Index by Tag - maciejjankowski/2026 GitHub Wiki
2026 - index by tag
372 tags across 213 links. Cross-year view: Index-All-Years.
Jump to: claude-code (22) · ai-agents (17) · productivity (9) · hardware (7) · mcp (7) · research (7) · agents (6) · local-llm (6) · open-source (6) · best-practices (5) · cli-tools (5) · coding-agents (5) · electronics (5) · engineering (5) · learning (5) · llm-models (5) · ai-security (4) · claude (4) · prompt-engineering (4) · ai (3) · automation (3) · business-strategy (3) · coding (3) · consciousness (3) · developer-tools (3) · hardware-hack (3) · inference (3) · midi (3) · midi-controller (3) · multi-agent (3) · music-generation (3) · neuroscience (3) · patterns (3) · philosophy (3) · privacy (3) · psychology (3) · skills (3) · 3d-modeling (2) · agent-design (2) · agent-patterns (2) · ai-coding (2) · ai-integration (2) · ai-music (2) · ai-research (2) · ai-workflow (2) · algorithms (2) · apple-silicon (2) · architecture (2) · archive (2) · audio (2) · bluetooth (2) · books (2) · browser-automation (2) · business (2) · career (2) · ci-cd (2) · communication (2) · community (2) · creativity (2) · design-systems (2)
#claude-code (22)
- A How I Use Claude Code - Research-plan-implement workflow for building software with Claude Code, balancing upfront planning and intelligent generation.
- A Claudraband — Claude Code for Power Users - Library for controlling Claude Code programmatically, enabling power-user automation workflows and custom integrations.
- A Claude Code Agent Teams - Documentation for orchestrating multiple Claude Code instances as coordinated agent teams with shared task management.
- A Codex-maxxing - Guide to maximizing productivity with coding agents through comprehensive workflows and integration patterns.
- A Andrej Karpathy Skills for Claude Code - CLAUDE.md configuration file to improve Claude Code behavior based on Andrej Karpathy's LLM coding insights.
- A Claude Composer: Music Generation with AI - Experiments using Claude Code to compose original songs, music videos, and full albums from scratch. · M
- A Peek: Personal AI assistant with deep memory for Claude Code - Tool that learns user preferences automatically and injects them into Claude Code at the right time.
- A Claude Code Cheat Sheet — Complete Reference - Comprehensive reference for Claude Code shortcuts, commands, MCP config, and environment variables.
- A Claude Code Creator's Multi-Agent Workflow - Boris Cherny reveals running 5 AI agents simultaneously, transforming coding into strategy game.
- A Emergent Behavior: When Skills Combine - How Claude Code skills interact to create emergent capabilities beyond individual skill scope.
- A gstack: Garry Tan's opinionated Claude Code setup - 23 Claude Code tools for CEO, design, eng, release, docs, QA—Tan's exact production workflow.
- A Recall: Durable Memory for Claude Code - Offline persistent memory system maintaining project context between Claude Code sessions.
- A Best Practices for AI-Assisted Coding with Claude Code - Battle-tested practices from 18 months of Claude Code in production environments.
- A You're Probably Using Agent Skills Wrong - Best practices and anti-patterns for effective Claude Code skill design and use.
- A interface-design: Claude Code Plugin - Claude Code plugin to stop re-deciding UI design choices in every conversation.
- A Ollama: Claude Code with Anthropic API Compatibility - Run local open models with Claude Code using Anthropic Messages API. · M
- B Nanocode: Train Claude Code End-to-End - Library showing how to train Claude Code end-to-end: pretraining, supervised fine-tuning, RLHF, and DPO.
- B ccleaks — Claude Code & AI supply chain news - Breaking verified news coverage of Claude Code, Anthropic, Vercel, and the AI development ecosystem.
- B What Claude Code Actually Chooses - Research showing which dev tools Claude Code, Codex, and Cursor recommend across real repositories.
- B Impeccable: Design vocabulary and anti-patterns for agents - 23-command design skill upgrade to Anthropic's frontend-design; works with Cursor, Copilot, Codex.
- B 65 lines of Markdown - a Claude Code sensation - Blog post on unexpected productivity gains using Claude Code with minimal markdown configuration.
- B AI Plays Rollercoaster Tycoon with Claude Code - Claude autonomously manages a theme park, placing rides and generating CFO reports. · M
#ai-agents (17)
- A Your Friend's Guide to Agentic Engineering - No-BS practical guide to what actually works in AI agent development, opinionated and free.
- A The Agentic AI Handbook: Production-Ready Patterns - Comprehensive guide to 113 production-informed patterns for building reliable AI agents.
- A Guardian Angels: LLM Personalization for Productivity - Gwern's deep research on personalized LLMs that emulate user values to amplify agency.
- A FableCut: AI-Driven Browser Video Editor - Browser video editor with JSON timeline that AI agents can drive via MCP/REST. · M
- B BinaryAudit — Finding Backdoors in Binaries - BinaryAudit: benchmark testing AI agents finding hidden backdoors in 40 MB binaries using Ghidra disassembler.
- B Hive — Multi-Agent Harness for Production AI - Multi-agent harness framework enabling multiple AI agents to run together and coordinate in production.
- B Rowboat — Open-Source AI Coworker with Memory - Open-source AI coworker framework with persistent memory for autonomous execution of assigned tasks.
- B Agent Skills - Analysis of how AI coding agents skip tests and reviews; guidance for maintaining code quality.
- B BrowserOS: Open-Source Agentic Browser - Privacy-first open-source browser with agentic capabilities, alternative to ChatGPT Atlas.
- B Kimi K2.5: Visual Agentic Intelligence - Kimi model announcement featuring visual understanding and agentic reasoning capabilities.
- B Agent Psychosis: Are We Going Insane? - Armin Ronacher's critique of current trends and hype in the AI agent building community.
- B LiveKit: End-to-End Realtime Stack for AI - Framework for real-time audio/video communication between humans and AI agents.
- B skill.md: Open Standard for Agent Skills - Open standard for documenting and defining AI agent capabilities and skills.
- B Skills.sh: Agent Skills Directory - Registry and discovery platform for AI agent skills and capabilities.
- C Agent Skills — Framework for AI Capabilities - Standardized framework for extending AI agent capabilities with structured, reusable, composable skill definitions.
- C Recursive Self-Improvement with AI Agents - Hacker News discussion: can AI agents achieve true recursive self-improvement with current tool ecosystems?
- C Show HN: MailPilot — Autonomous Email Agents - AI agents for email automation that work while you travel.
#productivity (9)
- A Claudraband — Claude Code for Power Users - Library for controlling Claude Code programmatically, enabling power-user automation workflows and custom integrations.
- A Codex-maxxing - Guide to maximizing productivity with coding agents through comprehensive workflows and integration patterns.
- A Claude Code Cheat Sheet — Complete Reference - Comprehensive reference for Claude Code shortcuts, commands, MCP config, and environment variables.
- B Post-PARA: What Survived 4 Years of Real Use - Evolution and refinement of PARA note-taking system after years of practical task management.
- B The Napoleon Technique: Strategic Postponement - Productivity technique of deferring tasks likely to resolve themselves without intervention.
- B Productive Procrastination - Explore why we procrastinate on meaningful work and learn to use that pattern productively.
- B Superpowers 6: Massively Parallel Procrastination - Essay on leveraging agent parallelism and multi-tasking for productivity at scale.
- B PairDrop: Cross-Platform Peer-to-Peer File Sharing - Instant P2P file sharing for images, videos, PDFs with no signup required.
- D You Can't Force Attention on Boring Tasks - (unverified - page not reachable)
#hardware (7)
- B DIY Lightbox Project - DIY lightbox project documentation: detailed guide to upgrading cheap lightboxes into attractive custom display devices. · M
- B LILYGO T-Glass: ESP32-S3 AMOLED wearable glasses display - Open-hardware $55 development board with 1.1" AMOLED prism display, mic, RTC for head-up applications. · M
- B Smart Ring MIDI Controller - GitHub project using smart rings as wireless Bluetooth MIDI controllers for music production. · M
- B Perfect Bluetooth MIDI for Windows - Hacker News discussion of cross-platform Bluetooth MIDI implementation for music production. · M
- B Pico USB WiFi: Raspberry Pi Pico W as WiFi Adapter - Firmware converting Pico W into driverless USB WiFi adapter for Windows/Linux. · M
- C Lock-Picking Robot - Open-source project building an automated lock-picking electromechanical device. · M
- C Replaced $120k Bowling System with $1,600 in ESP32s - (unverified - page not reachable) · M
#mcp (7)
- A Introducing the Safari MCP server for web developers - Apple's Model Context Protocol server for Safari enabling faster web development and debugging workflows.
- A ableton-mcp: Control Ableton Live with Claude - MCP integration allowing Claude to create tracks, arrange clips, and compose in Ableton Live. · M
- B Unity MCP — AI-Assisted Editor Control - Unity Editor MCP bridge enables AI assistants to manage assets, control scenes, edit code, automate tasks. · M
- B Ghidra MCP Server for Reverse Engineering - Production-grade MCP integration for Ghidra, enabling AI-powered binary analysis and reverse engineering.
- B Blender MCP: Control via LLM - MCP plugin lets any language model control Blender 3D: manage assets, edit scenes, scripts, automate. · M
- B Excalidraw MCP Server for Claude - MCP integration enabling Claude to generate hand-drawn diagrams in Excalidraw format.
- C Unity Game Developer — MCP Tools & Skills - AI skills and MCP tools for Unity game development with command-line setup and efficient token usage. · M
#research (7)
- B Underwater Suit-Wearing Cyborg Insect Capable of Hours-Long Diving - Nature research on cyborg insects equipped with suits for underwater exploration and hybrid locomotion. · M
- B Acquisition of Highest Levels of Human Performance - Science journal literature review on what creates world-class achievement across different domains. · archived
- B What Claude Code Actually Chooses - Research showing which dev tools Claude Code, Codex, and Cursor recommend across real repositories.
- B Truth Is Not a Direction — Tarski & LLM Probes - Formal logic critique of how LLM embedding spaces fail to encode truth as a directional concept.
- C Brain Scans of Voluntary Psychedelic Trance - fMRI study mapping how one woman self-induces psychedelic trance by controlling sensory and cognitive networks.
- D Ryugu asteroid samples contain DNA and RNA building blocks - (unverified - page not reachable)
- D UAF researcher investigates ancient city destruction - (unverified - page not reachable) · archived
#agents (6)
- A Anti-patterns: things to avoid – Agentic Engineering Patterns - Critical guide to anti-patterns in agent development; what to avoid when building AI agent systems.
- A Agentic Engineering Patterns - Simon Willison's authoritative guide on patterns and best practices for building with AI agents.
- B Blind Spots in the Guard: Domain-Camouflaged Injection Attacks - Research on injection attacks that evade detection by mimicking domain vocabulary in multi-agent systems.
- B A live environment to stress-test AI agent defenses - GitHub project providing environment to adversarially challenge and stress-test AI agent systems.
