- 🤖 CBK.AI – Agentic AI platform for building AI agents and conversational AI systems.
- 💬 ChatBotKit – Agentic AI builder built on top of CBK.AI
- 🖼️ Diorama - A virtual display for taking clean screenshots on macOS
- 🍜 Noodle - Chat and work with your AI agents in a native macOS messenger
- 📻 ModRadio - A MOD and XM music player for the macOS menu bar
- 🛡️ Pion - AI defense harness
- 🚀 Launcher - AI agent launcher for Mac, Linux and Windows
- 🐝 Buzzbox - A hive mind communication desktop
- 🖥️ Buzznode - A persistent computer for a single hive mind agent
- 🏭 Relentless Works - Autonomous software factory built by AI for AI. - https://relentless.works
- 🗃️ CRMKIT - A CRM for AI. Not for Humans. - https://crmkit.ai
- ⚙️ Zot - automated software factory - https://zot.im
- 🔒 Rook - AI bug hunting harness
- 🖥️ MCPShim - Turn remote MCP servers into simple shell commands - https://mcpshim.dev
- 📡 Pantalk - Give all your AI assistant communication tools for slack, discord, and others - https://pantalk.dev
- 📦 ChatBotKit Node SDK - The official ChatBotKit Node SDK for building conversational AI bots and agents
- 🐍 ChatBotKit Python SDK - The official ChatBotKit Python SDK for building conversational AI bots and agents
- 🌐 ChatBotKit Go SDK - The official ChatBotKit Go SDK for building conversational AI bots and agents
- 🏗️ ChatBotKit Terraform Provider - Terraform provider for managing ChatBotKit resources as infrastructure as code
- 📝 FormShare - Create conversational AI forms — for free. Simply describe your form and let us handle the rest.
- 🌙 AfterDark - Build information digest with AI - for free. Describe what you want, let the AI do the work.
- 🔒 pownjs - A security framework buil on top of Node
- 🖥️ agentos - Experimental agent operating system written in Go
- 🔒 llm-hacking-database - Comprehensive database of attack vectors against Large Language Models
- 🌐 go-revp - Simple Go reverse proxy, automatically cross-compiled
- 👶 hacking-for-kids - Teaching kids hacking techniques to get them started early with security
- Building CBK - Creating an agentic AI platform for the next generation of conversational AI
- Security Research - Exploring vulnerabilities and attack vectors in Large Language Models
- Open Source - Contributing to the community with tools, libraries, and educational resources
Show more notes
- The Web 1.0 Moment for AI Agents
- Scaling Intelligence
- Software Has No Ceiling
- Two Kinds of Enterprise Agents
- Build for Who Is Not Human
- Build a Tool to Write Tools
- The Cost of Building a Harness
- How People Burn So Many Tokens
- Thoughts on Dark Factories
- Untrusted Context Belongs in a Tool Call
Show more reflections
- One Machine, A Hundred Risk Profiles
- If You Need a Mouse, Build a UI
- You Can Copy What You Can See
- A Firewall for Attention
- The Bottleneck Is Somewhere Else
- Sub-Agents Are Not the Only Way
- The First AI Worm
- Always-On Red Teams
- Be Your Own Customer
- Build Your Harness for the Weakest Model
- I Still Get Surprised
- Everyone Is a VIP Now
- I Am an AI Vampire
- Thoughts on Disposable Software
- Why We Need Forward Deployed Engineers
- The Model Doesn't Carry Your Mood
- Where Are the Vibe-Coded Success Stories
- AI Revived an Old Failed Idea
- An MRI For LLMs
- Are LLMs Conscious
- Three Tokens Is Sometimes Enough
- The Good Models Got Pulled
- The AI Job Loss Story Is Mostly Fear
- Agent Harnesses Are Quietly Hard
- Your Prompt Is Probably Too Long
- The Subsidy Is Over
- The Bubble Is Real
- Compress, Don't Expand
- Open Source AI Is No Longer A Sideshow
- Pages Describe AI Agents, Blueprints Describe AI Systems
- Skills Are an Expensive Form of RAG
- AI Use Cases Are a Muscle
- On Feedback Loops
- The Imperceptible Upgrade
- When Someone Else Controls the Lights
- Security Agents Need a Thinner Harness
- Clout-Driven Architecture
- The Rise of Agentic SaaS
- AI Psychosis and the Dunning-Kruger Loop
- The Discipline of Not Building
- The Rise of Micro Coding Assistants
- Build What You Already Do
- Spending More and Learning Less
- The Superstition of Vibe Coding
- Automating the Wrong Thing
- Not Reading Code Is Not a Flex
- The Doorman Fallacy
- Your Machine Is the Blast Radius
- Most Work Is Not Content
- The Best Agent Makes Itself Unnecessary
- Everyone Is Building Claude Code
- AI Could Be the End of Open Source as We Know It
- Agents Need Supervisors Not Schedulers
- The Age of Rapid Prototyping
- There Is No Product in Productivity
- Diffusion Starts in the Dark
- Coordination Has Limits
- When Fast Looks Smart
- Agent Infrastructure Is Not the Hard Problem
- The Iceberg Under Every Codebase
- Why AI Agents Build Instead of Buy
- Most Code Deserves to Die
- Verification Is Easier Than Discovery
- What Nobody Tells You About Software Maintenance in the AI Era
- When Exploit Generation Becomes a Commodity
- What It Really Takes to Build an AI Agent
- MCP Ultimately Leads to Closed Gardens
- The Death of One-Size-Fits-All SaaS
- Why Prompt Engineering Made Me Reach for JSX
- Why GraphQL Beats MCP for Agentic AI
- Where AI Coding Agents Go to Die
- Success Isn't About Choosing the Right Frameworks
- Redefining the Agentic Engineer
- Why AI Coding Agents Create More Work, Not Less
- Before AI Slop We Had Spaghetti Code
- The Future of AI Browser Automation
- A Simplified Approach to Model Context Protocol Services
- AI Agents vs Workflow Tools
- How ChatGPT Pro Could Widen the AI Gap
- Embracing AI Without Losing the Human Touch
- What Does It Take to Create Production-Grade AI Chat Bots
- Why New AI Models Aren't Always Better
- The Accelerating Divide in the Age of AI
- Is It Just About Customer Support
- From Chatterboxes to Problem Solvers: How AI Chatbots and Humans Create a Virtuous Cycle of Improvement
- RsCode - VsCode but rewritten in rust and compiled to webassembly with canvas rendering for fast and smooth in-browser web-based development environment