The image captures something difficult to articulate about the AI-augmented cognitive experience. Ice skating is fast, fluid, and requires real skill — it is not nothing. But the skater never touches the water beneath the ice. She moves across the surface with competence; she does not engage what the surface conceals. Applied to thinking, the image names a genuine phenomenon that productivity metrics cannot detect: the person is producing work, but the work is not descending into the layers where the deepest thinking happens.
The structural mechanism is clear in Murdoch's framework. The depth beneath the surface is where the ego encounters reality's resistance — the inarticulate pre-verbal material, the problem that refuses to simplify, the material that pushes back. These encounters are uncomfortable. The ego flees them when alternatives are available. AI provides the alternative: smooth, articulate, plausible output that allows the person to skate over the top without ever descending.
The first-person testimony is important evidence. Builders using AI often report both the exhilaration (productive flow, impressive output) and the disquiet (a sense that something is being bypassed). The disquiet is difficult to articulate because the productivity is real — the person is not wrong to feel effective. But Murdoch's framework identifies what is being bypassed: the descent into the productive depth where genuine thought occurs. The builder can feel this as absence even when she cannot name it.
The long-term consequence is capability decay. Skating preserves existing capabilities — the person continues to exercise her current understanding — but it does not develop new ones. Genuine development requires the descent into depth, the encounter with what exceeds current understanding, the struggle that forces expansion. A year of skating leaves the person where she started, even as her output volume grows. The tool's capability has grown; hers has not. The gap between her apparent competence (visible in outputs) and her actual competence (available when the tool is absent) widens invisibly.
The phrase 'skating over the surface of my own mind' appears in Segal's You On AI, attributed to a builder reflecting honestly on AI-augmented work. The Murdochian reading treats this first-person testimony as diagnostic of a phenomenon her framework predicts: the ego's triumph over the inner life, facilitated by a tool that makes the triumph easy and invisible.
Surface vs. depth. Real thinking happens beneath the surface, in the encounter with resistant material; AI enables surface-only performance.
The ego prefers surface. The ego's defining preference is for the smooth flow that avoids encounter with its limits.
Phenomenology is misleading. The skating experience is pleasant and feels productive; the absence of depth is not immediately registered.
Capability decays. Skating preserves existing capabilities but does not develop new ones; over time, the gap between apparent and actual competence widens.
Whether all AI-augmented cognitive work involves skating, or whether some deliberate practices preserve the descent into depth, is contested. The practical question is what such practices look like — how a person can use AI for execution while preserving the depth-encounters that develop capability. Early empirical work is exploring this, but reliable answers will require years of longitudinal study.