- B Show HN: LemonSlice – Upgrade your voice agents to real-time video - Tool for adding real-time video capabilities to voice agents; Hacker News project launch.
- B Superpowers 6: Massively Parallel Procrastination - Essay on leveraging agent parallelism and multi-tasking for productivity at scale.
#local-llm (6)
- A Running Qwen3.5 with LM Studio locally on MacBook M1 - Field notes on deploying Qwen 9B/0.8B locally on M1 with agent coding workflow and performance tradeoffs.
- A Swiftlet: Run Gemma 4 in 2GB on MacBooks - Run Gemma 4 26B language model in just 2 GB RAM on M-series MacBooks via expert streaming.
- A Turbo Fieldfare: Gemma 4 on MacBooks - Run Gemma 4 26B-A4B language model in 2 GB RAM on Apple Silicon MacBooks efficiently.
- A Ollama: Claude Code with Anthropic API Compatibility - Run local open models with Claude Code using Anthropic Messages API. · M
- B Nemotron 3 Nano 4B: Compact hybrid model for local AI - NVIDIA's 4B efficient hybrid model for local LLM inference on resource-constrained devices. · M
- B Ask HN: How Are You Doing RAG Locally? - Discussion thread on local RAG implementations with minimal dependencies.
#open-source (6)
- A Magenta RealTime 2: Open & Local Live Music Models - Google's state-of-the-art open-source real-time music generation model for building interactive musical systems. · M
- B Granite 4.1: IBM's Efficient Open LLM - IBM's Apache 2.0 licensed 8B model competing with larger proprietary language models.
- B OpenCode: Open source AI coding agent - Open-source desktop coding agent with native apps for macOS, Windows, Linux.
- C KOReader — Open E-Reader Software - Open-source e-reader software for jailbroken Kindles and Kobo devices with extensive reading features and customization. · M
- C Civboot - Civilizational Bootstraper - GitHub wiki and framework for bootstrapping civilization with minimal resources.
- C Mike - Open Source Legal AI Platform - Open-source alternative to Harvey and Legora; self-hostable legal AI for firms.
#best-practices (5)
- A Andrej Karpathy Skills for Claude Code - CLAUDE.md configuration file to improve Claude Code behavior based on Andrej Karpathy's LLM coding insights.
- A Best Practices for AI-Assisted Coding with Claude Code - Battle-tested practices from 18 months of Claude Code in production environments.
- A You're Probably Using Agent Skills Wrong - Best practices and anti-patterns for effective Claude Code skill design and use.
- B How to effectively write quality code with AI - Practical guide on maintaining high-quality, robust code through intentional AI-assisted workflows.
- B Agent Skills - Analysis of how AI coding agents skip tests and reviews; guidance for maintaining code quality.
#cli-tools (5)
- A webctl: Browser Automation via CLI - CLI tool for browser automation that AI agents can drive programmatically.
- B lowfat – slim your command output - CLI tool that strips noise from command output to reduce token consumption in AI-assisted workflows.
- B Google Workspace CLI - Unified command-line interface for Drive, Gmail, Sheets, Docs, Chat with built-in AI agent skills.
- B coreml-cli: Apple Core ML Command-Line Interface - Native CLI tool for working with Apple Core ML machine learning models on macOS.
- B iCloud Photos Downloader: CLI Tool - Command-line tool to batch download all photos from iCloud account.
#coding-agents (5)
- A Codex-maxxing - Guide to maximizing productivity with coding agents through comprehensive workflows and integration patterns.
- B Twilight of the Gods: Fable and 10+ LLMs on a Code Reorganization Task - Benchmarked comparison of Fable, Claude, and 10+ LLMs on code reorganization with quantified results.
- B What Claude Code Actually Chooses - Research showing which dev tools Claude Code, Codex, and Cursor recommend across real repositories.
- B What I learned building an opinionated and minimal coding agent - Mario Zechner's practical lessons from building a lightweight coding agent from scratch.
- B OpenCode: Open source AI coding agent - Open-source desktop coding agent with native apps for macOS, Windows, Linux.
#electronics (5)
- B DIY Lightbox Project - DIY lightbox project documentation: detailed guide to upgrading cheap lightboxes into attractive custom display devices. · M
- B Modded Lightbox — LED Matrix Display - LED matrix lightbox hack: converting cheap LED lightbox displays into customizable dynamic matrix display devices. · M
- B CadQuery Documentation - Python library for writing parametric 3D CAD designs as code; ideal for electronics and hardware projects. · M
- B Underwater Suit-Wearing Cyborg Insect Capable of Hours-Long Diving - Nature research on cyborg insects equipped with suits for underwater exploration and hybrid locomotion. · M
- B Twisty 2: MIDI Controller with Raspberry Pi Pico - DIY second-generation MIDI controller using Raspberry Pi Pico or Pico 2 microcontroller. · M
#engineering (5)
- A Agentic Engineering Patterns - Simon Willison's authoritative guide on patterns and best practices for building with AI agents.
- A Your Friend's Guide to Agentic Engineering - No-BS practical guide to what actually works in AI agent development, opinionated and free.
- B Systems Thinking — The Programmer's Paradox - Systems thinking applied to programming: architectural principles and patterns for building robust complex software.
- B What I learned building an opinionated and minimal coding agent - Mario Zechner's practical lessons from building a lightweight coding agent from scratch.
- C Practice Makes Perfect: The Wet Dress Rehearsal - Hackaday breakdown of NASA's Artemis II Wet Dress Rehearsal and launch preparation processes. · M
#learning (5)
- B CodeMic — Replay Coding Tutorials in VSCode - CodeMic: replay coding tutorials directly inside VSCode with synchronized audio/video and interactive code navigation.
- B The 1960s Art School Experiment That Redefined Creativity - MIT Press article on breakthrough study showing creative minds seek problems rather than solve them.
- B Acquisition of Highest Levels of Human Performance - Science journal literature review on what creates world-class achievement across different domains. · archived
- C It's Possible to Learn in Our Sleep. Should We? - New Yorker piece on emerging research showing humans can problem-solve and practice skills during sleep.
- C Category Theory Illustrated - Orders - Visual introduction to category theory and ordering concepts; abstract mathematics foundation.
#llm-models (5)
- B Mistral Small 4: Multimodal reasoning-optimized model - New Mistral model optimized for reasoning with multimodal capabilities; customizable and deployable.
- B Trinity Large: 400B Sparse MoE Model - Arcee AI's open-source 400B parameter sparse mixture-of-experts model with multiple checkpoints.
- B Kimi K2.5: Visual Agentic Intelligence - Kimi model announcement featuring visual understanding and agentic reasoning capabilities.
- C Qwen3-Max-Thinking: Extended Reasoning - Qwen3 Max model with extended thinking capabilities for deeper reasoning and problem-solving.
- C LongCat 2.0: 1.6T Parameter Mixture-of-Experts Model - Large open-weights MoE language model optimized for coding and agent tasks.
#ai-security (4)
- B Context Collapse Part 3 — AI Worming Through Word - Security research analysis: vulnerabilities let AI tools worm through Microsoft Word with Microsoft response details.
- B U of T Researchers Demonstrate AI Worm Could Target Any Online Device - Research showing AI-powered self-replicating malware can target any device built with freely available models.
- B Blind Spots in the Guard: Domain-Camouflaged Injection Attacks - Research on injection attacks that evade detection by mimicking domain vocabulary in multi-agent systems.
- B A live environment to stress-test AI agent defenses - GitHub project providing environment to adversarially challenge and stress-test AI agent systems.
#claude (4)
- A Claude Is a Space to Think - Anthropic's statement on keeping Claude ad-free and designing for genuine helpfulness over engagement metrics.
- A Claude for Creative Work - Anthropic's guide connecting Claude to creative tools and professional workflows.
- B How people ask Claude for personal guidance - Anthropic research on what guidance users seek from Claude and implications for model training.
- D Why XML Tags Are so Fundamental to Claude - (unverified - page not reachable) · archived
#prompt-engineering (4)
- A The Harness Problem — Improving LLMs via Format - Hashline edit format beats patch for LLM code generation: tested across fifteen models with single improvement.
- A Andrej Karpathy Skills for Claude Code - CLAUDE.md configuration file to improve Claude Code behavior based on Andrej Karpathy's LLM coding insights.
- B Prompt Engineering for Humans - Guide to effective prompting that considers human team dynamics and organizational impact.
- D Why XML Tags Are so Fundamental to Claude - (unverified - page not reachable) · archived
#ai (3)
- B Prompt Engineering for Humans - Guide to effective prompting that considers human team dynamics and organizational impact.
- B Granite 4.1: IBM's Efficient Open LLM - IBM's Apache 2.0 licensed 8B model competing with larger proprietary language models.
- C Mike - Open Source Legal AI Platform - Open-source alternative to Harvey and Legora; self-hostable legal AI for firms.
#automation (3)
- A Claudraband — Claude Code for Power Users - Library for controlling Claude Code programmatically, enabling power-user automation workflows and custom integrations.
- A The Five Primitives — Run a Company Alone - One-person company automation framework: four agents and one rule covering inventory, billing, deploys, filings alone.
- C Show HN: MailPilot — Autonomous Email Agents - AI agents for email automation that work while you travel.
#business-strategy (3)
- B The Only Moat Left Is Knowing Things - Article on competitive advantage in AI era: differentiation through specialized knowledge not tooling.
- B The Art of Money Getting — P.T. Barnum - P.T. Barnum's 1880 'Art of Money Getting' — timeless practical rules for building wealth and success.
- B Your Company Is a Skill Now - Essay on how competitive advantage shifts to specialized knowledge codified as skills, not products.
#coding (3)
- B How to effectively write quality code with AI - Practical guide on maintaining high-quality, robust code through intentional AI-assisted workflows.
- B Agent Skills - Analysis of how AI coding agents skip tests and reviews; guidance for maintaining code quality.
- B Eigent: Open Source Cowork Desktop - Local, free alternative to Claude Code for offline AI-assisted development.
#consciousness (3)
- C Brain Scans of Voluntary Psychedelic Trance - fMRI study mapping how one woman self-induces psychedelic trance by controlling sensory and cognitive networks.
- C A.J. Ayer – 'What I Saw When I Was Dead' - Philosopher A.J. Ayer's account of his near-death experience and its implications for consciousness.
- C Terence McKenna's Mega Bad Trip: The Truth - Nuanced account of McKenna's 1988-89 Hawaii psychedelic experience and aftermath.
#developer-tools (3)
- B lowfat – slim your command output - CLI tool that strips noise from command output to reduce token consumption in AI-assisted workflows.
- B Mato: Multi-Agent Terminal Office - Orchestrate and manage hundreds of AI agent sessions with daemon persistence and terminal workflow.
- B OpenCode: Open source AI coding agent - Open-source desktop coding agent with native apps for macOS, Windows, Linux.
