Skip to content
AI NEWS.ccAI正在发生 未来正在更新
Bookmarks
THE JOURNAL
AI NEWS JOURNAL
AI, IN PERSPECTIVE

Search AI news

Search headlines, summaries and articles to connect the dots.

Newest first

llama.cpp releases

b12202

The llama.cpp v1.220.2 release introduces enhanced BF16 precision support across multiple components, aligning with reference implementations by moving specific operations like GLM4 family projections, MiMoVL vision matmuls, and Qwen3-TTS speaker encoder conv1d to BF16 where appropriate. Backends such as CUDA, Hexagon, OpenCL, WebGPU, and Vulkan now uniformly handle GGML_PREC_BF16 similarly to F32, while CPU, SpaceMIT, and ET backends extend flash attention support to BF16 with F32 accumulation. Flash attention tests now include BF16 cases, and the KQ multiplication in Llama is updated to request BF16 precision. A TODO remains in clip.cpp for future work.

llama.cpp releases

b12207

A new fast path for d_state=64 has been added to the SSM_SCAN kernel in ggml-cuda, improving generation throughput for Zamba2-7B-Instruct from 11 to 15.1 tok/s—approximately a 35% increase—by eliminating CPU fallbacks for this state size. The change was verified with real-world Zamba2 hyperparameters and maintains correctness, though a separate 28% of scan calls still fall back to CPU due to an undiagnosed issue.

llama.cpp releases

b12208

The ggml library has been updated to make tied selections in the radix Top-K algorithm deterministic on CUDA and Metal backends by replacing atomic compaction loops with a deterministic scheme using simdgroup scans and coalesced memory accesses. This change ensures byte-identical outputs across runs, improves performance by 20-45% on M5 Ultra hardware, and enables deterministic testing for Metal and ROCm. Redundant CUDA kernel resets were removed, and Vulkan/Metal changes were reduced or skipped.

Hacker News Show HN

Show HN: An open-source IDE for software architecture

Now, with this open-source tool everyone can learn,build,share,and validate the system design.This is next step, drawing design on blank page to making it a workflow .Do explore and let me know how it is... Comments URL: https://news.ycombinator.com/item?id=50412135 Points: 1 # Comments: 0

Hacker News Show HN

Show HN: Coldframe, a honeypot that poses as a supermarket fridge controller

I built Coldframe, a honeypot that pretends to be a supermarket refrigeration controller. Inspired by the events on industrial fridge installations in the past weeks, I wanted to see what people would try against a convincing fake.To learn more about this I built a honeypot that resembled perfectly the Danfoss AK-SM 800A and Dixell XWEB interfaces. Both of them are recognized on Shodan as real ones (an not as honeypots). It’s one Python file using only the standard library, and it builds on eppser/danfoss-ak-sm-decoy and public research by Synacktiv, Claroty and Ironhackers.Things that were fun to get right:- Matching real units. I used Shodan records of real devices to match header order, page titles, certificate fields and TLS versions. The Danfoss persona sends nginx headers in the same order as a real AK-SM 800A. The Dixell persona sends no Server header and pads its page to exactly 8,202 bytes.- Recognising exploit attempts. It recognises CVE-2025-41450, the Danfoss “code of the day” login bypass, and records whether the submitted code is correct for today. It also tags attempts targeting three other Danfoss CVEs and the XWEB vulnerabilities.- Letting them in. Default logins open a fake store dashboard with defrost buttons, so you see the second move, not just the first. Nothing behind it is real.Two instances have been live for a day. So far they’ve only seen generic scanners and Shodan’s crawler, which is called refrigerator.census.shodan.io.Feedback welcome, especially from anyone who has seen these devices in the field! Comments URL: https://news.ycombinator.com/item?id=50406292 Points: 15 # Comments: 1

Hacker News Show HN

Show HN: Digdeep.fyi – A DNS Debug Tool for DNSSEC, Mail, HTTP

A fully client side, fast DNS debug tool. Queries DNS via DoH resolvers and analyzes DNSSEC configs (inspired by DNSViz), mail records (MX, SRF, DKIM, DMARC, etc), SRC, HTTPS records and runs multiple compliance checks. Comments URL: https://news.ycombinator.com/item?id=50406439 Points: 1 # Comments: 0

