The Last Scarcity

The capacity to want something over time — you imagine having it. You plan for it, you make sacrifices toward it. The object accumulates meaning from this process. When you finally arrive, you don't just get the thing. You get the thing plus everything you invested in the wanting. Those two things can't be separated.

This piece by @fede_intern has been circulating in AI circles this week, and it deserves the attention. The core argument: what atrophies in a world of frictionless AI isn't access to goods or even employment — it's desire itself. The psychological muscle that lets you stay oriented toward something difficult and distant.

It's a good essay. The addiction analogy is the strongest section — the observation that frictionless environments don't deliver pleasure so much as they train you to treat discomfort as a prompt for relief. That maps onto real neuroscience, and the piece is smart enough not to overstate the parallel.

Where it wobbles is the quiet assumption that AGI closes the skill gap to zero — that you'll no longer need to bring attention or judgement to AI interactions. Anyone who has spent serious time with current tools knows the opposite: the gap between a thoughtful prompt and a lazy one is vast, and arguably widening as models become more capable. More power demands more discernment about how to use it, not less.

There's also a class blind spot worth naming. The examples of "good friction" — handmade objects, wine vintages, difficult shared hobbies — are drawn from a life where most unwanted friction has already been solved. For a lot of professionals, the friction AI removes is medical paperwork, legal complexity, and the forty-five minutes spent wrestling a spreadsheet formula. That's not meaning-generating difficulty. That's just difficulty.

Still, the central warning holds: that continuous, frictionless satisfaction might not feel like loss at any given moment, which is precisely what makes it hard to notice. For professionals thinking about how to adopt AI thoughtfully, that's a useful lens. Not every inefficiency is waste. Some of it is load-bearing.