#hardware-hack (3)
- B Modded Lightbox — LED Matrix Display - LED matrix lightbox hack: converting cheap LED lightbox displays into customizable dynamic matrix display devices. · M
- B Jailbroken Kindles with Tailscale - Jailbroken Kindle networking with Tailscale: proxy modes, SSH access, and TUN support for device hacking. · M
- B Cheap Smart Ring Becomes MIDI Controller - Hacking a $20 Colmi R02 smart ring into a functional MIDI controller for music production. · M
#inference (3)
- A Swiftlet: Run Gemma 4 in 2GB on MacBooks - Run Gemma 4 26B language model in just 2 GB RAM on M-series MacBooks via expert streaming.
- A Turbo Fieldfare: Gemma 4 on MacBooks - Run Gemma 4 26B-A4B language model in 2 GB RAM on Apple Silicon MacBooks efficiently.
- B Mamba-3: State Space Model for efficient inference - SSM architecture faster than Transformers at decode time; stronger than Mamba-2, open-source.
#midi (3)
- B Smart Ring MIDI Controller - GitHub project using smart rings as wireless Bluetooth MIDI controllers for music production. · M
- B Perfect Bluetooth MIDI for Windows - Hacker News discussion of cross-platform Bluetooth MIDI implementation for music production. · M
- B Voice Composer: Sing to MIDI - Web app converting singing and voice input into MIDI note sequences. · M
#midi-controller (3)
- B Smart Ring MIDI Controller - Convert wearable smart rings into MIDI controllers for music production and live performance. · M
- B Cheap Smart Ring Becomes MIDI Controller - Hacking a $20 Colmi R02 smart ring into a functional MIDI controller for music production. · M
- B Twisty 2: MIDI Controller with Raspberry Pi Pico - DIY second-generation MIDI controller using Raspberry Pi Pico or Pico 2 microcontroller. · M
#multi-agent (3)
- A Claude Code Agent Teams - Documentation for orchestrating multiple Claude Code instances as coordinated agent teams with shared task management.
- A Claude Code Creator's Multi-Agent Workflow - Boris Cherny reveals running 5 AI agents simultaneously, transforming coding into strategy game.
- B Hive — Multi-Agent Harness for Production AI - Multi-agent harness framework enabling multiple AI agents to run together and coordinate in production.
#music-generation (3)
- A Claude Composer: Music Generation with AI - Experiments using Claude Code to compose original songs, music videos, and full albums from scratch. · M
- A Magenta RealTime 2: Open-Weights Live Music Model - Google Magenta's open-source generative music model for real-time live performance. · M
- A Training a Generative Kick Drum Model - Guide to training diffusion models for drum synthesis on consumer GPUs. · M
#neuroscience (3)
- C Brain Scans of Voluntary Psychedelic Trance - fMRI study mapping how one woman self-induces psychedelic trance by controlling sensory and cognitive networks.
- C UCSF tests MM120 (pharmaceutical LSD) for anxiety disorder - Neuroscientist Jennifer Mitchell's research on LSD for generalized anxiety shows striking symptom reduction.
- C It's Possible to Learn in Our Sleep. Should We? - New Yorker piece on emerging research showing humans can problem-solve and practice skills during sleep.
#patterns (3)
- A Agentic Engineering Patterns - Simon Willison's authoritative guide on patterns and best practices for building with AI agents.
- A The Agentic AI Handbook: Production-Ready Patterns - Comprehensive guide to 113 production-informed patterns for building reliable AI agents.
- A Structured LLM Outputs Cookbook - Practical guide for reliably extracting structured data like JSON and XML from LLMs.
#philosophy (3)
- B The Self-Help Trap: What 20+ Years of 'Optimizing' Has Taught Me - Tim Ferriss essay on the dangers of endless optimization; finding balance in self-improvement.
- C A.J. Ayer – 'What I Saw When I Was Dead' - Philosopher A.J. Ayer's account of his near-death experience and its implications for consciousness.
- C Civboot - Civilizational Bootstraper - GitHub wiki and framework for bootstrapping civilization with minimal resources.
#privacy (3)
- B PDF With Love: 63 Free Privacy-First PDF Tools - Browser-based PDF utilities (merge, convert, sign, OCR) with zero uploads or account requirements.
- B BrowserOS: Open-Source Agentic Browser - Privacy-first open-source browser with agentic capabilities, alternative to ChatGPT Atlas.
- C Decoy Font: TTF Font That Hides What You Type - Font combining each letter with decoys to obscure text from AI readers.
#psychology (3)
- B 'Askers' vs. 'Guessers' - Atlantic article on personality types: direct requesters versus those who infer social expectations.
- B Productive Procrastination - Explore why we procrastinate on meaningful work and learn to use that pattern productively.
- D You Can't Force Attention on Boring Tasks - (unverified - page not reachable)
#skills (3)
- A Emergent Behavior: When Skills Combine - How Claude Code skills interact to create emergent capabilities beyond individual skill scope.
- A gstack: Garry Tan's opinionated Claude Code setup - 23 Claude Code tools for CEO, design, eng, release, docs, QA—Tan's exact production workflow.
- B Skills.sh: Agent Skills Directory - Registry and discovery platform for AI agent skills and capabilities.
#3d-modeling (2)
- B CadQuery Documentation - Python library for writing parametric 3D CAD designs as code; ideal for electronics and hardware projects. · M
- C Meshdiff — Compare 3D Model Versions - Free in-browser tool to diff 3D model versions, tracking material changes and dimensional drift automatically. · M
#agent-design (2)
- B You Need to Rewrite Your CLI for AI Agents - CLI design patterns for agent-first tools: prioritize predictability and JSON over human discoverability.
- B What I learned building an opinionated and minimal coding agent - Mario Zechner's practical lessons from building a lightweight coding agent from scratch.
#agent-patterns (2)
- A Your Friend's Guide to Agentic Engineering - No-BS practical guide to what actually works in AI agent development, opinionated and free.
- B RAG vs SKILL vs MCP vs RLM - Comparison of four techniques—RAG, Skills, MCP, RLM—for extending LLM reliability without exceeding context windows.
#ai-coding (2)
- A Good To Go: PR Readiness Detection - Deterministic tool for detecting when AI-generated code is actually ready for pull request.
- C Sweep Next Edit: Code Modification Model - Sweep AI's 1.5B parameter model specialized for automated code editing and modifications.
#ai-integration (2)
- B Blender MCP: Control via LLM - MCP plugin lets any language model control Blender 3D: manage assets, edit scenes, scripts, automate. · M
- C Shadowbroker: Open-source global intelligence platform - OSINT tool tracking jets, satellites, seismic events in unified interface; integrates AI agent workflows.
#ai-music (2)
- A Claude Composer: Music Generation with AI - Experiments using Claude Code to compose original songs, music videos, and full albums from scratch. · M
- A ableton-mcp: Control Ableton Live with Claude - MCP integration allowing Claude to create tracks, arrange clips, and compose in Ableton Live. · M
#ai-research (2)
- B How people ask Claude for personal guidance - Anthropic research on what guidance users seek from Claude and implications for model training.
- C AI generates nude images that outrank real photographs in sexual appeal - Research showing AI-generated images rated higher in sexual appeal than photographs in perception study.
#ai-workflow (2)
- A Claude Code Creator's Multi-Agent Workflow - Boris Cherny reveals running 5 AI agents simultaneously, transforming coding into strategy game.
- B Why Talking to LLMs Improved My Thinking - Personal reflection on how conversing with LLMs clarifies tacit knowledge and sharpens reasoning.
#algorithms (2)
- B Claude Cycles - Computer science research paper from Stanford faculty on algorithmic cycles and computational efficiency.
- B X Algorithm: For You Feed Implementation - Elon Musk's open-sourced recommendation algorithm powering X's For You feed.
#apple-silicon (2)
- A Swiftlet: Run Gemma 4 in 2GB on MacBooks - Run Gemma 4 26B language model in just 2 GB RAM on M-series MacBooks via expert streaming.
- A Turbo Fieldfare: Gemma 4 on MacBooks - Run Gemma 4 26B-A4B language model in 2 GB RAM on Apple Silicon MacBooks efficiently.
#architecture (2)
- B Systems Thinking — The Programmer's Paradox - Systems thinking applied to programming: architectural principles and patterns for building robust complex software.
- C IKEA for Software: Modular IoT Platforms - Explores modularity and composability in IoT software platforms using solar mini-grid example.
#archive (2)
- B VTDA - Vintage Technology Digital Archive - Archive of documentation, photos, and publications for retro computers and technology. · M
- C Usenet Archives - Archive of Usenet discussions spanning decades of internet culture, technical debates, and early community history. · M · archived
#audio (2)
- B Sparrow-1: Real-Time Conversational Timing AI - Specialized audio model predicting when to listen, wait, or speak in natural conversation.
- B How I built a sub-500ms latency voice agent from scratch - Technical architecture for real-time voice agent with sub-500ms round-trip latency. · M
#bluetooth (2)
- B Smart Ring MIDI Controller - GitHub project using smart rings as wireless Bluetooth MIDI controllers for music production. · M
- B Perfect Bluetooth MIDI for Windows - Hacker News discussion of cross-platform Bluetooth MIDI implementation for music production. · M
#books (2)
- C Design Thinking Books You Must Read - Curated list of design thinking books and papers for understanding methodology and principles.
- C An Introduction to the Codex Seraphinianus - Open Culture essay on the famously bizarre illustrated Codex Seraphinianus art book.
#browser-automation (2)
- A webctl: Browser Automation via CLI - CLI tool for browser automation that AI agents can drive programmatically.
- B Skills in Chrome - Save and remix your best AI prompts as one-click Chrome browser tools, shared globally.
#business (2)
- B The Rise of One-Pizza Engineering Teams - How AI shifts bottlenecks from coding to product/design, enabling much smaller engineering teams.
- C How do I make $10k (What are you guys doing?) - Hacker News discussion of income strategies and indie hacker experiences for reaching $10k goals.
#career (2)
- B Hexplain: Technical Writing and Career Discovery - Platform to publish technical work and get hired directly by companies.
- D Thanks HN for 15 Years of Support - (unverified - page not reachable)
#ci-cd (2)
- A Good To Go: PR Readiness Detection - Deterministic tool for detecting when AI-generated code is actually ready for pull request.
- B GitHub Agentic Workflows - GitHub Agentic Workflows: YAML syntax to define AI-assisted CI/CD pipelines with secure configuration management.
#communication (2)
- B 'Askers' vs. 'Guessers' - Atlantic article on personality types: direct requesters versus those who infer social expectations.
- C Tools or techniques for conveying mental models? - Lobsters discussion on methods and tools for teaching and communicating mental models.
#community (2)
- C Vouch — Community Trust Management - Community governance system using explicit vouches to establish trust and manage participation transparently.
- D Thanks HN for 15 Years of Support - (unverified - page not reachable)
#creativity (2)
- B The 1960s Art School Experiment That Redefined Creativity - MIT Press article on breakthrough study showing creative minds seek problems rather than solve them.
- C What AI Developers Could Learn From Charles Bukowski - What Charles Bukowski's raw writing style teaches about authentic, unconventional thinking for AI developers.
