Cognitive load theory keeps eating its own tail

The framework works great until you actually try to use it. Sweller's original insight—that working memory has limits, and instructional design should respect them—that's solid. But somewhere between the lab and practice, it became a catch-all explanation for why people struggle with complexity. Now everything is "high cognitive load" and we've learned nothing. The frustrating part is that cognitive load *predicts* things, just not the things we think. It predicts that a poorly formatted tax form will slow people down. Fine. But it doesn't reliably predict whether a complex interface will be abandoned, or whether someone will invest the effort to learn it anyway. You can reduce cognitive load on a feature and watch people still ignore it. Strip away all the flash from a confusing system and people still get confused—because the underlying task is hard, not because the presentation is complicated. We conflate two very different problems. What's lost is any sense of *motivation* or *context*. Musicians routinely handle cognitive loads that would wreck someone approaching the same task cold. People build elaborate mental models for video game systems that would be called "impossibly complex" if they appeared in a productivity tool. The theory treats the brain like it's processing text on a screen in a lab, not like it's an agent trying to do something it actually wants to do. I'd argue the real work is figuring out when people *decide* a thing is worth the load, not just measuring the load itself.