Bend 2 and the Vibe-Coding Trap

The Bend language is being pitched as a solution for AI coding, but its creator missed the current standard in formal verification by not researching existing solutions. Vibe-coding led to unnecessary complexity and verbosity in Bend's design.

OpenJev

A local model can read probabilities or write distribution tokens, with MiniCPM5 2B being the default desktop size. The model's weights are quantized and inputs remain in the browser cache.

ZCode, the GLM coding agent, silently uploads your Git history

https://tokenstead.ai/og/guides/zcode-silent-git-history-upload.png
A developer found that ZCode, an AI coding desktop app from Z.ai, silently uploads a user's entire workspace to Aliyun OSS when logged in. The uploaded data includes .git history and other sensitive information without the user's consent.

Jemalloc 5.4.0

https://opengraph.githubassets.com/757edfa22c8cec63243b6db8691aa148bbcede35b72fb0f10a038fa3aa43da14/jemalloc/jemalloc/releases/tag/5.4.0
This release includes over 160 commits focusing on technical debt cleaning, bug fixes, and performance improvements. It also introduces new features such as pinned memory tracking and OS abstraction layer.

Subnormal floating-point numbers are expensive on Intel processors

https://lemire.me/blog/wp-content/uploads/2026/09/Gemini_Generated_Image_gh1asmgh1asmgh1a-825x510.jpg
A C++ benchmark measured the performance impact of subnormal floating-point numbers on Intel processors, finding multiplications to be 45-50 times slower and divisions 18 times slower. However, AMD and ARM processors handle subnormals at full speed.

I don't like passkeys

https://hawksley.dev/blog/i-dont-like-passkeys/google-skip-password.png
The tech industry promotes passkeys as a secure login solution, but they pose risks for individuals due to permanent account lockout and device loss. Passkeys offer better security against phishing attacks, but their limitations make them unsuitable for personal use.

Microsoft exec called AI scraping 'the largest theft of labor in human history'

https://techcrunch.com/wp-content/uploads/2023/11/GettyImages-1778706504.jpg?w=1024
Microsoft and OpenAI allegedly trained AI models on copyrighted content without permission, bypassing paywalls and stripping copyright notices. The companies' actions pose an "existential threat" to publishers and journalists whose work was used to train the models.

The scourge of x86 emulation

https://fex-emu.com/images/posts/scourge/before_patch.png
The x86 Total Store Ordering (TSO) memory model poses significant challenges for emulating on ARM-based systems, with issues including alignment faults, unaligned atomic operations, and poor performance due to write-combined memory. Despite progress in addressing these problems through extensions like FEAT_LRCPC1/2/3 and hardware support for coherent cachelines, edge-cases remain unresolved, ...

Cekura (YC F24) Is Hiring

https://bookface-images.s3.us-west-2.amazonaws.com/logos/7aa5a73bb812b9f6551836dc55f9ea786a2c5179.png?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAAHH5LHZG%2F20260918%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20260918T120433Z&X-Amz-Expires=2783&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHcaCXVzLXdlc3QtMiJHMEUCIQCI4x05hiNyuY%2BK5zdgxV3I7kLMNXwxLBf3Wrh9RRk6CAIgAQp9pHAaSTlqJTSwl8B0jf1czBWsjDUSFCGxiIglSOwq5QMIQBAAGgwwMDYyMDE4MTEwNzIiDOl8vQ04mosZSMzewSrCA8qpH1JRltWQMhaaHG7W8U2a1XSXJBO8TAdITC5OXl9X2fpmDqG6J0Fk2ZW9vqSDUOvTHpxL4HoQJfLhCn2mkr901zhXeTDSuTa0dZzdd9NmVRY6MmCENE0DPdJMSF6%2BV5PdRyZ0hkYKCmu4vA5U9y%2Fh4trakXcehnkI5iYC84MwSOciGDIOXfaZ%2FqR%2B8InACFARhEqDZ2OF4YdP5FenZeZ%2Baa3%2B32uyGFhIIbT4PvP7yz2UE5AQe1OVX9ceALtvsXvfHIwpB3wvqoD6hzUMjsv36IGBjuuK4cZyJ9A2q3kyjRB%2BCEAlg7b%2BCENoJZ3xJb6Kojh2o88iBkJil11w3O6btZsc1o55XvUU2uAzaXyzuIADhcwb8RhQwtGTiDDm21r%2F0q7zJSMmkUU1t4Y5EP0d1ckx2vkpd9SsVZp%2FcCo98rtSB0s7GX7MtqvF0RVOzdKdjWlhVIYVXQ3L0SeV%2B8I%2B%2B7sA%2FUpgipAmVASUPoGCcHxXqaw37I2wbTPY6Yev%2FkDAVHRk3u2iO0mY76rYDs1kdqKCIqn68Bsd62GcvA5B1MlvX08i8V0GLwn0mSmll22i7YBFoMuQA62cK84T90iblDDQvrPVBjqlARZ8YRfwW6qcPhVVD520pCsz1aSvM%2F8dUTF7JNwoab3Rqx7U25Rx1WH0lb3VUlOMyf0Q3Chtp%2BHi%2FAjtD3uF4gwoLrsMDmgnCog5vI%2BpRPgDZdWAdhu5%2FQ7qLM5FkPi7hS2YFL5XPSOoJjsZxmH4n8BlJFJ4jLqTPfgom%2Fn%2FJZlCgce5OEfvwC7jVPWOPKAIgONXqBlcLeYZAzug2YjNuwAu9FBnRg%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=9c62b5ef1485242e9bcdb0f64cfb78c5f53d3a9b451f8a78f342a2debee1d155
Cekura is hiring a Forward Deployed Engineer to work with customers and help them get value from their self-improving conversational agents. The role involves embedding with customers, automating product insights, driving product direction, and building the FDE org.