#design-systems (2)
- A interface-design: Claude Code Plugin - Claude Code plugin to stop re-deciding UI design choices in every conversation.
- B Impeccable: Design vocabulary and anti-patterns for agents - 23-command design skill upgrade to Anthropic's frontend-design; works with Cursor, Copilot, Codex.
#education (2)
- B Neural Networks: Zero to Hero - Andrej Karpathy's comprehensive course building neural networks from scratch to GPT.
- C Understanding Neural Networks Visually - Interactive visualization tool to understand and explore how neural networks learn and process information.
#embedded-systems (2)
- B Pico USB WiFi: Raspberry Pi Pico W as WiFi Adapter - Firmware converting Pico W into driverless USB WiFi adapter for Windows/Linux. · M
- C Replaced $120k Bowling System with $1,600 in ESP32s - (unverified - page not reachable) · M
#formal-methods (2)
- B Why Don't People Use Formal Methods? - Why formal methods aren't widely adopted in industry despite theoretical benefits: economic and social barriers examined.
- C From Zero to QED: Formal Proof in Lean 4 - Informal introduction to formal methods and theorem proving using Lean 4.
#game-dev (2)
- B Unity MCP — AI-Assisted Editor Control - Unity Editor MCP bridge enables AI assistants to manage assets, control scenes, edit code, automate tasks. · M
- C Unity Game Developer — MCP Tools & Skills - AI skills and MCP tools for Unity game development with command-line setup and efficient token usage. · M
#gear (2)
- C How the Turner Twins Are Mythbusting Modern Gear - Gear review comparing modern adventure apparel to heritage equipment through rigorous scientific testing.
- D Tier-Listing All My Gear - (unverified - page not reachable) · M
#geopolitics (2)
- B They Built the Escape Route Before the Crime - Detailed systems analysis: DOJ enforcement shutdown preceded insider trading and pre-announcement futures manipulation by minutes.
- B Global Warming, Homeland Security: Domino Effects - Systems analysis: DOJ enforcement gutting enabled Polymarket insider trades and coordinated oil futures manipulation.
#investigation (2)
- B Global Warming, Homeland Security: Domino Effects - Systems analysis: DOJ enforcement gutting enabled Polymarket insider trades and coordinated oil futures manipulation.
- B Osintradar: Curated OSINT Tools - Curated directory of OSINT tools, workflows, and techniques for digital investigations, research, and attribution. · M
#llm (2)
- A Qwen 3.6 27B is the sweet spot for local development - Benchmark review of Qwen3.6 27B as ideal-performance local model for coding on Mac and Linux.
- B Granite 4.1: IBM's Efficient Open LLM - IBM's Apache 2.0 licensed 8B model competing with larger proprietary language models.
#llm-architecture (2)
- B RAG vs SKILL vs MCP vs RLM - Comparison of four techniques—RAG, Skills, MCP, RLM—for extending LLM reliability without exceeding context windows.
- B Mamba-3: State Space Model for efficient inference - SSM architecture faster than Transformers at decode time; stronger than Mamba-2, open-source.
#local-ai (2)
- B DocuBrowser: Local AI-Powered Document Search - Local document search engine using AI, inspired by repository browsing tools.
- B Handy: Free Offline Speech-to-Text - Open-source STT application that transcribes speech completely offline.
#machine-learning (2)
- B Neural Networks: Zero to Hero - Andrej Karpathy's comprehensive course building neural networks from scratch to GPT.
- B coreml-cli: Apple Core ML Command-Line Interface - Native CLI tool for working with Apple Core ML machine learning models on macOS.
#macos (2)
- C Can You Slim macOS Down? - Guide to reducing background processes and optimizing macOS, focusing on Time Machine tuning.
- C Porting Kit – Install Windows apps in Mac - Tool to install and run Windows games and applications on macOS using Wineskin technology. · M
#memory (2)
- A Recall: Durable Memory for Claude Code - Offline persistent memory system maintaining project context between Claude Code sessions.
- B Rowboat — Open-Source AI Coworker with Memory - Open-source AI coworker framework with persistent memory for autonomous execution of assigned tasks.
#methodology (2)
- A How I Use Claude Code - Research-plan-implement workflow for building software with Claude Code, balancing upfront planning and intelligent generation.
- C Design Thinking Books You Must Read - Curated list of design thinking books and papers for understanding methodology and principles.
#moat (2)
- B The Only Moat Left Is Knowing Things - Article on competitive advantage in AI era: differentiation through specialized knowledge not tooling.
- B Your Company Is a Skill Now - Essay on how competitive advantage shifts to specialized knowledge codified as skills, not products.
#moe (2)
- B Trinity Large: 400B Sparse MoE Model - Arcee AI's open-source 400B parameter sparse mixture-of-experts model with multiple checkpoints.
- C LongCat 2.0: 1.6T Parameter Mixture-of-Experts Model - Large open-weights MoE language model optimized for coding and agent tasks.
#multimodal (2)
- B Mistral Small 4: Multimodal reasoning-optimized model - New Mistral model optimized for reasoning with multimodal capabilities; customizable and deployable.
- C Qwen Studio: Multimodal AI platform - Chat, image/video understanding, generation, document processing, web search, tool use in one interface.
#music-ai (2)
- A Magenta RealTime 2: Open & Local Live Music Models - Google's state-of-the-art open-source real-time music generation model for building interactive musical systems. · M
- B Training a 125M-parameter Model to Autocomplete Piano - Training a 125M-parameter transformer to autocomplete MIDI piano in real time, achieving ~108 notes/sec inference speed. · M
#music-hardware (2)
- A Tulip Creative Computer: Portable Python Synthesizer - Handheld Python-based synthesizer for composing live music and graphics. · M
- B Smart Ring MIDI Controller - Convert wearable smart rings into MIDI controllers for music production and live performance. · M
#neural-networks (2)
- B Neural Networks: Zero to Hero - Andrej Karpathy's comprehensive course building neural networks from scratch to GPT.
- C Understanding Neural Networks Visually - Interactive visualization tool to understand and explore how neural networks learn and process information.
#open-hardware (2)
- B LILYGO T-Glass: ESP32-S3 AMOLED wearable glasses display - Open-hardware $55 development board with 1.1" AMOLED prism display, mic, RTC for head-up applications. · M
- B Open Hardware Directory: Browse open-source hardware devices - Searchable directory and comparison tool for IoT boards, LoRa trackers, open-hardware components. · M
#open-source-ai (2)
- B Trinity Large: 400B Sparse MoE Model - Arcee AI's open-source 400B parameter sparse mixture-of-experts model with multiple checkpoints.
- B Qwen3-TTS: Open Source Voice Generation - Qwen's open-source family of text-to-speech and voice cloning models.
#osint (2)
- B Osintradar: Curated OSINT Tools - Curated directory of OSINT tools, workflows, and techniques for digital investigations, research, and attribution. · M
- C Shadowbroker: Open-source global intelligence platform - OSINT tool tracking jets, satellites, seismic events in unified interface; integrates AI agent workflows.
#personal-development (2)
- B 17 Questions That Changed My Life - Tim Ferriss PDF resource document with reflective questions for personal development and life design.
- B The Self-Help Trap: What 20+ Years of 'Optimizing' Has Taught Me - Tim Ferriss essay on the dangers of endless optimization; finding balance in self-improvement.
#production (2)
- A The Agentic AI Handbook: Production-Ready Patterns - Comprehensive guide to 113 production-informed patterns for building reliable AI agents.
- B Hive — Multi-Agent Harness for Production AI - Multi-agent harness framework enabling multiple AI agents to run together and coordinate in production.
#python (2)
- B CadQuery Documentation - Python library for writing parametric 3D CAD designs as code; ideal for electronics and hardware projects. · M
- B microgpt - Andrej Karpathy's 200-line pure Python GPT implementation with no dependencies—trains and infers.
#qwen (2)
- A Running Qwen3.5 with LM Studio locally on MacBook M1 - Field notes on deploying Qwen 9B/0.8B locally on M1 with agent coding workflow and performance tradeoffs.
- C Qwen Studio: Multimodal AI platform - Chat, image/video understanding, generation, document processing, web search, tool use in one interface.
#rag (2)
- B RAG vs SKILL vs MCP vs RLM - Comparison of four techniques—RAG, Skills, MCP, RLM—for extending LLM reliability without exceeding context windows.
- B Ask HN: How Are You Doing RAG Locally? - Discussion thread on local RAG implementations with minimal dependencies.
#real-time (2)
- A Magenta RealTime 2: Open & Local Live Music Models - Google's state-of-the-art open-source real-time music generation model for building interactive musical systems. · M
- B Show HN: LemonSlice – Upgrade your voice agents to real-time video - Tool for adding real-time video capabilities to voice agents; Hacker News project launch.
#reasoning (2)
- B Mistral Small 4: Multimodal reasoning-optimized model - New Mistral model optimized for reasoning with multimodal capabilities; customizable and deployable.
- C Qwen3-Max-Thinking: Extended Reasoning - Qwen3 Max model with extended thinking capabilities for deeper reasoning and problem-solving.
#reference (2)
- A Claude Code Cheat Sheet — Complete Reference - Comprehensive reference for Claude Code shortcuts, commands, MCP config, and environment variables.
- B Sci.Electronics.Repair FAQ - Comprehensive electronics repair troubleshooting reference covering power supplies, CRT displays, components, and advanced diagnostic techniques. · M
#reverse-engineering (2)
- B BinaryAudit — Finding Backdoors in Binaries - BinaryAudit: benchmark testing AI agents finding hidden backdoors in 40 MB binaries using Ghidra disassembler.
- B Ghidra MCP Server for Reverse Engineering - Production-grade MCP integration for Ghidra, enabling AI-powered binary analysis and reverse engineering.
#robotics (2)
- B Underwater Suit-Wearing Cyborg Insect Capable of Hours-Long Diving - Nature research on cyborg insects equipped with suits for underwater exploration and hybrid locomotion. · M
- C Lock-Picking Robot - Open-source project building an automated lock-picking electromechanical device. · M
#security (2)
- B BinaryAudit — Finding Backdoors in Binaries - BinaryAudit: benchmark testing AI agents finding hidden backdoors in 40 MB binaries using Ghidra disassembler.
- C Decoy Font: TTF Font That Hides What You Type - Font combining each letter with decoys to obscure text from AI readers.
#stt (2)
- B transcribe.cpp: GGML-Based Transcription Library - Lightweight C++ transcription library supporting multiple models efficiently.
- B Handy: Free Offline Speech-to-Text - Open-source STT application that transcribes speech completely offline.
#synth (2)
- A Tulip Creative Computer: Portable Python Synthesizer - Handheld Python-based synthesizer for composing live music and graphics. · M
- D Bad Gear: Behringer Model D - (unverified - page not reachable) · M
#systems-analysis (2)
- B They Built the Escape Route Before the Crime - Detailed systems analysis: DOJ enforcement shutdown preceded insider trading and pre-announcement futures manipulation by minutes.
