A technology leader's framework for AI cost, risk, and right-sizing. Why the price Is not the price For a growing share of enterprise workloads, a frontier language model is the most expensive and highest-risk route to an inferior result. The macro data already shows it. An NBER study published in February 2026 found 89 percent of firms reporting no impact of AI on productivity over the past three years, and more than 90 percent reporting none on employment. [1] A METR random

Why engineers who can name the business outcome make better calls, and how technology leaders build that knowledge An engineer who knows what the company is trying to do, and who the work reaches, makes better decisions without being told. Strategy decks don’t build that knowledge. Stories do, along with a few habits that keep the business in front of the team. It matters more now that AI writes a growing share of the code, because judgment about what matters is the part of t

What a CEO gets from the technology organization once AI decisions move to the top AI decisions have moved to the CEO’s desk, and in most companies the technology organization hasn’t moved with them. The result is one of three outcomes: fast and ungoverned, governed and slow, or fast and governed. Only fast and governed holds up at scale, and it depends on one condition: technology leaders who know what the company is trying to do well enough to shape AI decisions before the