Hacker News (Front Page) (RSS/Atom feed)

Document poisoning in RAG systems: How attackers corrupt AI's sources I'm the author. Repo is here: [https://github.com/aminrj-labs/mcp-attack-labs/tree/main/…][1] The lab runs entirely on LM Studio + Qwen2.5-7B-Instruct (Q4_K_M) + ChromaDB — no cloud APIs, no GPU required, no API keys. From zero to seeing the poisoning succeed: git clone, make setup, make attack1. About 10 minutes. Two things worth flagging upfront: - The 95% success rate is against a 5-document corpus (best case for the attacker). In a mature collection you need proportionally more poisoned docs to dominate retrieval — but the mechanism is the same. - Embedding anomaly detection at ingestion was the biggest surprise: 95% → 20% as a standalone control, outperforming all three generation-phase defenses combined. It runs on embeddings your pipeline already produces — no additional model. All five layers combined: 10% residual. Happy to discuss methodology, the PoisonedRAG comparison, or anything that looks off. Comments URL: [https://news.ycombinator.com/item?id=47350407][2] Points: 13 # Comments: 1 [1]: https://github.com/aminrj-labs/mcp-attack-labs/tree/main/… [2]: https://news.ycombinator.com/item?id=47350407 https://aminrj.com/posts/rag-document-poisoning/

Shall I implement it? No Article URL: [https://gist.github.com/bretonium/291f4388e2de89a43b25c13…][1] Comments URL: [https://news.ycombinator.com/item?id=47357042][2] Points: 23 # Comments: 1 [1]: https://gist.github.com/bretonium/291f4388e2de89a43b25c13… [2]: https://news.ycombinator.com/item?id=47357042 https://gist.github.com/bretonium/291f4388e2de89a43b25c13…

Illinois introduces OS-level age verification law Article URL: [https://legiscan.com/IL/bill/SB3977/2025][1] Comments URL: [https://news.ycombinator.com/item?id=47357294][2] Points: 13 # Comments: 0 [1]: https://legiscan.com/IL/bill/SB3977/2025 [2]: https://news.ycombinator.com/item?id=47357294 https://legiscan.com/IL/bill/SB3977/2025

AI error jails innocent grandmother for months in North Dakota fraud case Article URL: [https://www.grandforksherald.com/news/north-dakota/ai-err…][1] Comments URL: [https://news.ycombinator.com/item?id=47356968][2] Points: 27 # Comments: 10 [1]: https://www.grandforksherald.com/news/north-dakota/ai-err… [2]: https://news.ycombinator.com/item?id=47356968 https://www.grandforksherald.com/news/north-dakota/ai-err…

Bringing Chrome to ARM64 Linux Devices Article URL: [https://blog.chromium.org/2026/03/bringing-chrome-to-arm6…][1] Comments URL: [https://news.ycombinator.com/item?id=47356392][2] Points: 6 # Comments: 1 [1]: https://blog.chromium.org/2026/03/bringing-chrome-to-arm6… [2]: https://news.ycombinator.com/item?id=47356392 https://blog.chromium.org/2026/03/bringing-chrome-to-arm6…

Runners Are Discovering It's Surprisingly Easy to Churn Butter on Their Runs Article URL: [https://www.runnersworld.com/news/a70683169/how-to-make-b…][1] Comments URL: [https://news.ycombinator.com/item?id=47355781][2] Points: 5 # Comments: 0 [1]: https://www.runnersworld.com/news/a70683169/how-to-make-b… [2]: https://news.ycombinator.com/item?id=47355781 https://www.runnersworld.com/news/a70683169/how-to-make-b…

An old photo of a large BBS Article URL: [https://rachelbythebay.com/w/2022/01/26/swcbbs/][1] Comments URL: [https://news.ycombinator.com/item?id=47356017][2] Points: 11 # Comments: 2 [1]: https://rachelbythebay.com/w/2022/01/26/swcbbs/ [2]: https://news.ycombinator.com/item?id=47356017 https://rachelbythebay.com/w/2022/01/26/swcbbs/

The Bitter Lesson Has No Utility Function Article URL: [https://gfrm.in/posts/bitter-lesson-missing-half/index.html][1] Comments URL: [https://news.ycombinator.com/item?id=47355077][2] Points: 6 # Comments: 1 [1]: https://gfrm.in/posts/bitter-lesson-missing-half/index.html [2]: https://news.ycombinator.com/item?id=47355077 https://gfrm.in/posts/bitter-lesson-missing-half/index.html

White House plan to break up iconic U.S. climate lab moves forward Article URL: [https://www.science.org/content/article/white-house-plan-…][1] Comments URL: [https://news.ycombinator.com/item?id=47355702][2] Points: 11 # Comments: 0 [1]: https://www.science.org/content/article/white-house-plan-… [2]: https://news.ycombinator.com/item?id=47355702 https://www.science.org/content/article/white-house-plan-…