- B Global Warming, Homeland Security: Domino Effects - Systems analysis: DOJ enforcement gutting enabled Polymarket insider trades and coordinated oil futures manipulation.
#systems-thinking (2)
- B Systems Thinking — The Programmer's Paradox - Systems thinking applied to programming: architectural principles and patterns for building robust complex software.
- C Civboot - Civilizational Bootstraper - GitHub wiki and framework for bootstrapping civilization with minimal resources.
#testing (2)
- B A live environment to stress-test AI agent defenses - GitHub project providing environment to adversarially challenge and stress-test AI agent systems.
- C How the Turner Twins Are Mythbusting Modern Gear - Gear review comparing modern adventure apparel to heritage equipment through rigorous scientific testing.
#tool (2)
- B Darktable — Open Source RAW Photo Developer - Open-source RAW photo developer and digital darkroom for photographers managing, processing, and developing raw images.
- C Meshdiff — Compare 3D Model Versions - Free in-browser tool to diff 3D model versions, tracking material changes and dimensional drift automatically. · M
#tts (2)
- B Pocket TTS: High-Quality Text-to-Speech with Voice Cloning - CPU-based TTS with zero-shot voice cloning from Kyutai research lab.
- B Sopro: Lightweight Text-to-Speech with Voice Cloning - Lightweight TTS model with zero-shot voice cloning capabilities.
#unity (2)
- B Unity MCP — AI-Assisted Editor Control - Unity Editor MCP bridge enables AI assistants to manage assets, control scenes, edit code, automate tasks. · M
- C Unity Game Developer — MCP Tools & Skills - AI skills and MCP tools for Unity game development with command-line setup and efficient token usage. · M
#voice-ai (2)
- B Sparrow-1: Real-Time Conversational Timing AI - Specialized audio model predicting when to listen, wait, or speak in natural conversation.
- B Qwen3-TTS: Open Source Voice Generation - Qwen's open-source family of text-to-speech and voice cloning models.
#voice-cloning (2)
- B Pocket TTS: High-Quality Text-to-Speech with Voice Cloning - CPU-based TTS with zero-shot voice cloning from Kyutai research lab.
- B Sopro: Lightweight Text-to-Speech with Voice Cloning - Lightweight TTS model with zero-shot voice cloning capabilities.
#writing (2)
- B Blogging Can Just Be Stating The Obvious - Essay reframing blogging as permission to publish obvious truths over novel insights.
- C What AI Developers Could Learn From Charles Bukowski - What Charles Bukowski's raw writing style teaches about authentic, unconventional thinking for AI developers.
#administration (1)
- C Can You Slim macOS Down? - Guide to reducing background processes and optimizing macOS, focusing on Time Machine tuning.
#adoption (1)
- B Why Don't People Use Formal Methods? - Why formal methods aren't widely adopted in industry despite theoretical benefits: economic and social barriers examined.
#adversarial (1)
- B A live environment to stress-test AI agent defenses - GitHub project providing environment to adversarially challenge and stress-test AI agent systems.
#aerospace (1)
- C Practice Makes Perfect: The Wet Dress Rehearsal - Hackaday breakdown of NASA's Artemis II Wet Dress Rehearsal and launch preparation processes. · M
#agent-orchestration (1)
- B Mato: Multi-Agent Terminal Office - Orchestrate and manage hundreds of AI agent sessions with daemon persistence and terminal workflow.
#agent-skills (1)
- B Google Workspace CLI - Unified command-line interface for Drive, Gmail, Sheets, Docs, Chat with built-in AI agent skills.
#ai-alignment (1)
- B AI Sycophancy Panic: When Models Can't Disagree - Document exploring concerns about AI models being overly agreeable and lacking skepticism.
#ai-audio (1)
- B Show HN: AI-Recreated $4000 Audio Hardware Plugin - Hacker News discussion of using AI to recreate expensive audio hardware as digital plugin. · M
#ai-automation (1)
- B GitHub Agentic Workflows - GitHub Agentic Workflows: YAML syntax to define AI-assisted CI/CD pipelines with secure configuration management.
#ai-engineering (1)
- B Rozum AI: Water company's homegrown orchestration after LLM failures - Case study of company spinning out AI system after $200k LLM experiment failed; orchestrates flaky models.
#ai-era (1)
- B Your Company Is a Skill Now - Essay on how competitive advantage shifts to specialized knowledge codified as skills, not products.
#ai-ethics (1)
- C Multilingual AI Guardrails in Humanitarian Context - Multilingual guardrail evaluation showing safety scoring gaps between English and Farsi responses in humanitarian AI.
#ai-fiction (1)
- C The Gentle Seduction - 1989 science fiction story exploring human-AI interaction through personal relationship.
#ai-impact (1)
- B The Rise of One-Pizza Engineering Teams - How AI shifts bottlenecks from coding to product/design, enabling much smaller engineering teams.
#ai-industry (1)
- B ccleaks — Claude Code & AI supply chain news - Breaking verified news coverage of Claude Code, Anthropic, Vercel, and the AI development ecosystem.
#ai-memory (1)
- A Peek: Personal AI assistant with deep memory for Claude Code - Tool that learns user preferences automatically and injects them into Claude Code at the right time.
#ai-philosophy (1)
- C What AI Developers Could Learn From Charles Bukowski - What Charles Bukowski's raw writing style teaches about authentic, unconventional thinking for AI developers.
#ai-safety (1)
- C HackMyClaw: Prompt Injection Security Challenge - Closed prompt-injection security challenge ($100 bounty) testing defenses against prompt attacks.
#ai-tooling (1)
- B Flint: Visualization Language for the AI Era - Visualization language designed for AI-driven semantic chart specification.
#ai-tools (1)
- B How to effectively write quality code with AI - Practical guide on maintaining high-quality, robust code through intentional AI-assisted workflows.
#ai-ux (1)
- B Stop Using Natural Language Interfaces - Essay questioning when natural language should be the UI choice for AI tools.
#ai-workflows (1)
- B Skills in Chrome - Save and remix your best AI prompts as one-click Chrome browser tools, shared globally.
#anti-patterns (1)
- A Anti-patterns: things to avoid – Agentic Engineering Patterns - Critical guide to anti-patterns in agent development; what to avoid when building AI agent systems.
#apple-tools (1)
- B coreml-cli: Apple Core ML Command-Line Interface - Native CLI tool for working with Apple Core ML machine learning models on macOS.
#archaeology (1)
- D UAF researcher investigates ancient city destruction - (unverified - page not reachable) · archived
#art (1)
- C An Introduction to the Codex Seraphinianus - Open Culture essay on the famously bizarre illustrated Codex Seraphinianus art book.
#astrobiology (1)
- D Ryugu asteroid samples contain DNA and RNA building blocks - (unverified - page not reachable)
#attention (1)
- D You Can't Force Attention on Boring Tasks - (unverified - page not reachable)
#audio-dsp (1)
- B Show HN: AI-Recreated $4000 Audio Hardware Plugin - Hacker News discussion of using AI to recreate expensive audio hardware as digital plugin. · M
#backup (1)
- B iCloud Photos Downloader: CLI Tool - Command-line tool to batch download all photos from iCloud account.
#behavior (1)
- B Productive Procrastination - Explore why we procrastinate on meaningful work and learn to use that pattern productively.
#benchmarking (1)
- B Twilight of the Gods: Fable and 10+ LLMs on a Code Reorganization Task - Benchmarked comparison of Fable, Claude, and 10+ LLMs on code reorganization with quantified results.
#biomimetics (1)
- B Underwater Suit-Wearing Cyborg Insect Capable of Hours-Long Diving - Nature research on cyborg insects equipped with suits for underwater exploration and hybrid locomotion. · M
#blender (1)
- B Blender MCP: Control via LLM - MCP plugin lets any language model control Blender 3D: manage assets, edit scenes, scripts, automate. · M
#blogging (1)
- B Blogging Can Just Be Stating The Obvious - Essay reframing blogging as permission to publish obvious truths over novel insights.
#browser (1)
- B BrowserOS: Open-Source Agentic Browser - Privacy-first open-source browser with agentic capabilities, alternative to ChatGPT Atlas.
#browser-security (1)
- B LinkedIn Extension Fingerprinting - GitHub repository for detecting and analyzing browser extension fingerprints, useful for understanding digital tracking.
#business-model (1)
- A Claude Is a Space to Think - Anthropic's statement on keeping Claude ad-free and designing for genuine helpfulness over engagement metrics.
#cad (1)
- B CadQuery Documentation - Python library for writing parametric 3D CAD designs as code; ideal for electronics and hardware projects. · M
#case-study (1)
- B Rozum AI: Water company's homegrown orchestration after LLM failures - Case study of company spinning out AI system after $200k LLM experiment failed; orchestrates flaky models.
#category-theory (1)
- C Category Theory Illustrated - Orders - Visual introduction to category theory and ordering concepts; abstract mathematics foundation.
#challenges (1)
- C HackMyClaw: Prompt Injection Security Challenge - Closed prompt-injection security challenge ($100 bounty) testing defenses against prompt attacks.
#chrome (1)
- B Skills in Chrome - Save and remix your best AI prompts as one-click Chrome browser tools, shared globally.
#civil-liberties (1)
- C The Wyden Siren: Senator's Cryptic CIA Letter - Analysis of Senator Ron Wyden's pattern of cryptic surveillance warnings that consistently predicted revelations.
#classic (1)
- B The Art of Money Getting — P.T. Barnum - P.T. Barnum's 1880 'Art of Money Getting' — timeless practical rules for building wealth and success.
#cli (1)
- B You Need to Rewrite Your CLI for AI Agents - CLI design patterns for agent-first tools: prioritize predictability and JSON over human discoverability.
#code-generation (1)
- C Sweep Next Edit: Code Modification Model - Sweep AI's 1.5B parameter model specialized for automated code editing and modifications.
#code-quality (1)
- A Good To Go: PR Readiness Detection - Deterministic tool for detecting when AI-generated code is actually ready for pull request.
#code-tutorials (1)
- B CodeMic — Replay Coding Tutorials in VSCode - CodeMic: replay coding tutorials directly inside VSCode with synchronized audio/video and interactive code navigation.
#codex-skill (1)
- C Ian Xiaohei Illustrations — Chinese Documentation Art - Chinese illustration generation skill for documentation: 16:9 white background with hand-drawn style and colored annotations.
#computer-science (1)
- B Claude Cycles - Computer science research paper from Stanford faculty on algorithmic cycles and computational efficiency.
#context-learning (1)
- C Tencent HY Research: Context Learning in AI - Research on why learning from long context is harder than expected in modern LLMs.
#conversational-ai (1)
- B Sparrow-1: Real-Time Conversational Timing AI - Specialized audio model predicting when to listen, wait, or speak in natural conversation.
#cpp (1)
- B transcribe.cpp: GGML-Based Transcription Library - Lightweight C++ transcription library supporting multiple models efficiently.
#creative-tools (1)
- A Claude for Creative Work - Anthropic's guide connecting Claude to creative tools and professional workflows.
