Popular：
Virtualization DNS security formal verification reachability analysis compiler errors macro conflict web extension development framework Bitmap Graphics API inconsistencies All Tags

VM Cloning and Linux Random Number Generation: Security Implications and Solutions

2025-04-12
VM Cloning and Linux Random Number Generation: Security Implications and Solutions

This document analyzes the security implications of restoring multiple VM clones from a single snapshot. Linux exposes three main RNG interfaces: /dev/random, /dev/urandom, and the getrandom syscall. Cloning VMs leads to inconsistent RNG states due to multiple parameters (like timer data or CPU HWRNG instruction outputs) being mixed into each result. The article examines different implementations of RNGs in newer and older kernels and proposes solutions: reinitializing the RNG after restore, using the virtio-rng device, and leveraging the VMGenID mechanism (introduced in Linux 5.18 and later) to address inconsistent RNG states after cloning.

Read more
(github.com)
Development VM cloning Linux RNG

PyReason: Explainable Inference Software for Graph-Based Reasoning

2025-04-12
PyReason: Explainable Inference Software for Graph-Based Reasoning

PyReason is a visual inference tool that uses logical rules and facts to reason over graph structures. Supporting annotated, real-valued graphs and temporal logic, it offers Python library installation, multi-core parallel support (Python 3.9 and 3.10 only), and comprehensive documentation and code examples. The software is published with a paper and licensed under trademark permission from the Arizona Board of Regents/Arizona State University.

Read more
(github.com)
Development graph reasoning temporal logic

Rust CUDA: Bringing High-Speed GPU Computing to Rust

2025-04-11
Rust CUDA: Bringing High-Speed GPU Computing to Rust

The Rust CUDA project aims to make Rust a top-tier language for extremely fast GPU computing using the CUDA Toolkit. It provides tools for compiling Rust to highly optimized PTX code and libraries for interfacing with existing CUDA libraries. Addressing past challenges in integrating Rust with CUDA, it offers a comprehensive suite of crates covering various aspects of the CUDA ecosystem, including GPU-side functions, CUDA driver API wrappers, and OptiX support for ray tracing. While still in early development, the project seeks to propel the Rust GPU computing industry forward.

Read more
(github.com)
Development

Container Tools: Automating Minimal Debian Container Image Builds

2025-04-11
Container Tools: Automating Minimal Debian Container Image Builds

Container Tools is a project automating the creation of minimal Debian-based root filesystems using debootstrap. It supports customization with specific packages and configurations, and integrates security scanning for containerized environments. Easily extensible to other distros and projects, it addresses the bloat, network inefficiency, and slow iteration times of traditional Dockerfile builds. It creates lightweight, efficient container images by streamlining the build process, including only necessary components. Pre-built images with Java, Kafka, and more are available. The final output is a .tar file importable and runnable via `docker import`.

Read more
(github.com)
Development container images automated build

ParticleOS: A Fully Customizable Immutable Linux Distribution

2025-04-11
ParticleOS: A Fully Customizable Immutable Linux Distribution

ParticleOS is a unique immutable Linux distribution that lets users build and sign their own images, giving them complete control over system configuration. Users choose the base distribution (currently Arch and Fedora are supported) and the packages they want. System updates are handled by cloning the repository and running mkosi commands. Building systemd from source is recommended to ensure all features work correctly. ParticleOS uses the user's keys for Secure Boot signing and provides detailed installation instructions, including USB drive installation and systemd-homed configuration. In virtual machines, the default root password and username are both 'particleos'.

Read more
(github.com)
Development immutable system secure boot

Mbed TLS Port for Classic Mac OS: A Retro-Tech Challenge

2025-04-11
Mbed TLS Port for Classic Mac OS: A Retro-Tech Challenge

A developer successfully ported Mbed TLS to Classic Mac OS 7/8/9, a remarkable feat. The project overcame numerous hurdles, including the limitations of C89/C90 compilers lacking modern C features and the idiosyncrasies of the Mac's file system. The developer implemented 64-bit integer emulation and a custom entropy collection system, ultimately enabling a basic HTTPS GET request on a classic Mac. While security limitations exist, the project showcases a passion for retro technology and impressive programming skills.

