proomt

Search

Search posts, papers, and topics

All posts

Mozilla Automation TeamTom Ritter3 min readintermediate

webaudio fingerprinting on alibaba

Summary

Firefox made WebAudio output deterministic in v118, collapsing most variance to three CPU‑dependent buckets (x86 without FMA, x86/x64 with FMA, ARM NEON). Telemetry shows 99.24% of users fall into these buckets; a long tail of 23 distinct values affects <0.1% of users. Ongoing bugs aim to merge the remaining buckets. Alibaba’s fingerprint script was extracted (two SHA‑256 hashes) and shown to be…

  • WebAudio fingerprinting yields only coarse CPU‑class signals; it does not uniquely identify users.
  • Firefox’s Fingerprinting Protection (v118) normalizes the audio processing path, reducing variance dramatically.
  • Remaining variance is tied to low‑level CPU instruction set differences (FMA vs. no‑FMA, NEON).
  • Telemetry can surface unexpected outliers (23 values from 48 users) likely caused by hardware quirks.

Demonstrates that a well‑engineered mitigation can render a known fingerprinting vector practically useless, and shows how telemetry can guide incremental hardening of privacy defenses.

6/10

Related reading

  1. TeleAntiFraud 2.0: A Refreshable, Profile-Grounded, and Audio-Based Benchmark for Telecom Fraud Detection

    The paper introduces TeleAntiFraud 2.0, a monthly refreshed, frozen audio benchmark of 900 Chinese telecom calls (600 fraud, 300 near‑domain non‑fraud) built with a Mixed‑Tree generation pipeline. Experiments show models that score perfectly on unrelated negatives fall to ~0.66 Macro‑F1 on near‑domain cases, exposing shortcut learning and prediction collapse.

    Hugging Face Daily Papersarxiv.org1 minpaper
  2. High-speed Release Trains – These Weeks in Firefox: Issue 209

    Firefox 155 shipped via a new 2‑week release train; the week’s highlights include a WebExtensions theme API addition, multi‑context address bar in Nightly, DevTools stylesheet editing in the JS debugger, and extensive New Tab page UI/telemetry work (stocks widget, privacy widget, policy‑locked settings). The post also calls out many volunteer contributors and a handful of infrastructure tweaks (T…

    Mozilla Automation Teammozilla.org9 min
  3. ProgramDistill: From Interactive Web Apps to Verifiable Reference-Guided SWE Tasks

    ProgramDistill is a new benchmark that automatically extracts 1,975 replay‑verified feature behaviors from 26 real web apps, builds 4,063 coding‑agent tasks, and measures how well state‑of‑the‑art agents (e.g., GPT‑6 Astra, Claude Opus 5) can reconstruct full or partial applications, revealing steep drops in success as restoration depth grows.

    Hugging Face Daily Papersarxiv.org1 minpaper
  4. FRAUDSkill: Structured Frozen-Weight Skill Optimization for Audio Anti-Fraud Detection

    FRAUDSkill is a framework that keeps a pretrained audio‑language model frozen and learns an external layer of skill programs, routing policies, and decision rules to meet a structured anti‑fraud detection protocol. On the TeleAntiFraud benchmark it reaches 73.5% Macro‑F1 (≈32% improvement) while cutting invalid predictions to 1.94%.

    Hugging Face Daily Papersarxiv.org1 minpaper