Posts
All the articles I've posted.
Software 3.0 - Programming in the Age of AI
Updated: at 01:39 PMExploring Andrej Karpathy's vision of Software 3.0 where prompts become programs, LLMs act as utilities, and we design for partial autonomy in an AI-first world.
From Vibe Coding to Agentic Engineering- The Mindset Shift for AI Development
Updated: at 04:09 PMHow I moved from quick-and-dirty prototyping to building robust, intentional AI agent systems — and why both mindsets matter.
Building Multi-Agent Research Systems - Lessons from Anthropic's Architecture
Updated: at 12:59 AMHow multi-agent research systems work and what it takes to build one: the orchestrator-worker pattern, prompting subagents, evaluation, token costs, and production pitfalls, drawing on Anthropic's published design.
Claude Prompt Engineering Best Practices - What Works with Claude 4
Updated: at 10:59 AMPractical Claude prompt engineering best practices, written for Claude 4: be explicit, give context, show examples, control the output format, and let Claude verify its own work. With code examples and a migration guide from earlier Claude models.
Emerging User Interaction Patterns in Agentic Applications
Updated: at 04:21 PMExploring the fundamental shift in how users interact with intelligent software agents, from command-based interfaces to collaborative, intent-driven experiences that are reshaping human-computer interaction.
Generating an llms.txt file for your Astro Website
Updated: at 11:30 AMA practical guide to creating an llms.txt file for Astro websites to make your content accessible to language models without complex crawling solutions - simple, efficient, and perfect for static sites.
Prompt Engineering- The Critical Skill for Building Reliable Agentic Workflows
Updated: at 11:29 AMAn exploration of why prompt engineering is essential for creating reliable AI agents - covering deterministic behavior in probabilistic systems, tool use precision, chain-of-thought integrity, and practical techniques for measuring effectiveness.
Agent != Conversational
Updated: at 10:59 AMExploring why agentic AI systems shouldn't be limited to conversational interfaces, and how focusing on alternative interaction models can deliver the 300%+ efficiency gains that make agents truly transformative.
Interesting Finds of the Week - Claude Code, OWL, Agentic Systems, and MCP Servers
Updated: at 10:59 AMA summary of interesting AI-related projects and articles this week, including Claude Code analysis, OWL framework, agentic systems architecture, and awesome MCP servers.
What is MCP (Model Context Protocol)? The Foundation for Agentic AI
Updated: at 10:00 AMHow Model Context Protocol (MCP) provides the critical foundation for building capable AI agents. Explore why structured context exchange is the key to unlocking truly agentic systems, how MCP enables tool use, multi-agent collaboration, and the rise of complex agent ecosystems.