Commentary on AI, coding with agents, and the future of software — from the workshop floor.
Anthropic just baked a new pattern into Claude Code — the Advisor strategy. Sonnet runs your task end to end and only calls Opus when it hits a wall. Better performance, lower cost, and one line in your API call changes everything about how we build AI agents.
Read full article →A famous Tamil poet's wisdom about avoiding negative words in songs applies perfectly to AI prompts. Learn why framing instructions positively leads to better results than using restrictions like "don't" and "avoid."
Read full article →Enterprise AI crossed a decisive threshold on December 3, 2025. AWS unveiled its comprehensive agentic AI strategy with Frontier Agents, Nova 2 models, and AgentCore platform—marking a fundamental shift from AI as assistant to AI as autonomous collaborator.
Read full article →A training session revelation about AI coding assistants. When asked if an AI could read Excel files, the answer surprised me—it couldn't, but it wrote a script to do it anyway. A glimpse into how AI agents are becoming meta-developers, reshaping the future of software development.
Read full article →How AI's evolution in software development mirrors the T-1000 from Terminator 2. Exploring three phases—Mimic, Morph, and Infiltration—that show how AI is reshaping code review and the developer's role from code writer to system architect.
Read full article →