Phasing is Morton's fourth hyperobject property. Hyperobjects do not hold steady. They flicker in and out of awareness — climate change visible during a wildfire season, invisible during a mild spring; nuclear radiation palpable during an accident, imperceptible during decades of contamination. Phasing is not a feature of the observer's attention but of the hyperobject's relationship to perceptual apparatus. The entity is always there, but 'there' is a spatiotemporal location the massively distributed entity does not consistently occupy from any observer's perspective. Applied to AI, phasing explains the oscillation Segal captures: 'terror and awe, sometimes in the same minute.' The smooth is devastating on Monday, invisible on Wednesday — and the oscillation is the phenomenological signature of hyperobject engagement.
There are two kinds of Tuesday. On one, AI-augmented work flows — collaboration produces genuine surprise, questions feel alive, attention feels intact. The smooth is invisible, the way gravity is invisible to a person on solid ground. On the other Tuesday, the work feels thin. Prompts feel mechanical. Responses that felt like partnership last week feel like autocomplete this week. The smooth phases into visibility. One catches a glimpse of the flatness, the shallowness, the hollowness of productivity without depth. These are not different days. They are different phases of the same hyperobject. The smooth is present on both Tuesdays, but it manifests differently, revealing different aspects at different moments — the way a three-dimensional object passing through a two-dimensional plane reveals different cross-sections at different points.
Phasing defeats stable assessment. It is impossible to arrive at a fixed judgment — 'AI is good,' 'AI is bad,' 'net positive with manageable costs,' 'catastrophe in slow motion' — because the entity being judged does not hold still long enough for judgment to stabilize. The person declaring on the good Tuesday that AI is the most generous expansion of human capability since writing is not wrong. The person declaring on the bad Tuesday that AI erodes cognitive foundations is not wrong either. Both respond to genuine phases of a genuine entity. Both mistake a phase for the totality. This is why the AI discourse resists resolution. Optimists and pessimists look at different phases of the same hyperobject.
Segal's Trivandrum account captures phasing with clinical precision — excitement and terror coexisting in the same room, in the same person, sometimes in the same sentence. 'I could not tell whether I was watching something being born or something being buried.' The sentence is not ambivalent. It is precise. It describes hyperobject phasing with fidelity no single-valence assessment could achieve. The response Segal proposes — build anyway, build with awareness — meets the phasing with action. Morton's complementary response emphasizes attention: stay with the phasing, stay with the oscillation, refuse the temptation to resolve it.
The temptation is enormous. The mind craves stable ground. The oscillation between terror and awe is expensive, and the mind will seize any narrative promising to end it. 'AI is transformative and the future is bright if we build responsibly' ends the oscillation on the upswing. 'AI is eroding everything that makes us human' ends it on the downswing. Both provide relief. Both are forms of what Morton calls beautiful soul syndrome — the aesthetic pleasure of a clean position exempting one from the discomfort of continuing to attend to an entity that will not hold still. The alternative is to stay in the oscillation, to attend to the phasing without resolving it, to hold terror and awe simultaneously as a perceptual practice adequate to an entity that phases.
Morton derived phasing from the frustrating intermittency of climate awareness. The crisis is undeniable during extreme weather, forgettable during normal seasons. The forgetting is not denial. It is the structure of the relationship between a phasing hyperobject and a finite perceiver whose attention cannot be sustained indefinitely at crisis intensity. The hyperobject does not disappear when it phases out. It withdraws into imperceptibility while continuing to operate.
Applied to AI, phasing explains why neither panic nor complacency stabilizes. The entity keeps revealing new faces — new capabilities, new costs, new ambiguities. The builder who declares 'we have this under control' has mistaken a phase for stability. The critic who declares 'this is the end of human thought' has mistaken a different phase for totality. The entity undulates. The only adequate posture is one that undulates with it — provisional, attentive, willing to revise as the next phase arrives.
Hyperobjects flicker in and out of awareness. Not because observers are inattentive but because the entity's relationship to perception is intermittent by structure.
Different phases reveal different faces. Flow and compulsion, exhilaration and despair, depth and shallowness — all genuine manifestations of the same entity at different moments.
Phasing defeats stable judgment. Fixed assessments mistake a phase for the totality; the entity will reveal a different face tomorrow.
The oscillation is not pathology. It is the phenomenological signature of finite perceivers engaging a phasing hyperobject.
Stay with the phasing. Morton's practice: attend to the oscillation without resolving it, hold contradictory truths simultaneously, refuse the comfort of clean positions.