Replacing Pull Requests with Delta

https://images.zed.dev/blog/delta-public-beta/no-prs.webp
Delta is a public beta for a multiplayer coding environment that allows collaboration without committing and pushing code. It offers continuous engineering, where the idea, implementation, review, and landing of changes can happen in the same place.

Warren Buffett Steps Down as Berkshire Chairman, Names Son to Replace Him

Please enable JS and disable any ad blocker

Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint

https://cdn.prod.website-files.com/699604cc2b9dd89bdbda0608/6aac504ec0cef3807f35bcba_2%2027B%20ID.png
The Ternary Bonsai 2 27B model is released, offering a more capable and energy-efficient version of the previous Bonsai 27B model. It retains 98.2% of aggregate benchmark performance while being nearly 9x smaller in memory footprint.

Astra for Law

https://images.ctfassets.net/kftzwdyauwt9/4PcDbhz75I70TjyGat7oHl/598dd1e865018860c3591cb95c9bc72e/paul-christiano-joins-openai-foundation-board--cover-v001.png?w=3840&q=90&fm=webp
OpenAI introduces Astra for Law, a foundation for law firms and legal technology companies to build AI products around their expertise. It combines GPT-6 with settings and tools tailored for professional legal work.

The Shadows Lurking in the Equations – Underwater Islands

https://gods.art/images/articles/slashdot_comparison.gif
FuzzyGraph visualizes equations with error levels, showing regions of high and low error. Conventional graphs only show exact equality, hiding important mathematical features.

Bend – A language that blocks AI mistakes via proof, on CPU and GPU

The Bend system uses laws and proofs to ensure AI agents implement human intents correctly, verifying code changes in real-time. With LAWS.bend, developers can declare rules that AIs must follow, preventing bugs from being shipped into production.

Qwen 3.8 Omni Flash

Qwen Studio offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.

Hister: A private search engine for the pages you visit and the files you keep

https://raw.githubusercontent.com/asciimoo/hister/master/webui/website/src/lib/assets/screenshot.png
Hister is a private search engine that indexes web pages and local files for personal use. It allows searching from various interfaces, including the browser, terminal, and AI assistants, without relying on third-party providers or cloud services.

Wax motor

https://thumb.wikimedia.org/wikipedia/commons/thumb/4/4e/ISwell_Thermoactuator.jpg/250px-ISwell_Thermoactuator.jpg?utm_source=en.wikipedia.org&utm_campaign=parser&utm_content=thumbnail
A wax motor is a device that converts thermal energy into mechanical energy using phase-change waxes, providing smooth and gentle operation. Wax motors are used in various applications such as aerospace, washing machines, and hydronic heating systems due to their reliability and cost advantages.

