><pubDate>Mon, 14 Sep 2026 18:34:27 GMT</pubDate><category>industry</category><source url="https://www.wsj.com/tech/openai-buys-startup-developing-smartphone-camera-63590370">Hacker News</source></item><item><title>v0.4.1</title><link>https://github.com/ggml-org/llama.cpp/releases/tag/v0.4.1</link><guid isPermaLink="false">842d340f487b0523b08d90cc49e55b58b00b83662144bb5a86c66e3e6723a658</guid><description>Overview llama.cpp 0.4.1 adds Maple 20B-A1B, Tencent Hy 4, and Spark2.5 support, improves JSON schema handling, chat parsing, logging, and server child-process management, and updates ggml to v0.24.0. API changes Changed llamasamplerchainn() to return int32t instead of int (#28631). Added serversubproc and waiter for router child-process monitoring in server</description><pubDate>Mon, 14 Sep 2026 18:27:29 GMT</pubDate><category>local-llm</category><source url="https://github.com/ggml-org/llama.cpp/releases/tag/v0.4.1">GitHub · ggml-org/llama.cpp</source></item><item><title>Google tests paying publishers for using its content in AI Mode and Gemini</title><link>https://searchengineland.com/google-tests-paying-publishers-for-using-its-content-in-ai-mode-ai-overviews-and-gemini-488382</link><guid isPermaLink="false">02c740ed65dfc602378bdbc17fd88e31eb6d240dfda54607b7ae4855c8cb618a</guid><description></description><pubDate>Mon, 14 Sep 2026 18:06:47 GMT</pubDate><category>models</category><source url="https://searchengineland.com/google-tests-paying-publishers-for-using-its-content-in-ai-mode-ai-overviews-and-gemini-488382">Hacker News</source></item><item><title>Show HN: Taskuary – turn your inbox into tasks for Claude Code, Codex, or Gemini</title><link>https://github.com/ldbumble/taskuary</link><guid isPermaLink="false">aaac4c73f6063b6f74e95841baed0e047e0994dec2ee7dd8c69169846fc84112</guid><description></description><pubDate>Mon, 14 Sep 2026 17:55:23 GMT</pubDate><category>models</category><source url="https://github.com/ldbumble/taskuary">Hacker News</source></item><item><title>Anthropic targets financial advisers with new Claude tool</title><link>https://www.reuters.com/business/anthropic-targets-financial-advisers-with-new-claude-tool-2026-09-14</link><guid isPermaLink="false">f62a11e18bd08cb3decb637a8bd8c58a49e67b03b2752f90df704648ce5efe36</guid><description></description><pubDate>Mon, 14 Sep 2026 17:52:02 GMT</pubDate><category>tools</category><source url="https://www.reuters.com/business/anthropic-targets-financial-advisers-with-new-claude-tool-2026-09-14">Hacker News</source></item><item><title>Show HN: CostClaw, a local audit of your Claude Code logs</title><link>https://www.costclaw.io</link><guid isPermaLink="false">cec4429b1320ecb82146a91d8689e149b78e2cbcddb9284227587d562e89849f</guid><description></description><pubDate>Mon, 14 Sep 2026 17:44:14 GMT</pubDate><category>models</category><source url="https://www.costclaw.io">Hacker News</source></item><item><title>A Year of Hacking with LLMs</title><link>https://sites.google.com/site/zhiniangpeng/blogs/Hacking-with-LLMs-Eng</link><guid isPermaLink="false">adea1e0cdaff8c2746d61f1edd1b333eb43b9924e2ca5140392de5c3178a4187</guid><description></description><pubDate>Mon, 14 Sep 2026 17:22:39 GMT</pubDate><category>industry</category><source url="https://sites.google.com/site/zhiniangpeng/blogs/Hacking-with-LLMs-Eng">Hacker News</source></item><item><title>Can AI agents conduct open-ended AI research?