10 essays · 3 categories
Concepts
Foundations · 3 essays
What perceived performance actually is, how humans perceive time, and where the canonical thresholds come from.
What perceived performance actually is
The dichotomy between objective and perceived performance, the 20 % just-noticeable difference, and where looking-fast stops being a substitute for being-fast and becomes a tax on the user's trust.
5 min read
How humans perceive time
Active vs. passive phases, the dopamine pathway, filled vs. empty duration, prospective vs. retrospective duration judgments, and the one-second active-to-passive transition.
7 min read
The canonical thresholds
Miller 1968's 17-transaction taxonomy, Card et al. 1991's perceptual / immediate-response / unit-task tiers, Doherty's 400 ms productivity cliff, and Nielsen's 1993 distillation. Why the clean 0.1 / 1 / 10 trichotomy is Nielsen's framing, not Miller's.
8 min read
Practice · 5 essays
Applying the science — anatomy of a wait, illusions you can exploit, where the perception layer breaks down, budgets that include perception, and the decision rule for which loading affordance to show in which time band.
The anatomy of a wait
Decompose every wait into pre-action signal → response → animation → completion. What you can tune at each stage. The tip-the-hand rule.
10 min read
Time perception illusions you can exploit
Anstis on contrast and motion. Harrison et al. on backwards-decelerating progress bars. The geometric-mean indifference threshold. When illusions are honest and when they cross into deception.
8 min read
When perceived performance hurts you
Eizenberg's argument, latency JNDs in direct manipulation, and where polished placeholders become polished lies.
8 min read
Performance budgets that include perception
INP and Web Vitals as a starting point. How to add perception to a quantitative budget. The role of adaptive loading and the performance scaler.
7 min read
What cue to show, when
The decision rule: nothing under 1 s, spinner 1–2 s, skeleton or determinate progress 2–10 s, engagement past 10 s. Spinners rehabilitated, the 500 ms delay gate, the prospective–retrospective trade in engaging loading, and empty-state hygiene.
11 min read
AI · 2 essays
Where perceived performance meets AI tools. Why AI waits are shaped differently from the web waits the rest of the platform is tuned for, and where the perception layer breaks down under generative uncertainty.
AI changes the shape of the wait
AI waits are not the deterministic page-load waits the rest of the platform is tuned for. Three properties change: duration variance across two orders of magnitude, conversational shape, and the answer arriving mid-wait. Which classical thresholds survive the move to AI, and which do not.
10 min read
Honesty in AI UX
Where AI perception engineering crosses into deception. Fake streaming, manipulated cadence, confident UX over uncertain output, hidden tool calls, cancellation theatre. The line between polish and lying — and how to stay on the right side of it.
10 min read