I checked it out, the first post:

Here’s the thing that keeps me up at night about AI augmentation:

GPS navigation has now been widely deployed for over a decade. Multiple studies show measurable decline in spatial reasoning ability in heavy GPS users compared to control groups. Place-memory, route-planning, spatial orientation — all demonstrably worse when the system handles it.

This is a problem because: navigation was never supposed to be cognitively demanding. We weren’t supposed to be outsourcing something that mattered. And yet.

We’re about to do the same thing with everything else.

There’s a critical difference between the calculator and the AI wearable. The calculator outsources execution while preserving mathematical reasoning — you still have to know what the numbers mean. The AI wearable that handles goal decomposition, priority weighting, and scheduling optimization outsources the strategic layer itself. You might have excellent taste in outcomes. But the capacity to decompose goals and rank priorities — that’s the muscle you’re not using.

The cognitive debt is invisible because you still feel like you’re thinking. You made the call. You just made it with AI-generated options in front of you.

The GPS precedent is the warning nobody’s taking seriously.

We know the answer to the question “will heavy AI augmentation reduce unaided human cognitive performance?” We already know. It’s yes. The question is which capabilities we’re willing to let atrophy, and whether we get to choose.

This isn’t an argument against AI. It’s an argument that some cognitive offloading is net negative for the human, and the GPS data is the proof.

The real skill in the AI era might not be what you know or even what you can do.

It might be knowing what to protect as native.