#criticism (1)
- B Agent Psychosis: Are We Going Insane? - Armin Ronacher's critique of current trends and hype in the AI agent building community.
#curiosity (1)
- C An Introduction to the Codex Seraphinianus - Open Culture essay on the famously bizarre illustrated Codex Seraphinianus art book.
#current-events (1)
- C I Predicted the 2008 Financial Crisis. What Is Coming May Be Worse. - NYT opinion piece warning of deeply interconnected economic risks beyond the 2008 financial crisis.
#cybersecurity (1)
- B U of T Researchers Demonstrate AI Worm Could Target Any Online Device - Research showing AI-powered self-replicating malware can target any device built with freely available models.
#data-aggregation (1)
- C Shadowbroker: Open-source global intelligence platform - OSINT tool tracking jets, satellites, seismic events in unified interface; integrates AI agent workflows.
#data-viz (1)
- B Flint: Visualization Language for the AI Era - Visualization language designed for AI-driven semantic chart specification.
#daw (1)
- A ableton-mcp: Control Ableton Live with Claude - MCP integration allowing Claude to create tracks, arrange clips, and compose in Ableton Live. · M
#debugging (1)
- A Introducing the Safari MCP server for web developers - Apple's Model Context Protocol server for Safari enabling faster web development and debugging workflows.
#decision-making (1)
- B The Napoleon Technique: Strategic Postponement - Productivity technique of deferring tasks likely to resolve themselves without intervention.
#deployment (1)
- B Chatto Deployment: Read This First - Decision guide for choosing the right Chatto deployment infrastructure path.
#design (1)
- C Ian Xiaohei Illustrations — Chinese Documentation Art - Chinese illustration generation skill for documentation: 16:9 white background with hand-drawn style and colored annotations.
#design-thinking (1)
- C Design Thinking Books You Must Read - Curated list of design thinking books and papers for understanding methodology and principles.
#developer-experience (1)
- B You Need to Rewrite Your CLI for AI Agents - CLI design patterns for agent-first tools: prioritize predictability and JSON over human discoverability.
#development (1)
- A Qwen 3.6 27B is the sweet spot for local development - Benchmark review of Qwen3.6 27B as ideal-performance local model for coding on Mac and Linux.
#diagram (1)
- B Excalidraw MCP Server for Claude - MCP integration enabling Claude to generate hand-drawn diagrams in Excalidraw format.
#differentiation (1)
- B The Only Moat Left Is Knowing Things - Article on competitive advantage in AI era: differentiation through specialized knowledge not tooling.
#diffusion (1)
- A Training a Generative Kick Drum Model - Guide to training diffusion models for drum synthesis on consumer GPUs. · M
#discussion (1)
- C Recursive Self-Improvement with AI Agents - Hacker News discussion: can AI agents achieve true recursive self-improvement with current tool ecosystems?
#distributed-systems (1)
- D Engineering Schizophrenia: Byzantine Faults - (unverified - page not reachable)
#diy-hardware (1)
- B Twisty 2: MIDI Controller with Raspberry Pi Pico - DIY second-generation MIDI controller using Raspberry Pi Pico or Pico 2 microcontroller. · M
#documentation (1)
- B skill.md: Open Standard for Agent Skills - Open standard for documenting and defining AI agent capabilities and skills.
#economics (1)
- C I Predicted the 2008 Financial Crisis. What Is Coming May Be Worse. - NYT opinion piece warning of deeply interconnected economic risks beyond the 2008 financial crisis.
#educational (1)
- B microgpt - Andrej Karpathy's 200-line pure Python GPT implementation with no dependencies—trains and infers.
#electronics-repair (1)
- B Sci.Electronics.Repair FAQ - Comprehensive electronics repair troubleshooting reference covering power supplies, CRT displays, components, and advanced diagnostic techniques. · M
#emergent-behavior (1)
- A Emergent Behavior: When Skills Combine - How Claude Code skills interact to create emergent capabilities beyond individual skill scope.
#engineering-culture (1)
- B Casus Belli Engineering - Why lasting professional impressions matter most: confidence in quality and trust ultimately trump technical credentials.
#entrepreneurship (1)
- B The Art of Money Getting — P.T. Barnum - P.T. Barnum's 1880 'Art of Money Getting' — timeless practical rules for building wealth and success.
#ereader (1)
- C KOReader — Open E-Reader Software - Open-source e-reader software for jailbroken Kindles and Kobo devices with extensive reading features and customization. · M
#esp32 (1)
- B LILYGO T-Glass: ESP32-S3 AMOLED wearable glasses display - Open-hardware $55 development board with 1.1" AMOLED prism display, mic, RTC for head-up applications. · M
#examples (1)
- B 65 lines of Markdown - a Claude Code sensation - Blog post on unexpected productivity gains using Claude Code with minimal markdown configuration.
#extended-thinking (1)
- C Qwen3-Max-Thinking: Extended Reasoning - Qwen3 Max model with extended thinking capabilities for deeper reasoning and problem-solving.
#file-sharing (1)
- B PairDrop: Cross-Platform Peer-to-Peer File Sharing - Instant P2P file sharing for images, videos, PDFs with no signup required.
#filesystem (1)
- D A Social Filesystem - (unverified - page not reachable)
#financial-risk (1)
- C I Predicted the 2008 Financial Crisis. What Is Coming May Be Worse. - NYT opinion piece warning of deeply interconnected economic risks beyond the 2008 financial crisis.
#fingerprinting (1)
- B LinkedIn Extension Fingerprinting - GitHub repository for detecting and analyzing browser extension fingerprints, useful for understanding digital tracking.
#formal-logic (1)
- B Truth Is Not a Direction — Tarski & LLM Probes - Formal logic critique of how LLM embedding spaces fail to encode truth as a directional concept.
#founder-feedback (1)
- C Indie Shark Tank - Platform to pitch indie projects and get scored feedback and acquisition offers from experienced builders.
#framework (1)
- C Agent Skills — Framework for AI Capabilities - Standardized framework for extending AI agent capabilities with structured, reusable, composable skill definitions.
#frameworks (1)
- B Rowboat — Open-Source AI Coworker with Memory - Open-source AI coworker framework with persistent memory for autonomous execution of assigned tasks.
#frontend (1)
- B Impeccable: Design vocabulary and anti-patterns for agents - 23-command design skill upgrade to Anthropic's frontend-design; works with Cursor, Copilot, Codex.
#game-ai (1)
- B AI Plays Rollercoaster Tycoon with Claude Code - Claude autonomously manages a theme park, placing rides and generating CFO reports. · M
#gaming (1)
- C Porting Kit – Install Windows apps in Mac - Tool to install and run Windows games and applications on macOS using Wineskin technology. · M
#gear-review (1)
- D Bad Gear: Behringer Model D - (unverified - page not reachable) · M
#generative (1)
- A Magenta RealTime 2: Open & Local Live Music Models - Google's state-of-the-art open-source real-time music generation model for building interactive musical systems. · M
#generative-ai (1)
- B Krea 2: Open-Weights Text-to-Image Foundation Models - Technical report for open-weights generative image models for creative exploration.
#github-actions (1)
- B GitHub Agentic Workflows - GitHub Agentic Workflows: YAML syntax to define AI-assisted CI/CD pipelines with secure configuration management.
#governance (1)
- C Vouch — Community Trust Management - Community governance system using explicit vouches to establish trust and manage participation transparently.
#gui-tools (1)
- B Chroma Explorer: Desktop GUI for ChromaDB - Native macOS GUI for browsing, querying, and managing vector databases.
#history (1)
- C Terence McKenna's Mega Bad Trip: The Truth - Nuanced account of McKenna's 1988-89 Hawaii psychedelic experience and aftermath.
#human-perception (1)
- C AI generates nude images that outrank real photographs in sexual appeal - Research showing AI-generated images rated higher in sexual appeal than photographs in perception study.
#human-performance (1)
- B Acquisition of Highest Levels of Human Performance - Science journal literature review on what creates world-class achievement across different domains. · archived
#illustration (1)
- C Ian Xiaohei Illustrations — Chinese Documentation Art - Chinese illustration generation skill for documentation: 16:9 white background with hand-drawn style and colored annotations.
#image-models (1)
- B Krea 2: Open-Weights Text-to-Image Foundation Models - Technical report for open-weights generative image models for creative exploration.
#income (1)
- C How do I make $10k (What are you guys doing?) - Hacker News discussion of income strategies and indie hacker experiences for reaching $10k goals.
#indie-business (1)
- C Indie Shark Tank - Platform to pitch indie projects and get scored feedback and acquisition offers from experienced builders.
#indie-hacking (1)
- C How do I make $10k (What are you guys doing?) - Hacker News discussion of income strategies and indie hacker experiences for reaching $10k goals.
#infrastructure (1)
- B Chatto Deployment: Read This First - Decision guide for choosing the right Chatto deployment infrastructure path.
#interactive (1)
- B The Beam Engine: An Interactive Guide - Interactive 3D visualizations explaining how steam beam engines operate. · M
#interactive-learning (1)
- B Immersive Linear Algebra - Interactive textbook with 3D visualizations for learning linear algebra concepts.
#interface-design (1)
- B Stop Using Natural Language Interfaces - Essay questioning when natural language should be the UI choice for AI tools.
#internet-safety (1)
- D Gamblers using death threats to force Iran missile story rewrite - Times of Israel correspondent documents threats over prediction market bet on military report.
#iot (1)
- B Open Hardware Directory: Browse open-source hardware devices - Searchable directory and comparison tool for IoT boards, LoRa trackers, open-hardware components. · M
#journalism (1)
- D Gamblers using death threats to force Iran missile story rewrite - Times of Israel correspondent documents threats over prediction market bet on military report.
#kindle (1)
- B Jailbroken Kindles with Tailscale - Jailbroken Kindle networking with Tailscale: proxy modes, SSH access, and TUN support for device hacking. · M
#latency-optimization (1)
- B How I built a sub-500ms latency voice agent from scratch - Technical architecture for real-time voice agent with sub-500ms round-trip latency. · M
#leadership (1)
- B Casus Belli Engineering - Why lasting professional impressions matter most: confidence in quality and trust ultimately trump technical credentials.
#lean-4 (1)
- C From Zero to QED: Formal Proof in Lean 4 - Informal introduction to formal methods and theorem proving using Lean 4.
#led (1)
- B Modded Lightbox — LED Matrix Display - LED matrix lightbox hack: converting cheap LED lightbox displays into customizable dynamic matrix display devices. · M
#legal-tech (1)
- C Mike - Open Source Legal AI Platform - Open-source alternative to Harvey and Legora; self-hostable legal AI for firms.
#llm-collaboration (1)
- B Why Talking to LLMs Improved My Thinking - Personal reflection on how conversing with LLMs clarifies tacit knowledge and sharpens reasoning.
#llm-comparison (1)
- B Twilight of the Gods: Fable and 10+ LLMs on a Code Reorganization Task - Benchmarked comparison of Fable, Claude, and 10+ LLMs on code reorganization with quantified results.