'AI is African intelligence': The workers who train AI are fighting back [https://archive.ph/yS8xb][1] Comments URL: [https://news.ycombinator.com/item?id=47353019][2] Points: 5 # Comments: 1 [1]: https://archive.ph/yS8xb [2]: https://news.ycombinator.com/item?id=47353019 https://www.404media.co/ai-is-african-intelligence-the-wo…

Launch HN: IonRouter (YC W26) – High-throughput, low-cost inference Hey HN — I’m Veer and my cofounder is Suryaa. We're building Cumulus Labs (YC W26), and we're releasing our latest product IonRouter ([https://ionrouter.io/][1]), an inference API for open-source and fine tuned models. You swap in our base URL, keep your existing OpenAI client code, and get access to any model (open source or finetuned to you) running on our own inference engine. The problem we kept running into: every inference provider is either fast-but-expensive (Together, Fireworks — you pay for always-on GPUs) or cheap-but-DIY (Modal, RunPod — you configure vLLM yourself and deal with slow cold starts). Neither felt right for teams that just want to ship. Suryaa spent years building GPU orchestration infrastructure at TensorDock and production systems at Palantir. I led ML infrastructure and Linux kernel development for Space Force and NASA contracts where the stack had to actually work under pressure. When we started building AI products ourselves, we kept hitting the same wall: GPU infrastructure was either too expensive or too much work. So we built IonAttention — a C++ inference runtime designed specifically around the GH200's memory architecture. Most inference stacks treat GH200 as a compatibility target (make sure vLLM runs, use CPU memory as overflow). We took a different approach and built around what makes the hardware actually interesting: a 900 GB/s coherent CPU-GPU link, 452GB of LPDDR5X sitting right next to the accelerator, and 72 ARM cores you can actually use. Three things came out of that that we think are novel: (1) using hardware cache coherence to make CUDA graphs behave as if they have dynamic parameters at zero per-step cost — something that only works on GH200-class hardware; (2) eager KV block writeback driven by immutability rather than memory pressure, which drops eviction stalls from 10ms+ to under 0.25ms; (3) phantom-tile attention scheduling at small batch sizes that cuts attention time by over 60% in the worst-affected regimes. We wrote up the details at cumulus.blog/ionattention. On multimodal pipelines we get better performance than big players (588 tok/s vs. Together AI's 298 on the same VLM workload). We're honest that p50 latency is currently worse (~1.46s vs. 0.74s) — that's the tradeoff we're actively working on. Pricing is per token, no idle costs: GPT-OSS-120B is $0.02 in / $0.095 out, Qwen3.5-122B is $0.20 in / $1.60 out. Full model list and pricing at [https://ionrouter.io][2]. You can try the playground at [https://ionrouter.io/playground][3] right now, no signup required, or drop your API key in and swap the base URL — it's one line. We built this so teams can see the power of our engine and eventually come to us for their finetuned model needs using the same solution. We're curious what you think, especially if you're running finetuned or custom models — that's the use case we've invested the most in. What's broken, what would make this actually useful for you? Comments URL: [https://news.ycombinator.com/item?id=47355410][4] Points: 6 # Comments: 0 [1]: https://ionrouter.io/ [2]: https://ionrouter.io [3]: https://ionrouter.io/playground [4]: https://news.ycombinator.com/item?id=47355410 https://ionrouter.io

Show HN: Aurion OS – A 32-bit GUI operating system written from scratch in C Hi HN! I'm 13 and I built Aurion OS as a solo learning project over 14 days (~12 hours/day). It's a 32-bit x86 operating system written entirely in C and x86 Assembly with no external libraries. What it has: Custom bootloader and kernel VESA framebuffer graphics (1920x1080, double-buffered) Window manager with draggable, overlapping windows macOS-inspired dock with transparency PS/2 keyboard and mouse drivers ATA hard drive driver with filesystem PCI bus enumeration RTL8139 network driver (WIP) Real-time clock Runs on just 16MB RAM (up to 10 windows simultaneously) Built-in apps: Terminal (with DOS mode), Notepad (save/load), Calculator, Paint (multiple colors and brush sizes), Snake game, Settings (theme switching), and System Info. Currently works best on QEMU, VirtualBox, and VMware. Real hardware support is still a work in progress. Next goal: TCP/IP networking stack. I'd love any feedback, suggestions, or criticism. This is my first OS project and I learned mass amounts while building it. Happy to answer any technical questions! Comments URL: [https://news.ycombinator.com/item?id=47355213][1] Points: 5 # Comments: 0 [1]: https://news.ycombinator.com/item?id=47355213 https://github.com/Luka12-dev/AurionOS