Four billion glass slabs, and every few years someone announces the replacement. Then the replacement ships at 600 grams, lasts two hours, and costs as much as a used car — and everyone acts surprised that humanity declines to strap it on. The failures share one ignored variable: mass. Somewhere around 45 grams sits the all-day eyewear limit, the weight at which a device stops being equipment and starts being clothing. Everything above that line is a headset, a session device you visit. Everything below is eyewear, an ambient device that lives on your face. Physics draws this line, not taste, and no industrial design negotiates with it.
Do the power budget and the line gets sharper. At 250–300 Wh/kg, ten grams of battery holds roughly 2.5–3 Wh. Spread across a twelve-hour day of intermittent use, the whole device — display, radio, compute, sensors — must average a few hundred milliwatts. A phone SoC idles above that before it renders a pixel. This is why on-device TOPS-per-watt is the binding curve of the decade, and why heavy lifting stays tethered to the phone or the cloud for the foreseeable future. The glasses are a terminal; the terminal must sip.
Then comes the waveguide trilemma: brightness, transparency, field of view — pick most of two. Diffractive gratings steer photons to the eye at single-digit efficiency, which means most of the light becomes heat and stray glow instead of image. Every demo that looks dim, narrow, or washed out is showing you the same shortfall from a different angle. Progress is real — each waveguide generation roughly doubles efficiency — but ship dates are set by curves, not demos. Watch lumens-per-watt and grating efficiency, not keynote renders.
Here is the underappreciated part: audio already proved the interface can vanish. Earbuds plus voice assistants handle navigation, messages, timers, capture, and calls with zero pixels involved — and billions of daily sessions say the formula works. Add models that see what you see and hear what you hear, and context plus audio covers the large majority of glanceable needs. The screen becomes the exception for rich confirmation, not the rule for every interaction. The slab’s replacement is already halfway shipped; it just doesn’t look like eyewear yet.
This is why roughly 90% of AR interface work fails on contact with reality: it ports the desktop into space. Floating windows, nested menus, pinch-to-resize spreadsheets — the whole WIMP inheritance, now with neck strain. Outdoor attention budgets in seconds, not sessions; nobody browses a file tree while crossing a street. The patterns that survive will be glanceable single glyphs, audio confirmations, and capture-first flows — interfaces measured by how rarely you notice them, which is the opposite of every demo metric the industry currently optimizes.
The stack that actually retires the slab has five compounding curves, and notably, no single miracle: sub-45g frames, phone-grade nearby compute, audio-first interaction, one contextual pixel layer for exceptions, and scene-understanding models running mostly on-device. Each curve is moving on its own schedule. The product exists at their intersection, which is why predicting the year from any single breakthrough keeps failing — you need the conjunction, and conjunctions arrive all at once after looking impossible for years.
Be honest about the timeline by watching the four numbers that matter: waveguide efficiency, battery Wh/kg, on-device TOPS/W, and display nits-per-watt. On current slopes they converge late this decade — which puts the slab’s landline era, the long fade where it still works but stops being where life happens, somewhere in the early 2030s. Anyone selling certainty about the year is selling something else. The curves are public; do the extrapolation yourself and discount every demo by half.
For builders, the implication is to stop designing screens and start designing disappearances. Ship the audio-first flows now, on hardware people already own. Learn the grammar of glanceable, ambient, interruptible computing while the display industry finishes its homework. Because the smartphone will not be beaten by a better screen. It will be retired the way the landline was — not by a superior handset, but by the slow realization that the job it did has dissolved into everything else, and the slab on the table is just where the glass used to live.