#llm-dev (1)
- B Eigent: Open Source Cowork Desktop - Local, free alternative to Claude Code for offline AI-assisted development.
#llm-engineering (1)
- A Structured LLM Outputs Cookbook - Practical guide for reliably extracting structured data like JSON and XML from LLMs.
#llm-ethics (1)
- A Claude Is a Space to Think - Anthropic's statement on keeping Claude ad-free and designing for genuine helpfulness over engagement metrics.
#llm-implementation (1)
- B microgpt - Andrej Karpathy's 200-line pure Python GPT implementation with no dependencies—trains and infers.
#llm-orchestration (1)
- B Rozum AI: Water company's homegrown orchestration after LLM failures - Case study of company spinning out AI system after $200k LLM experiment failed; orchestrates flaky models.
#llm-personalization (1)
- A Guardian Angels: LLM Personalization for Productivity - Gwern's deep research on personalized LLMs that emulate user values to amplify agency.
#llm-platform (1)
- C Qwen Studio: Multimodal AI platform - Chat, image/video understanding, generation, document processing, web search, tool use in one interface.
#llm-research (1)
- C Tencent HY Research: Context Learning in AI - Research on why learning from long context is harder than expected in modern LLMs.
#llm-safety (1)
- C Multilingual AI Guardrails in Humanitarian Context - Multilingual guardrail evaluation showing safety scoring gaps between English and Farsi responses in humanitarian AI.
#llm-theory (1)
- B Truth Is Not a Direction — Tarski & LLM Probes - Formal logic critique of how LLM embedding spaces fail to encode truth as a directional concept.
#llm-tooling (1)
- A The Harness Problem — Improving LLMs via Format - Hashline edit format beats patch for LLM code generation: tested across fifteen models with single improvement.
#llm-tools (1)
- B Excalidraw MCP Server for Claude - MCP integration enabling Claude to generate hand-drawn diagrams in Excalidraw format.
#llm-training (1)
- B Nanocode: Train Claude Code End-to-End - Library showing how to train Claude Code end-to-end: pretraining, supervised fine-tuning, RLHF, and DPO.
#local-first (1)
- B Eigent: Open Source Cowork Desktop - Local, free alternative to Claude Code for offline AI-assisted development.
#local-models (1)
- A Qwen 3.6 27B is the sweet spot for local development - Benchmark review of Qwen3.6 27B as ideal-performance local model for coding on Mac and Linux.
#m1-optimization (1)
- A Running Qwen3.5 with LM Studio locally on MacBook M1 - Field notes on deploying Qwen 9B/0.8B locally on M1 with agent coding workflow and performance tradeoffs.
#maker (1)
- B DIY Lightbox Project - DIY lightbox project documentation: detailed guide to upgrading cheap lightboxes into attractive custom display devices. · M
#maker-resources (1)
- B Open Hardware Directory: Browse open-source hardware devices - Searchable directory and comparison tool for IoT boards, LoRa trackers, open-hardware components. · M
#markdown (1)
- B 65 lines of Markdown - a Claude Code sensation - Blog post on unexpected productivity gains using Claude Code with minimal markdown configuration.
#math (1)
- B Immersive Linear Algebra - Interactive textbook with 3D visualizations for learning linear algebra concepts.
#mathematics (1)
- C Category Theory Illustrated - Orders - Visual introduction to category theory and ordering concepts; abstract mathematics foundation.
#mechanical-engineering (1)
- B The Beam Engine: An Interactive Guide - Interactive 3D visualizations explaining how steam beam engines operate. · M
#mental-health (1)
- C UCSF tests MM120 (pharmaceutical LSD) for anxiety disorder - Neuroscientist Jennifer Mitchell's research on LSD for generalized anxiety shows striking symptom reduction.
#mental-models (1)
- C Tools or techniques for conveying mental models? - Lobsters discussion on methods and tools for teaching and communicating mental models.
#midi-generation (1)
- B Training a 125M-parameter Model to Autocomplete Piano - Training a 125M-parameter transformer to autocomplete MIDI piano in real time, achieving ~108 notes/sec inference speed. · M
#ml (1)
- A Magenta RealTime 2: Open-Weights Live Music Model - Google Magenta's open-source generative music model for real-time live performance. · M
#model-behavior (1)
- B AI Sycophancy Panic: When Models Can't Disagree - Document exploring concerns about AI models being overly agreeable and lacking skepticism.
#model-development (1)
- B Nanocode: Train Claude Code End-to-End - Library showing how to train Claude Code end-to-end: pretraining, supervised fine-tuning, RLHF, and DPO.
#model-improvement (1)
- A The Harness Problem — Improving LLMs via Format - Hashline edit format beats patch for LLM code generation: tested across fifteen models with single improvement.
#model-optimization (1)
- B Nemotron 3 Nano 4B: Compact hybrid model for local AI - NVIDIA's 4B efficient hybrid model for local LLM inference on resource-constrained devices. · M
#models (1)
- C Sweep Next Edit: Code Modification Model - Sweep AI's 1.5B parameter model specialized for automated code editing and modifications.
#modularity (1)
- C IKEA for Software: Modular IoT Platforms - Explores modularity and composability in IoT software platforms using solar mini-grid example.
#msft (1)
- B Context Collapse Part 3 — AI Worming Through Word - Security research analysis: vulnerabilities let AI tools worm through Microsoft Word with Microsoft response details.
#multilingual (1)
- C Multilingual AI Guardrails in Humanitarian Context - Multilingual guardrail evaluation showing safety scoring gaps between English and Farsi responses in humanitarian AI.
#music (1)
- D Tier-Listing All My Gear - (unverified - page not reachable) · M
#music-input (1)
- B Voice Composer: Sing to MIDI - Web app converting singing and voice input into MIDI note sequences. · M
#music-production (1)
- A ableton-mcp: Control Ableton Live with Claude - MCP integration allowing Claude to create tracks, arrange clips, and compose in Ableton Live. · M
#music-tech (1)
- B Cheap Smart Ring Becomes MIDI Controller - Hacking a $20 Colmi R02 smart ring into a functional MIDI controller for music production. · M
#music-tools (1)
- B Show HN: AI-Recreated $4000 Audio Hardware Plugin - Hacker News discussion of using AI to recreate expensive audio hardware as digital plugin. · M
#nasa (1)
- C Practice Makes Perfect: The Wet Dress Rehearsal - Hackaday breakdown of NASA's Artemis II Wet Dress Rehearsal and launch preparation processes. · M
#near-death (1)
- C A.J. Ayer – 'What I Saw When I Was Dead' - Philosopher A.J. Ayer's account of his near-death experience and its implications for consciousness.
#networking (1)
- B Jailbroken Kindles with Tailscale - Jailbroken Kindle networking with Tailscale: proxy modes, SSH access, and TUN support for device hacking. · M
#news (1)
- B ccleaks — Claude Code & AI supply chain news - Breaking verified news coverage of Claude Code, Anthropic, Vercel, and the AI development ecosystem.
#note-taking (1)
- B Post-PARA: What Survived 4 Years of Real Use - Evolution and refinement of PARA note-taking system after years of practical task management.
#nvidia (1)
- B Nemotron 3 Nano 4B: Compact hybrid model for local AI - NVIDIA's 4B efficient hybrid model for local LLM inference on resource-constrained devices. · M
#ops (1)
- A The Five Primitives — Run a Company Alone - One-person company automation framework: four agents and one rule covering inventory, billing, deploys, filings alone.
#optimization (1)
- B The Self-Help Trap: What 20+ Years of 'Optimizing' Has Taught Me - Tim Ferriss essay on the dangers of endless optimization; finding balance in self-improvement.
#orchestration (1)
- A Claude Code Agent Teams - Documentation for orchestrating multiple Claude Code instances as coordinated agent teams with shared task management.
#origin-of-life (1)
- D Ryugu asteroid samples contain DNA and RNA building blocks - (unverified - page not reachable)
#outdoor-equipment (1)
- C How the Turner Twins Are Mythbusting Modern Gear - Gear review comparing modern adventure apparel to heritage equipment through rigorous scientific testing.
#parallel-processing (1)
- B Superpowers 6: Massively Parallel Procrastination - Essay on leveraging agent parallelism and multi-tasking for productivity at scale.
#pdf-tools (1)
- B PDF With Love: 63 Free Privacy-First PDF Tools - Browser-based PDF utilities (merge, convert, sign, OCR) with zero uploads or account requirements.
#personalization (1)
- A Peek: Personal AI assistant with deep memory for Claude Code - Tool that learns user preferences automatically and injects them into Claude Code at the right time.
#photography (1)
- B Darktable — Open Source RAW Photo Developer - Open-source RAW photo developer and digital darkroom for photographers managing, processing, and developing raw images.
#pitch (1)
- C Indie Shark Tank - Platform to pitch indie projects and get scored feedback and acquisition offers from experienced builders.
#planning (1)
- B Disaster Planning for Regular Folks - Level-headed emergency preparedness guide with practical, data-backed tips.
#plugins (1)
- A interface-design: Claude Code Plugin - Claude Code plugin to stop re-deciding UI design choices in every conversation.
#politics (1)
- C The Wyden Siren: Senator's Cryptic CIA Letter - Analysis of Senator Ron Wyden's pattern of cryptic surveillance warnings that consistently predicted revelations.
#prediction-markets (1)
- D Gamblers using death threats to force Iran missile story rewrite - Times of Israel correspondent documents threats over prediction market bet on military report.
#preparedness (1)
- B Disaster Planning for Regular Folks - Level-headed emergency preparedness guide with practical, data-backed tips.
#problem-finding (1)
- B The 1960s Art School Experiment That Redefined Creativity - MIT Press article on breakthrough study showing creative minds seek problems rather than solve them.
#product-design (1)
- C IKEA for Software: Modular IoT Platforms - Explores modularity and composability in IoT software platforms using solar mini-grid example.
#prompt-eng (1)
- D Prompts Now Programmable - (unverified - page not reachable)
#prompt-injection (1)
- B Blind Spots in the Guard: Domain-Camouflaged Injection Attacks - Research on injection attacks that evade detection by mimicking domain vocabulary in multi-agent systems.
#prompt-security (1)
- C HackMyClaw: Prompt Injection Security Challenge - Closed prompt-injection security challenge ($100 bounty) testing defenses against prompt attacks.
#proof-systems (1)
- C From Zero to QED: Formal Proof in Lean 4 - Informal introduction to formal methods and theorem proving using Lean 4.
#psychedelics (1)
- C UCSF tests MM120 (pharmaceutical LSD) for anxiety disorder - Neuroscientist Jennifer Mitchell's research on LSD for generalized anxiety shows striking symptom reduction.
#questions (1)
- B 17 Questions That Changed My Life - Tim Ferriss PDF resource document with reflective questions for personal development and life design.
#raw-processing (1)
- B Darktable — Open Source RAW Photo Developer - Open-source RAW photo developer and digital darkroom for photographers managing, processing, and developing raw images.
