Customer inattention pays for much of the economy; agents that make everyone optimise remove the subsidies

A surprising amount of the economy runs on the assumption that most people will not bother.

The 99d newsletter lists the examples. Credit card rewards are paid for by the customers who never optimise. Cable discounts exist because few people ring up and threaten to cancel. Mortgage pricing assumes only a limited number of borrowers will refinance when rates drop. Banks rely on cheap deposits to fund their lending. In the author's words, your laziness or indifference pays for my margin.

Most AI commentary stops at the first-order conclusion: personal agents make ruthless optimisation free for everyone, so everyone wins. 99d argues that this is the wrong conclusion. No system can carry a population made up entirely of optimisers, so the subsidies disappear. Morgan Stanley thinks AI agents could push refinancing frequency up to twice its current rate, leading to increases in mortgage rates overall. Apollo has speculated that agents sweeping cash automatically could drive up the cost of capital economy-wide.

The newsletter expects higher, not lower, barriers: agents deployed against agents, deals withdrawn, friction put back on purpose. And it expects edge to move to what an agent cannot replicate. In its simplest terms, "knowing a guy" or "showing up in a room". Relationships and reputation, in other words.

A note on the evidence: the Morgan Stanley and Apollo points are second-hand in the newsletter, and the rest is the author's reasoning, not data.

I would not take every forecast here as settled. The direction is hard to argue with, though, and the audit it implies is straightforward for anyone designing a business model.

Which of your margins are quietly paid for by customer inattention? And what happens when every customer has a tireless assistant?

Source: 99d, "Do you know a guy?" (2 October 2026), https://lnkd.in/grRmcX7m

#BusinessModelInnovation #AgenticAI #ConsumerBehaviour