</title><link>https://arxiv.org/abs/2607.27191</link><guid isPermaLink="false">1cf9928a3d86ee7f65735dfdca01788d2e6e375fe11368b1e390e7d27af25aae</guid><description></description><pubDate>Mon, 14 Sep 2026 17:10:50 GMT</pubDate><category>agents</category><source url="https://arxiv.org/abs/2607.27191">Hacker News</source></item><item><title>AI agents blew the whistle on their cheating colleagues</title><link>https://www.technologyreview.com/2026/09/14/1144037/ai-agents-blew-whistle-o-cheating-colleagues</link><guid isPermaLink="false">876402fef721cf37c77fe99418c9f717f4f5a7b813b1cb5645648bb02339ba87</guid><description></description><pubDate>Mon, 14 Sep 2026 17:01:39 GMT</pubDate><category>agents</category><source url="https://www.technologyreview.com/2026/09/14/1144037/ai-agents-blew-whistle-o-cheating-colleagues">Hacker News</source></item><item><title>Show HN: MCP Harbor – An MCP Registry</title><link>https://ai.mcpharbor.dev</link><guid isPermaLink="false">d0fdfef5044913ad6a38e1315d6610751d9b440468cffa7431a28bc6f0c3d6fc</guid><description></description><pubDate>Mon, 14 Sep 2026 16:59:00 GMT</pubDate><category>agents</category><source url="https://ai.mcpharbor.dev">Hacker News</source></item><item><title>Limen: One-human-many-agents harness built from files, Git and one CLI</title><link>https://github.com/overment/limen</link><guid isPermaLink="false">eed3d986e5f1e8fd6e384f2d93c9642f316a78807680601cb995925fa6960643</guid><description></description><pubDate>Mon, 14 Sep 2026 16:54:46 GMT</pubDate><category>agents</category><source url="https://github.com/overment/limen">Hacker News</source></item><item><title>A1ex: A simple LLM coding agent in Lua</title><link>https://github.com/ziyao233/a1ex</link><guid isPermaLink="false">630adab8ceddab5cb3f2f0b2d59c5f2b8079474cad9b786dc203138f02b5e748</guid><description></description><pubDate>Mon, 14 Sep 2026 16:51:42 GMT</pubDate><category>agents</category><source url="https://github.com/ziyao233/a1ex">Hacker News</source></item><item><title>Prompt steering can reduce DeepSeek v4 hallucinations to below GPT and Claude</title><link>https://propensitylabs.substack.com/p/how-to-get-deepseek-to-hallucinate</link><guid isPermaLink="false">ba20ac6f0fb68eb4ec0e4cdb9d768b9c35d8c6cfb6fa7254418e876be3886e38</guid><description></description><pubDate>Mon, 14 Sep 2026 16:51:24 GMT</pubDate><category>models</category><source url="https://propensitylabs.substack.com/p/how-to-get-deepseek-to-hallucinate">Hacker News</source></item><item><title>Accelerating Dropless MoE Training in JAX with NVIDIA Transformer Engine</title><link>https://developer.nvidia.com/blog/accelerating-dropless-moe-training-in-jax-with-nvidia-transformer-engine</link><guid isPermaLink="false">75ebfc4f40c5e28ba3dfdf08c6c5bd1b6a2f89ebc64c2f334ce4631d8aad3af3</guid><description>Mixture of experts (MoE) has become one of the defining architectural trends in large-scale AI model training. DeepSeek, Qwen, and Mixtral are examples of MoE...</description><pubDate>Mon, 14 Sep 2026 16:39:15 GMT</pubDate><category>hardware</category><source url="https://developer.nvidia.com/blog/accelerating-dropless-moe-training-in-jax-with-nvidia-transformer-engine">NVIDIA</source></item><item><title>Show HN: Cognition-Claude-proxy – Use Devin model catalog with Claude Code</title><link>https://github.com/coderexpert123/cognition-claude-proxy</link><guid isPermaLink="false">8a7733327eb5f66fb1e257c5f048c98eeef94038d57d993e3b6500fa0e83af48</guid><description></description><pubDate>Mon, 14 Sep 2026 16:33:33 GMT</pubDate><category>models</category><source url="https://github.com/coderexpert123/cognition-claude-proxy">Hacker News</source></item><item><title>Show HN: Rebuno - An open-source runtime for production agents</title><link>https://github.com/rebuno/rebuno</link><guid isPermaLink="false">9058d24670a028347a4ab399a87c4f6f102cbe2ccb2fee8608d70e4931ef6144</guid><description></description><pubDate>Mon, 14 Sep 2026 16:29:39 GMT</pubDate><category>agents</category><source url="https://github.com/rebuno/rebuno">Hacker News</source></item><item><title>When LLM judges agree, should we believe them?