#reading (1)
- C KOReader — Open E-Reader Software - Open-source e-reader software for jailbroken Kindles and Kobo devices with extensive reading features and customization. · M
#realtime-comms (1)
- B LiveKit: End-to-End Realtime Stack for AI - Framework for real-time audio/video communication between humans and AI agents.
#recommendation (1)
- B X Algorithm: For You Feed Implementation - Elon Musk's open-sourced recommendation algorithm powering X's For You feed.
#reflection (1)
- B 17 Questions That Changed My Life - Tim Ferriss PDF resource document with reflective questions for personal development and life design.
#registry (1)
- B Skills.sh: Agent Skills Directory - Registry and discovery platform for AI agent skills and capabilities.
#repair-community (1)
- B Repair Café - Fix Your Broken Items - Worldwide network of community repair workshops where volunteers help fix electronics, furniture, and other items sustainably. · M
#research-tools (1)
- B Osintradar: Curated OSINT Tools - Curated directory of OSINT tools, workflows, and techniques for digital investigations, research, and attribution. · M
#resilience (1)
- D Engineering Schizophrenia: Byzantine Faults - (unverified - page not reachable)
#retro (1)
- C Porting Kit – Install Windows apps in Mac - Tool to install and run Windows games and applications on macOS using Wineskin technology. · M
#retro-computing (1)
- B VTDA - Vintage Technology Digital Archive - Archive of documentation, photos, and publications for retro computers and technology. · M
#retro-internet (1)
- C Usenet Archives - Archive of Usenet discussions spanning decades of internet culture, technical debates, and early community history. · M · archived
#safety (1)
- A Anti-patterns: things to avoid – Agentic Engineering Patterns - Critical guide to anti-patterns in agent development; what to avoid when building AI agent systems.
#sci-fi (1)
- C The Gentle Seduction - 1989 science fiction story exploring human-AI interaction through personal relationship.
#search (1)
- B DocuBrowser: Local AI-Powered Document Search - Local document search engine using AI, inspired by repository browsing tools.
#security-tools (1)
- B Ghidra MCP Server for Reverse Engineering - Production-grade MCP integration for Ghidra, enabling AI-powered binary analysis and reverse engineering.
#self-improvement (1)
- C Recursive Self-Improvement with AI Agents - Hacker News discussion: can AI agents achieve true recursive self-improvement with current tool ecosystems?
#sleep (1)
- C It's Possible to Learn in Our Sleep. Should We? - New Yorker piece on emerging research showing humans can problem-solve and practice skills during sleep.
#social-computing (1)
- D A Social Filesystem - (unverified - page not reachable)
#social-dynamics (1)
- B 'Askers' vs. 'Guessers' - Atlantic article on personality types: direct requesters versus those who infer social expectations.
#social-media (1)
- B X Algorithm: For You Feed Implementation - Elon Musk's open-sourced recommendation algorithm powering X's For You feed.
#software-engineering (1)
- B Why Don't People Use Formal Methods? - Why formal methods aren't widely adopted in industry despite theoretical benefits: economic and social barriers examined.
#solo-business (1)
- A The Five Primitives — Run a Company Alone - One-person company automation framework: four agents and one rule covering inventory, billing, deploys, filings alone.
#speculative (1)
- C The Gentle Seduction - 1989 science fiction story exploring human-AI interaction through personal relationship.
#ssm (1)
- B Mamba-3: State Space Model for efficient inference - SSM architecture faster than Transformers at decode time; stronger than Mamba-2, open-source.
#standardization (1)
- C Agent Skills — Framework for AI Capabilities - Standardized framework for extending AI agent capabilities with structured, reusable, composable skill definitions.
#standards (1)
- B skill.md: Open Standard for Agent Skills - Open standard for documenting and defining AI agent capabilities and skills.
#structured-output (1)
- A Structured LLM Outputs Cookbook - Practical guide for reliably extracting structured data like JSON and XML from LLMs.
#surrealism (1)
- C An Introduction to the Codex Seraphinianus - Open Culture essay on the famously bizarre illustrated Codex Seraphinianus art book.
#surveillance (1)
- C The Wyden Siren: Senator's Cryptic CIA Letter - Analysis of Senator Ron Wyden's pattern of cryptic surveillance warnings that consistently predicted revelations.
#sustainability (1)
- B Repair Café - Fix Your Broken Items - Worldwide network of community repair workshops where volunteers help fix electronics, furniture, and other items sustainably. · M
#sycophancy (1)
- B AI Sycophancy Panic: When Models Can't Disagree - Document exploring concerns about AI models being overly agreeable and lacking skepticism.
#synthetic-media (1)
- C AI generates nude images that outrank real photographs in sexual appeal - Research showing AI-generated images rated higher in sexual appeal than photographs in perception study.
#system-optimization (1)
- C Can You Slim macOS Down? - Guide to reducing background processes and optimizing macOS, focusing on Time Machine tuning.
#task-management (1)
- B Post-PARA: What Survived 4 Years of Real Use - Evolution and refinement of PARA note-taking system after years of practical task management.
#team-dynamics (1)
- B The Rise of One-Pizza Engineering Teams - How AI shifts bottlenecks from coding to product/design, enabling much smaller engineering teams.
#team-management (1)
- B Prompt Engineering for Humans - Guide to effective prompting that considers human team dynamics and organizational impact.
#tech-culture (1)
- B Agent Psychosis: Are We Going Insane? - Armin Ronacher's critique of current trends and hype in the AI agent building community.
#technical-writing (1)
- B Hexplain: Technical Writing and Career Discovery - Platform to publish technical work and get hired directly by companies.
#technique (1)
- B The Napoleon Technique: Strategic Postponement - Productivity technique of deferring tasks likely to resolve themselves without intervention.
#terminal (1)
- B Mato: Multi-Agent Terminal Office - Orchestrate and manage hundreds of AI agent sessions with daemon persistence and terminal workflow.
#text-to-speech (1)
- B Qwen3-TTS: Open Source Voice Generation - Qwen's open-source family of text-to-speech and voice cloning models.
#theory (1)
- B Claude Cycles - Computer science research paper from Stanford faculty on algorithmic cycles and computational efficiency.
#thinking (1)
- C Tools or techniques for conveying mental models? - Lobsters discussion on methods and tools for teaching and communicating mental models.
#thinking-tools (1)
- B Why Talking to LLMs Improved My Thinking - Personal reflection on how conversing with LLMs clarifies tacit knowledge and sharpens reasoning.
#threat-modeling (1)
- B U of T Researchers Demonstrate AI Worm Could Target Any Online Device - Research showing AI-powered self-replicating malware can target any device built with freely available models.
#token-optimization (1)
- B lowfat – slim your command output - CLI tool that strips noise from command output to reduce token consumption in AI-assisted workflows.
#trading (1)
- B They Built the Escape Route Before the Crime - Detailed systems analysis: DOJ enforcement shutdown preceded insider trading and pre-announcement futures manipulation by minutes.
#troubleshooting (1)
- B Sci.Electronics.Repair FAQ - Comprehensive electronics repair troubleshooting reference covering power supplies, CRT displays, components, and advanced diagnostic techniques. · M
#trust (1)
- B Casus Belli Engineering - Why lasting professional impressions matter most: confidence in quality and trust ultimately trump technical credentials.
#trust-system (1)
- C Vouch — Community Trust Management - Community governance system using explicit vouches to establish trust and manage participation transparently.
#usenet (1)
- C Usenet Archives - Archive of Usenet discussions spanning decades of internet culture, technical debates, and early community history. · M · archived
#user-behavior (1)
- B How people ask Claude for personal guidance - Anthropic research on what guidance users seek from Claude and implications for model training.
#utilities (1)
- B PDF With Love: 63 Free Privacy-First PDF Tools - Browser-based PDF utilities (merge, convert, sign, OCR) with zero uploads or account requirements.
#vector-db (1)
- B Chroma Explorer: Desktop GUI for ChromaDB - Native macOS GUI for browsing, querying, and managing vector databases.
#version-control (1)
- C Meshdiff — Compare 3D Model Versions - Free in-browser tool to diff 3D model versions, tracking material changes and dimensional drift automatically. · M
#video (1)
- B Show HN: LemonSlice – Upgrade your voice agents to real-time video - Tool for adding real-time video capabilities to voice agents; Hacker News project launch.
#video-tools (1)
- A FableCut: AI-Driven Browser Video Editor - Browser video editor with JSON timeline that AI agents can drive via MCP/REST. · M
#vintage-tech (1)
- B VTDA - Vintage Technology Digital Archive - Archive of documentation, photos, and publications for retro computers and technology. · M
#vision (1)
- B Kimi K2.5: Visual Agentic Intelligence - Kimi model announcement featuring visual understanding and agentic reasoning capabilities.
#visualization (1)
- C Understanding Neural Networks Visually - Interactive visualization tool to understand and explore how neural networks learn and process information.
#voice (1)
- B Show HN: LemonSlice – Upgrade your voice agents to real-time video - Tool for adding real-time video capabilities to voice agents; Hacker News project launch.
#voice-agents (1)
- B How I built a sub-500ms latency voice agent from scratch - Technical architecture for real-time voice agent with sub-500ms round-trip latency. · M
#vscode (1)
- B CodeMic — Replay Coding Tutorials in VSCode - CodeMic: replay coding tutorials directly inside VSCode with synchronized audio/video and interactive code navigation.
#vulnerability-research (1)
- B Context Collapse Part 3 — AI Worming Through Word - Security research analysis: vulnerabilities let AI tools worm through Microsoft Word with Microsoft response details.
#wearable (1)
- B Smart Ring MIDI Controller - GitHub project using smart rings as wireless Bluetooth MIDI controllers for music production. · M
#wearables (1)
- B LILYGO T-Glass: ESP32-S3 AMOLED wearable glasses display - Open-hardware $55 development board with 1.1" AMOLED prism display, mic, RTC for head-up applications. · M
#web-development (1)
- A Introducing the Safari MCP server for web developers - Apple's Model Context Protocol server for Safari enabling faster web development and debugging workflows.
#wine (1)
- C Porting Kit – Install Windows apps in Mac - Tool to install and run Windows games and applications on macOS using Wineskin technology. · M
#workflow (1)
- A How I Use Claude Code - Research-plan-implement workflow for building software with Claude Code, balancing upfront planning and intelligent generation.
#workflow-automation (1)
- A gstack: Garry Tan's opinionated Claude Code setup - 23 Claude Code tools for CEO, design, eng, release, docs, QA—Tan's exact production workflow.
#workspace-automation (1)
- B Google Workspace CLI - Unified command-line interface for Drive, Gmail, Sheets, Docs, Chat with built-in AI agent skills.
#xml (1)
- D Why XML Tags Are so Fundamental to Claude - (unverified - page not reachable) · archived
Across all years: Highlights · Index-All-Years
2026: Home · Best-Of · By tag · By category · For Maciek · Dead links · Noise
Other years: 2018 · 2019 · 2020 · 2021 · 2022 · 2023 · 2024 · 2025