Top

Write while learningpurplesyringa.moe

A Necessary History of the Oddest Letter: Wlithub.com

A historical essay tracing the origin of the letter W from post‑Roman scribal practices, its adoption via double‑U in early medieval Latin and Old English, competition with the rune wynn, and its later spelling quirks in English.

GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skillsarxiv.org

GraphSkillEvo encodes LLM agent skills as directed graphs and applies population‑based evolution (mutation, crossover) to optimise them. Across five benchmarks it beats the SkillOpt baseline, gaining up to 4% accuracy on GPT‑5.4‑nano.

One-Electron Universewikipedia.org

The article explains the one‑electron universe hypothesis: Wheeler’s idea that all electrons and positrons are different segments of a single particle’s world‑line moving forward and backward in time, with historical notes on Feynman and Stueckelberg.

Compiler-style optimization for drawing via Skiaarxiv.org

The Millennium Problems for Biologymillenniumproblems.bio

A speculative list of twelve grand challenges for biology, each with detailed success criteria ranging from abiogenesis to engineered enzymes and regenerative medicine.

Key symbols we lost to time, pt. 2: The Mac sidearesluna.org

Back to the future: modules for Guix packages (2022)guix.info

Guix-Modules is a new tool that generates Environment Modules files from Guix packages, allowing HPC users to leverage Guix's reproducible software environments with their familiar `module` command interface. This bridges the gap between traditional HPC environment management and Guix's declarative, reproducible approach.

You can defeat the Dream Devourer from Chrono Trigger using an int overflowfandom.com

Pre-Greek: The lost language hidden within Ancient Greeklinguisticdiscovery.com

The article surveys the mysterious set of ~1,000 Ancient Greek words with no Indo‑European etymology, arguing they are borrowings from a now‑lost “Pre‑Greek” substrate spoken by earlier Neolithic farmers. It explains linguistic concepts (substrate, superstrate, adstrate), gives examples (olive, fig, cypress, etc.), and outlines how such borrowings reveal migration and contact patterns.

LLM Classification Is Feature Engineeringminimallysufficient.com

Self-Hosting Behind CGNATalvarezrosa.com

The Hierarchy of Moneygregorygundersen.com

When scanners miss the attack: how Cloudflare Client-Side Security protects storefrontscloudflare.com

Cloudflare’s Page Shield uses a graph‑neural‑network (GNN) to model JavaScript as a syntax‑tree graph, followed by a lightweight LLM for second‑opinion triage and an ensemble of frontier models for deep analysis. This pipeline caught eight malicious payloads across four distinct affiliate‑theft and backdoor techniques that traditional scanners missed, demonstrating the need for runtime, behavior‑…

Suspension of the de minimis administrative exemption for imports $800 or lesspersonalimportation.org

The page is a simple advocacy landing page for personal prescription importation, containing only contact info and a call‑to‑action. It provides no technical content, analysis, or engineering insight.

Frontier Labs Are Selling Garbage to Fools in Washingtonsubstack.com

Why Do Video Diffusion Models Violate Physics? Unveiling the Flaws in Attention Mechanismsarxiv.org

The paper investigates why video diffusion models often break physical laws, pinpointing excessive spatial decay from Rotary Position Embedding (RoPE) as the culprit. By analyzing cross‑attention trajectories and self‑attention patterns, the authors identify specific attention heads that drive motion planning. They propose a lightweight fix: scaling RoPE frequency per denoising step, which empiri…

Calibrating Teacher--Student Discrepancy for On-Policy Distillationarxiv.org

This paper introduces Calibrated On-Policy Distillation (Cal-OPD), a method to improve knowledge distillation by filtering out the teacher model's inherent deviations from the learning signal. Cal-OPD estimates and removes the teacher's self-deviation, leading to better student performance on mathematical reasoning tasks.

Landing the Space Shuttle – A Flying Machine and the Thrill of a Lifetimeeaa.org

A former NASA astronaut recounts the dead-stick landing of the Space Shuttle Atlantis from Mach 1 and 50,000 feet, a high-stakes maneuver with no go-around option. Extensive training, including 1,700 simulator approaches in a modified Gulfstream G-II, was crucial for mastering the shuttle's unique flight characteristics.

Show HN: A competition for small neural networks that play strategy gamestinybrains.dev

Apple releases iPhone Duo simulator and Xcode 27.1 betaapple.com

Apple announced a new iPhone Duo simulator and the Xcode 27.1 beta, but the release notes contain no technical details beyond the headline features.

Simple and Efficient Row-Level Securityacadia.engineering

Unable to summarize – the article content was not provided.

SiliconBench: Speed, Memory, and Fidelity for LLM Serving on Unified-Memory Desktopsarxiv.org

SiliconBench benchmarks nine Apple‑Silicon LLM serving engines on unified‑memory desktops, measuring speed, memory usage, and output fidelity across Qwen3, Qwen3.5, and Gemma‑4 models. It finds vllm‑metal doubles throughput at modest concurrency, memory budgets often fail to preserve headroom, and only three stacks satisfy all fidelity and model‑coverage requirements, with tensor‑parallel scaling…

Inside OpenAI’s agentic software factorypragmaticengineer.com

OpenAI has internalized Codex and ChatGPT Work as a universal coding agent, achieving >90% adoption across the company and reshaping tooling, CI/CD load, and engineering roles. The shift has forced a new agentic software factory, massive infra scaling, and a rethink of IDEs, PRs, and code reviews.

Negativland, Culture Jamming, and the Art of Making Something Newarchive.org

Negativland is a long‑running experimental sound collage collective that coined “culture jamming” and has built a niche audience through DIY releases, legal battles over sampling, and a massive archive of their freeform radio show. Their recent live show at the Internet Archive highlights remote collaboration and ongoing preservation efforts, but the piece offers no technical insight for software…

In-Context Robot Learning with VLM Agentsarxiv.org

GPT‑Policy is a framework that lets a large vision‑language model (e.g. GPT‑6 Astra) perform in‑context robot learning: a context compiler extracts visual transitions from demos, the VLM proposes tool actions, and a constrained controller verifies and executes them. Real‑robot experiments show that raw video demos improve success rates even without explicit action labels, and that providing align…

Learning Another Language May Be One of the Best Ways to Keep Your Brain Healthytheconversation.com

From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Interventionarxiv.org

The paper presents PARTS, a framework that augments a frozen pretrained robot policy with RL‑learned residuals on selected bottleneck subtasks, using local success rewards and minimal human resets. In real‑world bimanual and single‑arm tasks, PARTS more than doubles success rates with only minutes of robot rollouts, outperforming prior fine‑tuning methods.

One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agentsarxiv.org

The authors propose a category‑aware expert‑training loop (RRE) and a multi‑teacher on‑policy distillation (MOPD) to improve software‑engineering agents across heterogeneous task categories. On Pro‑618 and SWE‑bench Multilingual the final student model reaches 58.04% and 59.00% resolution, 5.39 and 2.78 points above the base.

GAVEL: Graph World Models for Verified and Efficient Long-Horizon LLM Task Planningarxiv.org

GAVEL augments LLM‑driven robot planners with an explicit graph world model that verifies actions, repairs violations, and reasons over belief distributions, boosting single‑task success from 41 % to 92 % and multi‑task success from 20 % to 93 % on BEHAVIOR‑1K.