Two documents landed within twenty-four hours of each other this week, and together they say more than either intended.

On Monday, more than 200 economists and technologists — over a dozen Nobel laureates among them, including Joseph Stiglitz, Daron Acemoglu, and Simon Johnson — published an open letter titled "We Must Act Now." Organized by the Stanford Digital Economy Lab, it warns that AI may become radically more powerful within the decade and could compress an economic transformation larger than the Industrial Revolution into a fraction of its timeframe. Large-scale job displacement on one side, major gains in living standards on the other — with the outcome depending on whether we build, in the letter's words, "the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society."

On Tuesday morning, Demis Hassabis published a framework calling for a US Frontier AI Standards Body modeled on FINRA: independent pre-release testing of frontier models for cyber, biological, and deception risks — voluntary at first, mandatory once the regime proves itself. His timeline for AGI: "probably only a few short years away." His metaphor of choice is not a better tool. It is fire, electricity — "we've essentially found a way to make sand think."

Note who is speaking. Not critics of the technology — the people building it. Jack Clark of Anthropic signed the letter. So did Google's Jeff Dean and OpenAI's CFO. Hassabis runs the lab whose models the proposed body would test. When the builders start designing their own constraints, that is not marketing. It is a statement that the current institutional setup cannot absorb what they know is coming.

Which brings me to the part of this week that worries me more than either document: society, at least, has started the right conversation. Most corporations have not.

The lens problem

After almost twenty-five years in professional services, advising organizations from family businesses to global groups, I see one pattern with almost no exceptions. Everyone looks at AI through their particular lens. The HR leader with two decades in HR sees AI-for-HR — a screening tool here, a chatbot there. The compliance officer augments compliance processes. The CFO pilots something in accounts payable. Each initiative is locally rational, professionally executed, and defensible in any budget review.

Collectively, they amount to answering a structural rupture with departmental patchwork.

The problem is not the pilots. It is what the pilots reveal: in most organizations, nobody's actual job is to ask what the company looks like when the assumptions underneath every function move at once. What happens to a leverage-based professional services model when junior work is automated. What happens to a middle-management layer built to relay information that machines now relay better. What happens to decision rights when analysis stops being scarce. These are not functional questions. They cut across the org chart — which is precisely why the org chart cannot answer them. The lens is the structure itself.

Where the advice should come from — and doesn't

The standard answer is that this is what business schools and strategy consultants are for. My assessment, after two and a half decades in and around that ecosystem: much of it has lost touch with the corporate reality it claims to shape. What the leading schools teach is calibrated to organizations with resources, talent density, and board sophistication that the median corporation simply does not have. It is Usain Bolt's coach lecturing on marginal performance gains to someone who runs an easy 2K twice a week — technically impressive, and aimed at a reality the audience doesn't inhabit.

One layer below, the technology consultancies sell implementation: tools, licenses, integration. That answers "how" for organizations that have never resolved "what" and "why."

So the macro layer produces warnings about the largest economic transformation since industrialization, and the micro layer buys a copilot license. The middle — the strategic redesign of actual companies — is nobody's job.

The obvious objection

The case against centralizing AI strategy is real: top-down transformation programs have a miserable track record, and functional experimentation at least ships something. Partly right. Experimentation beats paralysis, and I will take a pragmatic HR pilot over a 200-page strategy deck any day.

But the objection misses what this week's warnings are actually about. Functional pilots optimize the existing operating model. The letter and Hassabis's framework both describe a scenario in which the operating model itself becomes the variable — labor economics, coordination costs, the value of experience, the scarcity of judgment. You cannot pilot your way to a redesigned company. At some point, a decision-maker with real authority has to hold the whole picture and choose.

And if you take Hassabis's timeline even half seriously — half is my personal discount rate — waiting is also a choice. Just an unexamined one.

Direction

The economists asked for institutions equal to the transformation at the level of society. The same request applies one level down, and it lands on the CEO's desk, not the CIO's.

Three questions worth answering before the next board meeting. Who in your organization owns the holistic AI question — with decision rights, not a task force? Which of your core operating assumptions survive a world where analysis is abundant and execution is cheap? And if the honest answer to the first question is "nobody" — that is not a gap in your org chart. That is your strategy.

Society got its open letter this week. Most corporations still owe themselves one.