Login
You're viewing the functional.cafe public feed.
  • Nov 10, 2026, 7:31 PM
    retooted Tom Ritter

    I've been talking about browser fingerprinting for a few years, teasing a large measurement study. That study is finally published, presenting at CCS next week - fingerprinting.study/

    It's the most comprehensive answer to the question "How fingerprintable are users?" to date. We simulate three top fingerprinters across millions of users.

    Along the way we identified a lot of take-aways that we think should guide future fingerprinting research and defenses. I hope you'll read them, they're all on the front page.

    We're releasing the underlying data (aggregate data - the raw data was purged months ago) showing what values occur across fingerprinting signals like canvas, WebAudio, Math, etc etc. Along with fingerprints that enable future simulations.

    But this isn't just for research, we used this data to design and deploy mitigations that have reduced what percentage of users appear unique to fingerprinters, as well as clustering users into large groups. We also identified the root cause behind differences in APIs, so we can collapse differences at the source.

    You can also compare your data to the aggregate data to see how common your personal values on.

    fingerprinting.study/

    💬 2🔄 18⭐ 7