</title><link>https://www.amazon.science/blog/when-llm-judges-agree-should-we-believe-them</link><guid isPermaLink="false">5728c436a7598b96a221958e9aa58fbfcffa603952368f6713c396d941f2a1cc</guid><description></description><pubDate>Mon, 14 Sep 2026 16:29:30 GMT</pubDate><category>industry</category><source url="https://www.amazon.science/blog/when-llm-judges-agree-should-we-believe-them">Hacker News</source></item><item><title>I stress-tested Meta Muse until its agent control plane started timing out</title><link>https://blog.cygankiewicz.com/en/meta-muse-black-box-testing</link><guid isPermaLink="false">b4dc4435b90415a7f713f6911f9aa2d2a82f865c7658851a01d674fcb4b2c4a3</guid><description></description><pubDate>Mon, 14 Sep 2026 16:26:07 GMT</pubDate><category>agents</category><source url="https://blog.cygankiewicz.com/en/meta-muse-black-box-testing">Hacker News</source></item><item><title>Hacking AI customer service agents</title><link>https://www.intigriti.com/researchers/blog/hacking-tools/hacking-ai-customer-service-agents</link><guid isPermaLink="false">47f7e8228e0b9c7f9965f475e8cd05e5cfb743f23560d3d912333cb734fc94b6</guid><description></description><pubDate>Mon, 14 Sep 2026 16:26:03 GMT</pubDate><category>agents</category><source url="https://www.intigriti.com/researchers/blog/hacking-tools/hacking-ai-customer-service-agents">Hacker News</source></item><item><title>Running a small LLM in reMarkable 2</title><link>https://github.com/ykumards/unremarkable</link><guid isPermaLink="false">bb8b7cf5d9c517eeb5f8d2e1fedc7a1a153e093b1c5008515078ea0524afe6fd</guid><description></description><pubDate>Mon, 14 Sep 2026 16:20:11 GMT</pubDate><category>industry</category><source url="https://github.com/ykumards/unremarkable">Hacker News</source></item><item><title>Claude is a Contrarian</title><link>https://medium.com/@rdsubhas/claude-is-a-contrarian-dbce4de5cada</link><guid isPermaLink="false">c5797042bd82aba9395ab2516c58b1e26c8ecb25b6cdb67ed6cff08216b4d29c</guid><description></description><pubDate>Mon, 14 Sep 2026 16:14:24 GMT</pubDate><category>models</category><source url="https://medium.com/@rdsubhas/claude-is-a-contrarian-dbce4de5cada">Hacker News</source></item><item><title>Show HN: AgentDrive – persistent, versioned file storage for AI agents</title><link>https://tokencanopy.com/products/agentdrive</link><guid isPermaLink="false">59cb396054a6c678c8587bae7d6be56f7ffc365a5efb444072f9d40f06ae1127</guid><description></description><pubDate>Mon, 14 Sep 2026 16:09:02 GMT</pubDate><category>agents</category><source url="https://tokencanopy.com/products/agentdrive">Hacker News</source></item><item><title>Show HN: Local catalog of 3k agent skills with a static risk scan</title><link>https://github.com/xm1k3/ai-community-skills</link><guid isPermaLink="false">2f2010caf23c3045ae02022dc84bc1d890745b125fa0b1b0462b2a785b232b0d</guid><description></description><pubDate>Mon, 14 Sep 2026 15:49:57 GMT</pubDate><category>agents</category><source url="https://github.com/xm1k3/ai-community-skills">Hacker