Read more
(github.com)
Development Retro Programming

Pledge: A Lightweight Reactive Programming Framework for Swift

2025-04-10
Pledge: A Lightweight Reactive Programming Framework for Swift

Pledge is a lightweight, thread-safe reactive programming framework for Swift that simplifies state management and event propagation. Unlike other frameworks with steep learning curves, Pledge focuses on solving everyday problems faced by developers. It offers thread-safe implementation, priority-based notifications, customizable queues, batch updates, rate limiting, and common functional operators. Using `PLObservable` and `PLGlobalStore`, developers can easily implement the observer pattern and global state management, improving code efficiency and maintainability.

Read more
(github.com)
Development Reactive Programming State Management

Turn Docstrings into LLM Functions: Introducing the smartfunc Library

2025-04-10
Turn Docstrings into LLM Functions: Introducing the smartfunc Library

The smartfunc library ingeniously transforms docstrings into LLM functions, simplifying interaction with large language models. Using decorators and Jinja2 templating, it converts docstring text into prompts, interacting with various LLM providers (like OpenAI) via the underlying llm library. smartfunc supports Pydantic models for defining response structures, asynchronous functions, system prompts, and a debug mode for easy troubleshooting, significantly boosting development efficiency. While its functionality is relatively streamlined, its simplicity and ease of use make it ideal for rapid prototyping.

Read more
(github.com)
Development

TVMC: Time-Varying Mesh Compression using Volume-Tracked Reference Meshes

2025-04-10
TVMC: Time-Varying Mesh Compression using Volume-Tracked Reference Meshes

The TVMC project introduces a novel approach to time-varying mesh compression. It leverages volume-tracked reference meshes, employing a multi-step pipeline including ARAP volume tracking, MDS for reference center generation, transformation quaternion computation, creation of a volume-tracked self-contact-free reference mesh, mesh deformation, displacement field computation, and Draco-based compression and evaluation. The project supports Windows and Ubuntu, offering detailed Docker build and run instructions alongside instructions for running on a local machine.

Read more
(github.com)
Development mesh compression volume tracking

Photon: A Blazing-Fast Rust/WebAssembly Image Processing Library

2025-04-10
Photon: A Blazing-Fast Rust/WebAssembly Image Processing Library

Photon is a high-performance Rust image processing library compiling to WebAssembly for safe, fast image manipulation on the web and natively. Supporting formats like PNG, JPEG, and WebP, it boasts over 96 customizable functions, covering image correction, resizing, convolutions, channel manipulation, transformations, monochrome effects, color adjustments, filters, watermarking, and blending. Available natively, via WebAssembly in browsers and Node.js, version 0.3.2 adds duotone filters, image rotation, and dithering. Get started with its comprehensive documentation and tutorials.

Read more
(github.com)
Development

RCSS: Rust-Flavored CSS Preprocessor

2025-04-10
RCSS: Rust-Flavored CSS Preprocessor

RCSS is a styling language bringing Rust-inspired syntax to CSS. Combining Rust's robustness with SASS-like features such as nesting and variables, it aims for cleaner, more maintainable styles. The current implementation boasts Rust-like syntax, supporting variables, nesting, and functions (currently without arguments), along with a VS Code extension for syntax highlighting. Future plans include adding support for functions with arguments, importing, a formatter, improved CSS output formatting, and better error handling and debugging tools. RCSS boasts impressive compilation speed, completing in a few hundred microseconds.

