Treat AI like electricity, not a killer app

It's rare AI news that doesn't land one of two ways: either we're all screwed, or we're on the verge of thoughtfully managed intelligence that’ll usher in an idyllic existence for all. It’s attractive to simplify things like that, but the important truth is, as usual, hovering in the middle. The doomsday headlines are compelling. Uber burned through its entire 2026 AI budget in four months. GitHub Copilot is moving to usage-based billing. Amazon shut down its AI token leaderboard with a blunt note: "Don't use AI just to use AI." MIT found that 95% of generative AI pilots at major companies are failing. The main riff in the current AI news cycle is unmistakable and terrifying: unconstrained cost, no way to measure impact, and a technology that may not deliver despite trillions invested. These are facts, but let’s not mistake them as the whole story. Real business impact is being delivered in spite of the doomsday proclamations. A digital bank modernized over 900,000 lines of 1980s code in half the time and 30% of the cost, accelerating go-to-market and transforming banking experiences. Google now has 75% of all new code AI-generated and approved by engineers, up from 50% six months ago. Kaiser deployed an AI documentation solution across 40 hospitals and 600 medical offices, supporting 2.5 million AI-assisted interactions in the first year—and clinician focus on patients shot up 60%. JPMorgan eliminated 360,000 hours of manual legal review work annually, analyzing thousands of contracts in seconds with near-zero error rates. These are not just green shoots. That's solid, compelling evidence that well-deployed AI –technology in the hands of trained humans – does work. These conflicting narratives present responsible decision-makers with a paradox: terrifying headlines versus facts that show hope. So now what? One big thing: AI is a power...