AIWA "The One Thing" #23: Enterprise AI Has a J-Curve. The Hangover Proves It. Your CFO Is The Variable.

Show me your CFO's tech quotient and I'll show you where you are on the J-curve.

The token-maxxing hangover is real. So is Jevons Paradox.

Open models get cheaper, total consumption goes up not down and routing discipline becomes more critical not less.

Every CFO being surprised by a hangover bill tied to indiscriminate token use has a very important decision to make. One that will determine where they are and will be on the J-curve.

Every general purpose technology has lived through this dip. Electricity. The PC. The Internet.

Enterprise AI is no different.

The enterprises pulling out of the dip fastest are led by CFOs who know exactly which problems deserve frontier tokens and which ones don't. Those CFOs aren't reviewing a speculative R&D expense anymore.

They are managing a new class of operational spend. And their tech quotient is now the ceiling.

The Dip Was Always Coming. Play the Long Game.

General purpose technologies fundamentally change the economy but they take years to have their full impact.

Over a year ago, Stanford University economist Erik Brynjolfsson unpacked exactly this on Harvard Business School Managing the Future of Work podcast.

Describing his research that has documented the same pattern across every major GPT.

It’s always a J.

*"Erik Brynjolfsson on how AI is rewriting the rules of the economy" via HBS Future of Work Podcast

The first factories to install electric motors put them in the exact same place as the steam engine. Nothing changed. It took 30 years and a new generation of managers redesigning from scratch to unlock a "100 percent increase in productivity."

*"The Productivity J-Curve: How Intangibles Complement General Purpose Technologies" by Erik Brynjolfsson, Daniel Rock and Chad Syverson

His read in March 2025: we were close to the trough of the AI J-curve. The investments enterprises were making in workflow redesign, talent and business process weren't showing up in productivity numbers yet.

They weren't supposed to. That's the trough. The upswing follows.

Brynjolfsson was clear on one thing: the AI J-curve will be significantly shorter than electricity. The organizations that understood that over a year ago and kept investing are already further up the curve than those that panicked at the invoice.

*"AIWA TOT #09: The Gap is the Game."

Your CFO's Tech Quotient Is Now Your AI Ceiling.

The invoice isn't the problem. It's the tuition.

The era of token-maxxing is over. Indiscriminate frontier token use across every workflow regardless of complexity is what created the hangover. The antidote isn't spending less. It's spending smarter.

The CFO who reacts to a surprise Anthropic invoice without context cuts the wrong thing. The CFO who understands Jevons Paradox knows cheaper open models drive total consumption up not down. That's not a problem. It's a forcing function.

The decision now sitting on the CFO's desk: build or buy the orchestration layer. The harness. The router. The intelligent system that determines which workflows earn frontier tokens and which run on open models.

Frontier for the 10-20% requiring planning intelligence and complex decision-making. Open models for the other 80-90%.

That isn't a technology decision. It's a resource allocation decision. And it is the single most powerful lever a CFO can pull right now to accelerate an enterprise's move through the J-curve dip and into the upswing.

This isn't experimental anymore. It's operational. The CFO in that room now sets the ceiling.

*"AIWA TOT #17: Stop Asking 'What Is Our AI Strategy?' Start Asking What Ray Dalio Did In 2016."

Right Token. Right Job. Right Leader. Right Team.

Routing discipline is necessary. Not sufficient.

The enterprises coming out of the dip fastest have wired the human layer to match. Less activity metrics. More outcome metrics. Revenue per employee. Cost per outcome. Cycle time. Not AI sessions. Not tokens consumed.

The load-bearing polymath owns the business outcome end to end supported by the orchestration and routing harness beneath them. Leader of a cross-functional team that knows the difference between a chainsaw and a hedge clipper.

When and how to use either.

Less "what did you present this week?"

More "what did you ship and what outcomes did it generate?"

*"AIWA TOT #16: The Unit of Production Just Changed. Automate. Augment. Rewire. (From My RDSW Talk)"

That's not a staffing decision. That's a bet on the upswing.

Humans + Machines. Never Humans vs. Machines.