Read more
(github.com)
Development CSS preprocessor

WebGL Viewer for Sparse Voxel Scenes

2025-04-09
WebGL Viewer for Sparse Voxel Scenes

An interactive WebGL viewer for visualizing sparse voxel scenes from the Nvidia Sparse Voxels Rasterization paper. This viewer lets you explore and visualize the voxel radiance field from your web browser. Rendering is similar to the reference CUDA implementation. It features interactive camera controls (mouse and touch), performance metrics display (FPS), and supports loading custom PLY files. The project leveraged AI assistance, proving highly efficient for boilerplate code but less so for complex graphics debugging. Generated PLY files can be large; consider limiting voxel count.

Read more
(github.com)
Development Voxel Rendering

LispE: A Novel Lisp Dialect with Pattern Matching and Logic

2025-04-09
LispE: A Novel Lisp Dialect with Pattern Matching and Logic

LispE, a modern Lisp dialect developed by Naver, distinguishes itself from traditional Lisp implementations like Common Lisp, Scheme, and Clojure through innovative constructs: defpat, defmacro, and defpred. LispE extends Lisp's flexibility and macro system with advanced pattern matching, enhanced macro capabilities, and logic programming elements. defpat enables defining multiple functions under the same name, each triggered by a specific argument pattern; defmacro simplifies custom syntax creation using pattern matching and a $ operator; defpred integrates pattern matching with predicate-based evaluation and automatic backtracking, blending logic programming into the Lisp framework. Compared to other Lisps, LispE offers superior expressiveness and modularity.

Read more
(github.com)
Development

Running Windows XP and 2003 on the Original Apple TV!

2025-04-09
Running Windows XP and 2003 on the Original Apple TV!

A developer successfully booted Windows XP and Windows Server 2003 on the original Apple TV after two years of work! This feat overcame significant hurdles due to the device's EFI-only firmware, incompatible with standard Windows. Using a custom FreeLoader bootloader and drivers, the developer achieved a bootable system with desktop access, though some features like PCI, USB, and audio remain partially or fully broken.

Read more
(github.com)
Tech

CodeScientist: An AI-Powered Tool for Automated Scientific Discovery – Costs and Risks

2025-04-09
CodeScientist: An AI-Powered Tool for Automated Scientific Discovery – Costs and Risks

CodeScientist is an autonomous agent leveraging LLMs for automated scientific discovery. It generates, debugs, and runs experiments, but costs vary depending on debugging iterations, prompt size, etc., averaging around $4 per experiment. Users must carefully manage API keys and monitor usage to avoid high costs. The generated code might contain API keys; exclusion patterns are recommended to prevent accidental commits.

Read more
(github.com)
Development Cost Management

Visualizing Linux Kernel Security: A Defense Map and Hardening Checker

2025-04-09
Visualizing Linux Kernel Security: A Defense Map and Hardening Checker

Linux kernel security is intricate. This project presents a visual map detailing the relationships between vulnerability classes, exploitation techniques, detection mechanisms, and defense technologies. The map, written in DOT language and rendered with GraphViz, aids navigation of documentation and kernel source code. Complementing the map is a tool, `kernel-hardening-checker`, automating the verification of Linux kernel security hardening options, particularly those often disabled by default in major distributions, thereby enhancing system security.

Read more
(github.com)
Development Linux kernel security security hardening vulnerability analysis

Modernized Dockerfile Formatter: dockerfmt

2025-04-09
Modernized Dockerfile Formatter: dockerfmt

Introducing dockerfmt, a modernized Dockerfile formatter built on top of the buildkit parser. It offers improved support for RUN commands (though grouping and semicolons are not yet supported), basic inline comment support, and various command-line options for checking, writing, indentation, and newline handling. JS bindings are also provided for easy integration. While features like line wrapping for long JSON commands and the # escape=X directive are not yet implemented, dockerfmt provides a user-friendly and effective way to format your Dockerfiles.

Read more
(github.com)
Development formatter

Coroot: Actionable Observability Without Code Changes

2025-04-08
Coroot: Actionable Observability Without Code Changes

Coroot is an open-source observability platform that automatically gathers metrics, logs, and traces without requiring any code modifications, turning this data into actionable insights. Leveraging eBPF for zero-instrumentation monitoring, it provides a service map, predefined inspections, application health summaries, distributed tracing, log analysis, and profiling capabilities. Coroot quickly identifies and resolves application issues, integrates with Kubernetes and major cloud platforms, and offers SLO tracking and cost monitoring to help developers optimize application performance and reduce cloud costs.