Hacker News Show HN

Show HN: Neural Network Formal Verification

Article URL: https://github.com/giovanni-norbedo/awesome-nn-formal-verification Comments URL: https://news.ycombinator.com/item?id=50412710 Points: 1 # Comments: 0

Hacker News Show HN

Show HN: Shutdown Watch – App shutdowns, export deadlines, and alternatives

My recipe app picked up users when a similar app shut down, and I kept noticing people scrambling to export their data, so I created this project to scratch my own itch.The 3 main purposes this serves are:1. Give the public a way to find how to export their data or find a replacement if one of the apps they use shuts down.2. Give app creators a way to get their apps in front of people who are actively looking for a replacement for a similar app. Sponsored spots are labeled and never change the dates or export steps.3. Provide inspiration for app creators looking for ideas for what to build next. Just because apps shut down doesn't mean they were not good ideas; many times they shut down for other reasons, like poor execution, no marketing, over-spending, etc.Every date and export step links to the company's own notice.I'm always open to ideas to make it more useful. I read all the comments and constructive criticism and input are always welcomed. Comments URL: https://news.ycombinator.com/item?id=50412639 Points: 1 # Comments: 0

Hacker News Show HN

Show HN: Throng, a project first, terminal 2nd IDE for command line people

I've been beavering away at this since July, and its finally getting to a point where its stabilising.It started Windows only, but I recently converted over to Linux, so I've been ironing out the kinks in the last week. You can install it via apt or rpm, or .exe or .msi on Windows.Many more features to come, including native agent support and more preview features (images, PDF, document formats). Claude works pretty well in it, but I've not tried many other agents yet.I rarely leave throng now - 99% of my work happens inside it. I stopped using Visual Studio around 10 months ago (in favour of VSCode), and now I don't even use VSCode. Comments URL: https://news.ycombinator.com/item?id=50412226 Points: 1 # Comments: 0

Hacker News Show HN

Show HN: UK: Skip Royal Mail QR Code Queues with the Prince of Post

This scratches an itch I experience most months. When you get a "super handy QR Code" from a service, like Asos, Vinted and others that means you have to physically go to a post office to post your parcel. About ten years ago this was revolutionary convenience, but now it's a frustration.If you're like me you use Royal Mail's collection service to ship most of your parcels/returns from home, and post offices are open for ridiculously narrow times of day these days. Royal Mail don't provide an online facility to create a label from these in-person QR codes, so I built one instead.Behold, Citizen - Prince of Post! https://princeofpost.comI had some fun designing this, and it's something that I'll genuinely use all the time, so I thought I'd share it for others with this problem too. I couldn't find any prior art, but if you want to ruin my day please do tell me about anywhere that does this already :-) Comments URL: https://news.ycombinator.com/item?id=50406466 Points: 1 # Comments: 0

llama.cpp releases

b12175

A fix was implemented to override the hardware_concurrency value in CI tests, addressing an issue where test-llama-context and test-backend-ops hung due to incorrect thread count estimation on 8-CPU runners. The solution sets default values of 8 for TEST_ARG_HARDWARE_CONCURRENCY and 4 for TEST_ARG_PARALLEL_WORKERS specifically for the cpu-x64-high-perf configuration.

llama.cpp releases

b12176

The update qualifies calls to math functions like isnan and signbit with std:: to ensure compatibility with C++ standard library overloads, addressing issues on NetBSD 11.99.9/x86_64 with gcc 14.4.0. Additional C math function calls are standardized using std:: prefixes (e.g., std::fabs, std::copysign) instead of non-existent std::fabsf variants, improving accuracy by keeping computations in double precision. The change results in a minor behavioral difference with a relative error of at most 1.4e-7, well below test tolerances.

llama.cpp releases

b12177

A fix has been implemented to prevent crashes in the command-line interface when encountering invalid UTF-8 in prompts or files, with assistance from Claude. This update is part of the b12177 release of llama.cpp, which includes various binary distributions for macOS, Linux, Windows, Android, and other platforms.