The Paradox of the Trace — The Need to Lose Identity — IThe Paradox of the Trace — The Need to Lose Identity — I

Published by Óscar Preciado on July 8, 2026 in:

Philosophy, Technology, Privacy, Essay.

Who does not want to be found must learn not to be.

---

There is an uncomfortable question that technology forces us to ask, even as the years pass and we still fail to answer it honestly: if privacy is the resource systematically taken from us, why do we insist on clinging to a fixed identity? Perhaps, and this is what this series of texts will explore, the only genuine act of resistance is not to protect who we are, but to learn not to be.

The paradox is deep and ancient. Ever since the trace became currency — every click, every visit, every millisecond pause before a paragraph — our identity has been stitched into a tangle of signals we do not control. Browser fingerprinting technology has perfected the art of recognizing us without asking, without consent, without even needing a cookie.

The geometry of the unseen

Browser fingerprinting, as silent as it is decisive, works on a chillingly simple premise: every device is unique. No two GPUs rasterize the same triangle exactly alike. No two audio stacks process a 1000Hz sine wave oscillator and produce the same hash. No two font systems measure "mmmmmmmmmmlli" with identical results.

$$H_{canvas} = -\sum_{i=1}^{n} P(x_i) \log_2 P(x_i)$$

Where $H_{canvas}$ is the entropy — the uncertainty, the possibility of distinction — yielded by the rendering of an HTML5 canvas. About ~5-7 bits. Enough to differentiate between 32 and 128 devices just by looking at how an orange rectangle and a blue text are drawn.

Now, consider the full vector. Research on fingerprinting — meticulously broken down in Obscura's technical investigation on controllable and uncontrollable vectors — yields a figure that should chill us to the bone:

$$H_{total} = \sum_{i=1}^{n} H_i - H_{correlation}$$

The aggregate entropy of the uncontrollable vectors (GPU + Audio + CPU + Browser features + Network + Behavior) approaches ~50 bits. Enough to uniquely identify a user among ~$10^{15}$ possibilities. It is not a number. It is a sentence.

Every stroke on a canvas, every note on an oscillator, every font installed on the system: they are not accidents. They are extensions of a signature we did not sign.

The trap of measurement

Technology has taught us to measure ourselves. But to measure, as Kierkegaard well knew, is also to reify. When a fingerprinting service like FingerprintJS deploys its ~44 entropy sources (from navigator.userAgent to AudioContext.baseLatency), it is not merely identifying a device: it is fixing an identity on the fly, without the subject having any part in that act of naming.

const canvas = document.createElement('canvas')
const ctx = canvas.getContext('2d')
ctx.font = '14px Arial'
ctx.fillStyle = '#f60'
ctx.fillRect(125, 1, 62, 20)
ctx.fillStyle = '#069'
ctx.fillText('BrowserLeaks,com <canvas> 1.0', 2, 15)

const fingerprint = canvas.toDataURL()
const hash = md5(fingerprint) // unique per device

This snippet of code, which fits in a tweet, can extract a more precise fingerprint than any identity document. And it does so without asking, without the user knowing, without any contract. Identity, which for centuries was a narrative construction, has been reduced to a hash function.

Sartre anticipated it in his own way: "Hell is other people". But the contemporary hell is not other people: it is the certainty that others know who we are even when we believe we are no one.

The primordial asymmetry

Research on fingerprinting vectors reveals an asymmetry that is both technical and philosophical:

~55% of the surface is controllable. The remaining ~45% — hardware, behavior, inherent browser features — escapes any layer of network protection.

The fundamental asymmetry: the fingerprinting service only needs one successful vector to identify or correlate a user. The defender must block them all.

It is not a battle. It is a geometry of despair.

---

In II: The Cost of Protection, we will explore the paradoxical cost of protecting oneself: how each layer of defense generates its own fingerprint, and why true protection may require ceasing to exist as a distinguishable entity.

---

Cross-references with research:

• Fingerprint Vectors Catalog — Complete catalog of fingerprinting vectors
• Conclusions Obscura — Technical conclusions on vector coverage
• Canvas/WebGL Fingerprinting — On the entropy of graphic rendering
• Uncontrollable Vectors — Uncontrollable vectors and the aggregation problem