News</source></item><item><title>ProGantt: Gantt charts your AI agent can read and write via MCP</title><link>https://progantt.com</link><guid isPermaLink="false">31d7fab91d03d0433f272b380b339a8bd79f72529b79f42da80bf15f110a4ed6</guid><description></description><pubDate>Mon, 14 Sep 2026 15:48:32 GMT</pubDate><category>agents</category><source url="https://progantt.com">Hacker News</source></item><item><title>A "slowdown" might improve OpenAI and Anthropic products</title><link>https://killerstorm.github.io/ai/2026/09/14/slowbetter.html</link><guid isPermaLink="false">fbb2648375de5955c9d9a3bdf03039bc2349e025e820ba9d5b68ee8e3a0f48c5</guid><description></description><pubDate>Mon, 14 Sep 2026 15:34:55 GMT</pubDate><category>industry</category><source url="https://killerstorm.github.io/ai/2026/09/14/slowbetter.html">Hacker News</source></item><item><title>Show HN: Botbin – A zero-login HTML pastebin for AI agents</title><link>http://botbin.io</link><guid isPermaLink="false">a3e1ddcbcbef8669be9317b5069c980b4c5611b6ae386e82911d2a5dcdfc5a77</guid><description></description><pubDate>Mon, 14 Sep 2026 15:30:44 GMT</pubDate><category>agents</category><source url="http://botbin.io">Hacker News</source></item><item><title>Qwen Audio Agent – full duplex via ACP</title><link>https://qwenaudio.github.io/qwen-audio-agent</link><guid isPermaLink="false">d2c2629c200bd750b1f3082e82d6af669aa813096fbbc44f04a647b63a456953</guid><description></description><pubDate>Mon, 14 Sep 2026 15:28:23 GMT</pubDate><category>agents</category><source url="https://qwenaudio.github.io/qwen-audio-agent">Hacker News</source></item><item><title>b10964</title><link>https://github.com/ggml-org/llama.cpp/releases/tag/b10964</link><guid isPermaLink="false">db6c34d3204d219eb5fa14ef0ac9b62f8e7a7b26c302f080626f8c7318dfc24c</guid><description>llama.cpp : bump version to 0.4.1 (#28900) Website: link Attestations: link macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Linux: Ubuntu x64 (CPU) Ubuntu arm64 (CPU) Ubuntu s390x (CPU) Ubuntu x64 (Vulkan) Ubuntu arm64 (Vulkan) Ubuntu x64 (ROCm 10.0) Ubuntu x64 (OpenVINO) Ubuntu</description><pubDate>Mon, 14 Sep 2026 15:28:07 GMT</pubDate><category>local-llm</category><source url="https://github.com/ggml-org/llama.cpp/releases/tag/b10964">GitHub · ggml-org/llama.cpp</source></item><item><title>The Memory That AI Agents Love</title><link>https://medium.com/@majid.fekri/the-memory-that-ai-agents-love-b9ec0a708e7c</link><guid isPermaLink="false">517c32a9a897dd48a5b94185d2f238579f02ec31812332692d1e4a9ec1249116</guid><description></description><pubDate>Mon, 14 Sep 2026 15:25:01 GMT</pubDate><category>agents</category><source url="https://medium.com/@majid.fekri/the-memory-that-ai-agents-love-b9ec0a708e7c">Hacker News</source></item><item><title>Transitions.dev: UI transitions for AI agents</title><link>https://transitions.dev</link><guid isPermaLink="false">352476c117cdd6f24c22ae8e438f6e7762a20b5e6b51e428d894ccece29deeb9</guid><description></description><pubDate>Mon, 14 Sep 2026 15:18:58 GMT</pubDate><category>agents</category><source url="https://transitions.dev">Hacker News</source></item><item><title>Andon Labs Puts AI Agents in Charge of Real Businesses</title><link>https://spectrum.ieee.org/andon-labs-agentic-ai-businesses</link><guid isPermaLink="false">116cc28ea6686e17acae31912106c44182377de69f163b4e599c4c0c8349667c</guid><description></description><pubDate>Mon, 14 Sep 2026 15:08:02 GMT</pubDate><category>agents</category><source