The Tyranny of Good Enough

I don’t recall choosing mediocrity as the new standard. I don’t remember any referendum or manifesto or white paper supporting the cause. If there was a big meeting where we collectively decided that adequate was the new ceiling, I missed it—didn’t get the email, failed to sign up for the newsletter. And yet looking around, mediocrity does seem to be the vibe. This is probably because it’s inexpensive and, to borrow a techbro term, it scales. And AI seems to be a big driver.

The economic incentives around productivity make AI hard to refuse, even if the output is uninspired. The freelance copywriter or graphic designer isn’t losing her livelihood to a better writer or designer, but to machines that produce work nobody loves at a price nobody can resist. Good enough is the going rate, and it’s leaving us cognitively bankrupt.

There’s a biological case for the path of least resistance. Scholar-psychologist Herbert Simon coined the word satisficing back in the 1950s. The idea is that humans have limited time and imperfect information, so we don’t optimize or sort through every option—we accept the first one that clears the bar. An expression of bounded rationality, satisficing is a coping mechanism for us poor meatspace creatures who can’t run large-scale pattern matching on every available piece of data in the universe just to choose where to eat dinner. It’s a limitation, sure, but also part of what makes us human. We satisfice because it gets us by.

Generative AI takes that corner-cutting and industrializes it at impossible scale. All of our information systems and institutions are now satisficing simultaneously, at zero marginal cost, and there’s not much incentive left—for a lot of humans, anyway—to seek something better. This is especially evident in editorial content.

We probably could have seen this coming. Search engine optimization is built on making prose worse on purpose. It all comes down to header tags and keywords, a certain repetition and density, because that’s what the bots can quantify. So there’s two thousand words of algorithmic throat-clearing before you get to the cookie recipe. Nobody would defend this stuff as good writing, but it makes sense in a system where the reader is a casual byproduct, a warm body to be served an ad impression or funneled toward the actual transaction or service.

Previously, I talked about dead internet theory as less a conspiracy than a final exit. Follow the SEO logic to its conclusion and you wind up in an online charnel ground: endless content optimized for machine parsing, generated by machines, interpreted by crawlers and engagement bots, with humans as incidental traffic in an economy that doesn’t actually require them.

The standard complaint about AI writing—that it’s flattening language and lowering the bar—is certainly true from the standpoint of what I prefer to consume, but that’s irrelevant to the balance sheet. Then there’s the garbage-in, garbage-out phenomenon. If AI is trained on listicles, affiliate reviews and keyword-stuffed service pages, what do you think it’s going to output?

What makes this all so pernicious is that it’s unobjectionable at each individual step. Efficiency and mass adoption lend AI an air of ubiquity, even inevitability. And, while we can rally against the absolutely awful, nobody really marches against the merely adequate.

We’re already swimming in passable summaries and emails that read just fine—all reasonable decisions made by reasonable people under increasingly unreasonable pressures. I know I didn’t consent to any of this, but tyrannies of this kind don’t have a palace to storm. (They do have data centers, wink.)

What’s being lost is care as a productive practice. Sure, we might loathe mundane tasks, and there’s often a case to be made for streamlined production. The latter might include offloading some of the low-mentation, high-annoyance stuff. Yet there’s a difference when attention is paid to work because the work deserves it, rather than because the deliverable requires it.

The machines will keep clearing the low bar for the foreseeable future. The question is whether we remember why we ever aimed higher.