Read more
(github.com)
Development

HNSW: A Hierarchical Navigable Small World for Efficient Nearest Neighbor Search

2025-04-08
HNSW: A Hierarchical Navigable Small World for Efficient Nearest Neighbor Search

HNSW is a hierarchical navigable small world graph-based algorithm for nearest neighbor search of vector embeddings. It utilizes a hierarchical structure to speed up the search process. The algorithm builds sparse and dense graph structures at different levels, and searches efficiently from top to bottom. The code is concise, using modern C++ and Eigen for SIMD acceleration, requiring only about 500 lines of code.

Read more
(github.com)
Development nearest neighbor search vector embeddings

LLM Plugin: Summarize Hacker News Threads with Ease

2025-04-08
LLM Plugin: Summarize Hacker News Threads with Ease

A new LLM plugin, `llm-hacker-news`, lets you easily summarize Hacker News conversation threads. Simply install the plugin and use the command `llm -f hn:ID 'your instruction'` (e.g., `llm -f hn:43615912 'summary with illustrative direct quotes'`) to get a summary of the thread with the specified ID (found in the thread's URL). Installation and local setup instructions are provided in the README.

Read more
(github.com)
Development

Real-time Neuroplasticity: Giving Pre-trained LLMs Real-time Learning

2025-04-08
Real-time Neuroplasticity: Giving Pre-trained LLMs Real-time Learning

This experimental technique, called "Neural Graffiti," uses a plug-in called the "Spray Layer" to inject memory traces directly into the final inference stage of pre-trained large language models (LLMs) without fine-tuning or retraining. Mimicking the neuroplasticity of the brain, it subtly alters the model's "thinking" by modifying vector embeddings, influencing its generative token predictions. Through interaction, the model gradually learns and evolves. While not forcing specific word outputs, it biases the model towards associated concepts with repeated interaction. The aim is to give AI models more proactive behavior, focused personality, and enhanced curiosity, ultimately helping them achieve a form of self-awareness at the neuron level.

Read more
(github.com)
AI

Lightweight MCP Server: Real-time Weather Data for Claude

2025-04-07
Lightweight MCP Server: Real-time Weather Data for Claude

This project builds a lightweight Model Context Protocol (MCP) server enabling AI assistants like Claude to access and interpret real-time weather data. Users simply add the server to their Claude configuration, build the binary using `go build`, configure a weather API key, and can then query weather information for specific cities within Claude. The project features a modular design encompassing server handling, business logic, mock services for testing, and view templates, and is licensed under the MIT License.

Read more
(github.com)
Development MCP server real-time weather data

Kahuna: Your IndexedDB Swiss Army Knife

2025-04-07
Kahuna: Your IndexedDB Swiss Army Knife

Kahuna is a browser extension for Firefox and Chromium-based browsers that simplifies IndexedDB database management. It lets you create, modify, view, query, edit, import, and export IndexedDB data. Features include data filtering, pagination, JavaScript code execution, and import/export in various formats (Dexie, JSON, CSV). While documentation is a work in progress, Kahuna is a powerful tool for developers working with IndexedDB.

Read more
(github.com)
Development

OpenPrompt: Seamlessly Integrate Code into LLMs

2025-04-07
OpenPrompt: Seamlessly Integrate Code into LLMs

OpenPrompt simplifies the process of feeding code into large language models like Claude, GPT-4, and Grok. This tool rapidly serializes files and folders into XML, making it easy to upload your codebase. Available for Windows, macOS, and Linux (with executables provided), OpenPrompt lets you select directories, filter files, add instructions, and generate an XML prompt ready for pasting into your chosen LLM. Use cases include code reviews, documentation generation, refactoring assistance, bug hunting, learning new codebases, and architectural analysis.

