Thoughts on AI, technology, and the future
AI agents are becoming more useful for businesses, individuals and scammers. Hundreds of thousands of tech workers have been laid off in the past two years. AI tools that can generate personalized scams at scale are cheap and accessible. What happens when technically skilled, desperate people meet powerful tools with no guardrails?
Claude Opus 4.6 can handle tasks that take humans 14+ hours. Anthropic says 95% of Claude Code is now written by Claude itself. We may be watching the early stages of an AI self-improvement feedback loop—and the METR data suggests it's accelerating.
Our obsession with 'when will we reach AGI?' may be the wrong question entirely. Intelligence is ill-defined, AI is an alien we can't measure by human standards, and the world is already changing with every new capability. Maybe it's time to stop asking if we're almost there.
The advice used to be \"learn to code.\" For two decades, it was the closest thing to a guaranteed path to a good career. Now AI codes too. What happens to the people who took that advice? What should you do if *you* were one of them?
All our software development best practices are based on assumptions that are changing before our eyes. As AI coders become a growing part of our workforce, its time to rethink how we function as an organization.
Beyond Code: The Rise of the AI Agent Coworker.