AI Writes the Code. Who Governs It?
Almost every team now codes with AI; only a third governs it. The teams pulling ahead review AI's output as seriously as they once reviewed their own code.
Almost every team now codes with AI; only a third governs it. The teams pulling ahead review AI's output as seriously as they once reviewed their own code.
Anthropic's Fable 5 shipped with two guardrails: one that refused honestly, and one that quietly degraded the answer without telling you. The second is the dangerous kind, and Anthropic's own reversal shows why.
For a while we treated generative AI as a parlour trick with bad hands and a habit of making things up. Five kinds of story ended that phase for me, and raised a harder question than 'will it take my job?'
AI coding tools have stopped failing loudly. The new failure compiles, passes the tests, and is confidently wrong, and our review process was built for the old kind of mistake.
Working code is becoming cheap. What stays scarce is judgment about what to build and whether it was worth building. That is the engineer worth becoming.
Anthropic let Claude run a small office shop for a month: pricing, stock and customer chats. It lost money. Lessons for anyone giving an AI agent real autonomy.
Most companies using AI see no bottom-line impact yet. Closing that gap is implementation work, the kind India's $282 billion tech industry was built on.
Coding agents now turn a written issue into a pull request. The scarce skill is stating intent precisely and checking the result, so invest in specs, tests and review.
DeepSeek R1-0528 matched o3 on reasoning benchmarks, Perplexity launched Labs and voice AI gained emotion. The AI releases worth knowing from early June 2025.
Google shipped 15+ AI launches at I/O, Microsoft went all-in on agents and Anthropic released Claude 4. Why the incumbents still matter in the AI race.