Tag
#llm
16posts
GLM-5.2 just took the open-weights crown - and it thinks a lot
Z.ai's GLM-5.2 tops Artificial Analysis's open-weights leaderboard at 51 points, but HN is arguing about token bloat and whether benchmarks lie.
The Wildly Simple Architecture Running Our AI Moment
A ground-up walkthrough of how transformer-based LLMs actually work - tokens, embeddings, attention, and the next-token loop - and why HN engineers can't…
9Router: Free AI Coding With Auto-Fallback and 40% Token Savings
9Router is a local proxy that routes Claude Code, Cursor, and Cline through 40+ providers with auto-fallback, token compression, and zero-downtime quota…
Open CTFs Are Now Pay-to-Win, and Frontier AI Did It
A top-10 CTF competitor argues that Claude Opus 4.5 and GPT-5.5 have automated enough of the scoreboard that open CTFs no longer measure human security skill.
One Source, Two Surfaces: Anthropic's Financial Services Toolkit
10 Claude agents for investment banking, equity research, and fund admin, deployable as Cowork plugins or headless Managed Agents from a single YAML manifest.
OpenGame: The Agent That Runs the Game to See If It Works
OpenGame is a CLI agentic framework from CUHK MMLab that scaffolds, builds, and verifies browser games end-to-end from a plain-English prompt.
Four Principles to Stop LLMs Overengineering Your Code
A tiny CLAUDE.md and optional plugin that makes Claude Code ask, test, and edit surgically.
Context Hub: The Curated Doc Layer Every Coding Agent Needs
Context Hub (`chub`) gives AI coding agents current, versioned API docs on demand -- so they code from facts, not 18-month-old training weights.
MarkItDown: Everything Becomes Markdown
Microsoft's Python tool converts PDFs, Office docs, audio, and YouTube URLs to clean Markdown for LLM pipelines and RAG systems.
BitNet: The Era of 1-bit LLMs is Finally Here
Explore bitnet.cpp, Microsoft's official framework for 1-bit LLMs that replaces multiplications with additions for massive speedups.
Khoj: Self-Host Your Personal AI Second Brain Across Obsidian and Notion
Khoj is an open-source AI second brain to chat with local LLMs, index Obsidian and Notion notes, and run custom autonomous agents locally.
Context Engineering: The Discipline That Separates Good AI Agents from Great Ones
Open-source toolkit for Context Engineering that teaches AI agents to pull only the context they need - cited in academic research.
Massive Collection of 700+ AI System Prompts: Production Guide
Explore the massive collection of 700+ AI system prompts: Cursor, Manus, Windsurf, Devin, Claude Code, and v0 production agent architectures reverse-engineered.
5 Awesome Projects to Build with LM Studio's Local API
A quick guide on how to use LM Studio's Local Server for Python scripts, OpenClaw, AI coding assistants, and secure document chats.
LLM & RAG: The 'Smart Librarian' Mental Model
Why do LLMs hallucinate? A mastery guide to Retrieval Augmented Generation (RAG) - the architecture powering every serious AI product in 2026.
Prompt Engineering: The 'Director & Actor' Mental Model
Why does 'be concise' produce worse results than '3 bullet points'? A mastery guide to system prompts, few-shot examples, and chain-of-thought.