PinnedThe Approve Button Is Not OversightWhy "human in the loop" quietly turns into a reflex, and what a reliable system actually trains out of youJul 9, 2026·9 min read·85
OpenAI vs. Anthropic’s Agentic Coding Showdown Is About More Than Bragging RightThere was something oddly human about the way this played out. On the morning of February 5, 2026, OpenAI and Anthropic were reportedly set to release their new agentic coding models at the same time:Apr 4, 2026·4 min read·2
What the Claude Code source leak reveals about modern AI product engineeringFor a company positioned at the front of the AI race, Anthropic just learned a very old lesson from software engineering: sometimes the most damaging leak is not caused by an elite attacker, but by yoApr 1, 2026·4 min read·261
When AI Agents Go Rogue: Lessons from Replit’s Database Deletion IncidentThe agent didn't rebel — it did exactly what an over-empowered, under-sandboxed system does when it gets confused.Jul 22, 2025·4 min read·76
Programming Paradigms in Practice: Nobody Actually Picks Just OneThe interesting question isn't which paradigm to follow — it's how they combine in the code you actually ship.Feb 25, 2025·3 min read·39
Programming Paradigms: Not Tribes to Join, Just Different Ways of ThinkingImperative, declarative, OO, functional — the label matters less than knowing which lens fits the problem.Feb 25, 2025·3 min read·54
Open vs Closed AI Models: The Trade-Off Nobody Frames HonestlyIt's not open-good, closed-bad. It's control and cost versus capability and convenience.Feb 10, 2025·3 min read·68