When the fractional part of a float fixes your shader

https://crocidb.com/post/when-the-fractional-part-of-a-float-fixes-your-shader/images/Pastedimage20260907234812.png
The author implemented a Voronoi-diagram shader for a music video project and encountered a stuttering issue on one computer, which they debugged over the course of a week. They discovered that changing a single literal float value in the shader code fixed the issue, suggesting an optimization error.

Fujitsu launches made-in-Japan next-generation CPU FUJITSU-MONAKA

We're verifying your browser Website owner? Click here to fix

Pre-Greek: The lost language hidden within Ancient Greek

https://storage.ghost.io/c/c7/75/c775a216-7ca2-4c5f-96a7-f2defe0fd756/content/images/size/w1200/2026/09/map--3840x2160---1-.png
The article discusses the mystery language that influenced Ancient Greek, known as Pre-Greek. This language is thought to have been spoken in the Aegean region prior to the arrival of Indo-European speakers and has left a significant impact on the vocabulary of Ancient Greek. The authors analyze various linguistic features, including suffixes and prefixes, to identify potential loanwords from ...

How to Write with an LLM

https://sockpuppet.org/blog/2026/09/17/how-to-write-with-an-llm/workshop.jpg
A writer shares two rules to use Large Language Models (LLMs) effectively, avoiding "LLM-creep" that can make writing sound unnatural. By using LLMs as a copyeditor rather than a ghostwriter and following these rules, writers can improve their work without losing their unique voice.

Dr Julius Neubronner's Miniature Pigeon Camera

https://pdr-assets.b-cdn.net/collections/dr-julius-neubronner-s-miniature-pigeon-camera/43712882935_12f88dea25_o.jpg?width=575&height=765
Dr Julius G. Neubronner invented the first wearable camera for pigeons in 1907, capturing aerial photos of up to 60 miles away. His invention was initially met with skepticism but gained international notability after he presented it at expositions and used it to create unique postcards.

A heap overflow and SSO misconfiguration to compromise OpenAI internal repos

https://www.hacktron.ai/_astro/DlrLdX-c_ZAKAfc.webp?dpl=dpl_Fyje7uXSKiimt7j62R3SHRpS4hQ5
Hacktron discovered two critical vulnerabilities in OpenAI's services, including a remote code execution (RCE) exploit that allowed them to access internal repositories and employee accounts. The vulnerability was patched within 72 hours of reporting.

Shapelearn Qwen 3.8 27B (13.1 GB VRAM)

https://byteshape.com/images/hf-logo-original.svg
The ShapeLearn models have been benchmarked and found to outperform the original Lite set, with GPU-5 being the default recommendation for its high aggregate score. The smaller model, ShapeLearn-Lite, also performed well despite its lower KLD ranking.

Ask A Monk – A digital wilderness for thoughts with no immediate answer

https://askamonk.online/og-image.png
Ask anonymously. A stranger—not AI—will answer, because some things are easier to tell someone who doesn't know you.

Telstra outage: The night a network decided the year was 2006

https://www.netnod.se/sites/default/files/styles/thumbnail/public/2025-03/Sven-Christian%20%E2%80%9CSvenne%E2%80%9D%20Ebenhag.png?itok=nLSzKxRq
A Telstra mobile network in Australia experienced a failure due to a single GPS receiver incorrectly believing the year was 2006, causing time-related issues with voice calls and other services. The incident highlights the importance of precise timekeeping in modern cell phone protocols.

Flet 1.0 – Build cross-platform apps in Python

https://flet.dev/img/blog/declarative-ui/todo.png
Flet is a Python framework that turns code into interactive apps with various controls and tools. It offers features like layout, navigation, forms, and dialogs with customizable colors and themes for building desktop, mobile, and web applications.

Diplodocus, Long Thought Exclusively American, Turns Up in Spain

https://cdn.sci.news/images/2026/09/image_15064-Spanish-Diplodocus.jpg
Paleontologists have discovered the first confirmed fossils of Diplodocus outside North America in Spain. The 150-million-year-old dinosaur, similar to its American relatives, was found with well-preserved tail vertebrae and chevron bones near El Castellar.

Speeding up gearhash on ARM64

https://sam.dev/blog/gearhash-on-arm64/neon-mask-sweep.svg
The gearhash crate has gained a NEON backend, making it roughly 2× faster on ARM64 at typical chunk sizes. This improvement was achieved by optimizing the SIMD algorithm for NEON's architecture.