Read more
(github.com)
Development

GitMCP: Effortlessly Access GitHub Project Documentation with AI

2025-04-07
GitMCP: Effortlessly Access GitHub Project Documentation with AI

GitMCP is a free, open-source service that seamlessly transforms any GitHub project into a remote Model Context Protocol (MCP) endpoint, allowing AI assistants to effortlessly access and understand project documentation. Zero setup is required; GitMCP works out of the box and is completely free and private, collecting no personally identifiable information or queries. Users access GitHub repositories or GitHub Pages sites via simple URL formats. AI assistants can access project documentation through GitMCP, utilizing semantic search to optimize token usage. GitMCP acts as a bridge between your GitHub repository's documentation and AI assistants by implementing the MCP, ensuring efficient and accurate information delivery.

Read more
(github.com)
Development

LLM Elimination Game: Social Reasoning, Strategy, and Deception

2025-04-07
LLM Elimination Game: Social Reasoning, Strategy, and Deception

Researchers created a multiplayer "elimination game" benchmark to evaluate Large Language Models (LLMs) in social reasoning, strategy, and deception. Eight LLMs compete, engaging in public and private conversations, forming alliances, and voting to eliminate opponents until only two remain. A jury of eliminated players then decides the winner. Analyzing conversation logs, voting patterns, and rankings reveals how LLMs balance shared knowledge with hidden intentions, forging alliances or betraying them strategically. The benchmark goes beyond simple dialogue, forcing models to navigate public vs. private dynamics, strategic voting, and jury persuasion. GPT-4.5 Preview emerged as the top performer.

Read more
(github.com)
AI Multi-agent Strategic Game

mem-isolate: Safely Running Unsafe Code

2025-04-06
mem-isolate: Safely Running Unsafe Code

mem-isolate executes your function via a fork(), waits for the result, and returns it to the parent process, preventing unsafe code from affecting the parent's memory footprint. It handles memory leaks and heap fragmentation, enforcing memory purity even for impure functions. Currently supporting only Unix-like systems, it adds approximately 1ms overhead compared to direct function calls—a reasonable trade-off for memory safety in critical applications.

Read more
(github.com)
Development memory isolation inter-process communication safe programming

TripoSG: High-Fidelity 3D Shape Synthesis with Large-Scale Rectified Flow Models

2025-04-06
TripoSG: High-Fidelity 3D Shape Synthesis with Large-Scale Rectified Flow Models

TripoSG is a cutting-edge foundation model for high-fidelity image-to-3D generation. Leveraging large-scale rectified flow transformers, hybrid supervised training, and a high-quality dataset, it achieves state-of-the-art results. TripoSG generates meshes with sharp features, fine details, and complex structures, accurately reflecting input image semantics. It boasts strong generalization capabilities, handling diverse input styles. A 1.5B parameter model, along with inference code and an interactive demo, is now available.

Read more
(github.com)
AI AI generation

Erica Synths Opensources its DIY Eurorack Modules

2025-04-06
Erica Synths Opensources its DIY Eurorack Modules

Following the discontinuation of its legacy DIY eurorack projects, Erica Synths has decided to open-source them. They've provided Gerber files for PCB-made front panels and made slight design changes to eliminate custom and rare components (though some modules still require rare ICs available at www.ericasynths.lv). Each folder contains complete build information and files: schematics, front panel Gerber files, PCB Gerber files, BOMs, component placement with values and designators, and assembly manuals (note that some manuals are for older module versions). Erica Synths permits third-party use for personal, educational, or commercial purposes, but will provide no support; refer to forums (www.muffwiggler.com) for build nuances, troubleshooting, part availability, and alternatives. Open-sourced modules include: Bassline, BBD delay/flanger, Delay, Dual VCA, Envelope, MIDI-CV, Mixer, Modulator, Output, Polivoks VCF, Swamp, and VCO3.

Read more
(github.com)
Hardware
1 2 … 25 26 27 29 31 32 33 … 50 51