url="https://spectrum.ieee.org/andon-labs-agentic-ai-businesses">Hacker News</source></item><item><title>Show HN: Skillzero – save tokens by omitting skills from agent context</title><link>https://github.com/kurtextrem/skillzero</link><guid isPermaLink="false">d52b089c4bfa6264df6b3ed9838045a4780b238ac1e7a7348db79ca4e3c05583</guid><description></description><pubDate>Mon, 14 Sep 2026 15:06:04 GMT</pubDate><category>agents</category><source url="https://github.com/kurtextrem/skillzero">Hacker News</source></item><item><title>Has anyone made a legacy codebase more legible to AI coding agents?</title><link>https://news.ycombinator.com/item?id=49698073</link><guid isPermaLink="false">f4168ac93d24ff26dc220645bdc4302f4ddbe9f686c4d17b86e9c3524bce7901</guid><description></description><pubDate>Mon, 14 Sep 2026 15:00:32 GMT</pubDate><category>agents</category><source url="https://news.ycombinator.com/item?id=49698073">Hacker News</source></item><item><title>Your AI agent has a cache invalidation problem</title><link>https://www.drjoshcsimmons.com/writing/your-ai-agent-has-a-cache-invalidation-problem</link><guid isPermaLink="false">17967f939f3d2bd4df46330de5b87b5e6313d68eed22f8a32fbbf28b8a958b96</guid><description></description><pubDate>Mon, 14 Sep 2026 14:51:14 GMT</pubDate><category>agents</category><source url="https://www.drjoshcsimmons.com/writing/your-ai-agent-has-a-cache-invalidation-problem">Hacker News</source></item><item><title>Claude Code Over 6k issues labeled with has repro have been auto-closed</title><link>https://github.com/anthropics/claude-code/issues/87647</link><guid isPermaLink="false">5eef42654ca4cb8f6781f7103578611a18db401e509e793b31a8679fe550ebbe</guid><description></description><pubDate>Mon, 14 Sep 2026 14:43:19 GMT</pubDate><category>models</category><source url="https://github.com/anthropics/claude-code/issues/87647">Hacker News</source></item><item><title>RubyGems Open Source Supply Chain Security and OpenAI</title><link>https://rietta.com/blog/rubygems-supply-chain-openai</link><guid isPermaLink="false">88745fd4339ee51dddc63f7d0b0bbb93c49523315dd03a43fd218ed0919238c5</guid><description></description><pubDate>Mon, 14 Sep 2026 14:39:01 GMT</pubDate><category>security</category><source url="https://rietta.com/blog/rubygems-supply-chain-openai">Hacker News</source></item><item><title>Qwen 3.8 Flash via DwarfStar</title><link>https://twitter.com/antirez/status/2099487927793299936</link><guid isPermaLink="false">ce3af1cf96fd15bf1299c5d1ce2d29f1c2fe93802e385067266452a53332aa95</guid><description></description><pubDate>Mon, 14 Sep 2026 14:38:09 GMT</pubDate><category>models</category><source url="https://twitter.com/antirez/status/2099487927793299936">Hacker News</source></item><item><title>Quoting Laurie Voss</title><link>https://simonwillison.net/2026/Sep/14/laurie-voss</link><guid isPermaLink="false">340906303398794d2afc6905a152ee605e077da206ab2ea5f0d0c1e9ed6375d8</guid><description>The cost of writing code collapsed, and the cost of reviewing, fixing and operating it is following, and I'm assuming it gets there. What's left of making software is finding out what people actually want, defining it precisely, and making it pleasant to use. That cost is per piece of software and doesn't transfer, so as the amount of software goes to infini</description><pubDate>Mon, 14 Sep 2026 14:34:29 GMT</pubDate><category>agents</category><source url="https://simonwillison.net/2026/Sep/14/laurie-voss">Simon Willison</source></item>
  </channel>
</rss>
