• Core Discovery:

  • Reverse engineering of Microsoft Paint and Microsoft Photos reveals that AI images generated locally on Copilot+ PCs contain an invisible, server-issued GUID watermark embedded directly into the pixels.
  • While users can toggle visible Copilot watermarks in settings, the invisible pixel watermark cannot be disabled.

• Architecture and Workflow:

  • Local Model Execution: Paint ships with local ONNX models (.onnxe decrypted via XOR keys in segapi.dll) to run Stable Diffusion on the local NPU.
  • Mandatory Remote Moderation: Even for local generation, Paint sends the user's prompt and style over HTTPS to an Azure endpoint (/v1/paint-cocreator/moderate-prompt).
  • GUID Generation: The moderation server responds with a promptGenerationId and a unique watermarkId (GUID). Subsequent generation requests pass the prior ID (lastPromptGenerationId), linking sequential prompts.
  • Watermark Injection: The Watermarker.dll library embeds the 16-byte GUID into the pixel data via WmkWriteWatermark using a content-adaptive block-domain, SVD-style algorithm across 8x8 pixel blocks (modifying thousands of pixels).
  • Enforcement Differences: In Paint, if WmkWriteWatermark fails, the generation process aborts with an error rather than outputting an unwatermarked image. In Photos, it logs an error and still returns the image.

• C2PA Metadata & Soft Binding:

  • Paint submits the image to Azure (/v1/paint-cocreator/image-sign) to obtain a signed C2PA manifest embedded in a caBX PNG chunk.
  • The C2PA manifest contains a c2pa.soft-binding assertion (com.microsoft.invismark.1) holding the exact same watermark GUID embedded in the raw pixels, tying file-level metadata and pixel-level data together.

• Export Format Restrictions:

  • Direct saves and canvas exports restrict formats to C2PA-compatible types (PNG, JPEG, GIF, .paint).
  • Legacy formats like BMP are intentionally excluded because BMP cannot store embedded C2PA manifests without external files.

Hacker News Discussion

• Privacy & De-Anonymization Concerns:

  • Commenters heavily criticized the silent injection of unique GUIDs, noting it eliminates anonymity. If an image is published online, the GUID can be traced via Microsoft servers back to the user account, timestamp, prompt, and device.
  • Parallels were drawn to modern government surveillance and legal risks (e.g., subpoenas identifying meme creators or political dissidents).

• Comparisons to Historical Tracking (Printer Yellow Dots):

  • Many users compared this mechanism to machine identification codes (yellow tracking dots) used by color laser printers for decades, famously used to identify leakers like Reality Winner.
  • Others noted that embedded UUIDs have quietly existed in document formats (DOCX, PDF) and OS telemetry for a long time.

• Bypass and Neutralization Ideas:

  • Replacing or shimming Watermarker.dll with a no-op implementation or intercepting network requests to supply zeroed-out GUIDs.
  • Applying image transformations such as lossy recompression, slight pixel noise, smart directional blur, or re-running through local denoisers to break the watermark pattern.
  • Switching entirely to standalone open-source tools (e.g., ComfyUI, Automatic1111) and Linux to avoid proprietary OS-level telemetry.

指出将信号隐藏在系统提示符中使得其他隐私声明更难以相信，强调了透明度的重要性。

这句话提到了OpenAI与企业客户合作，以更长期的方法来提高安全性，同时支持企业的隐私要求。需要深入了解这些长期方法的细节和效果。

大多数人认为大型科技公司提供的AI服务必然会收集和存储用户数据以改进产品，但作者指出苹果声称其Siri AI采用隐私保护设计，只在用户提问时才访问数据。这一声明挑战了当前AI行业普遍依赖数据收集的做法，暗示苹果可能找到了一种既能提供AI功能又能保护隐私的新模式。

本地编码模型的核心优势在于隐私保护、零成本和完全离线能力。对于处理敏感代码或需要稳定网络环境的开发者来说，这些优势尤为重要。在选择编码工具时，应权衡这些因素与云端模型的便利性和高级功能。

大多数人认为在AI竞赛中，苹果会像其他科技巨头一样，为了提升AI功能而牺牲部分隐私保护。然而，苹果却强调隐私是其AI策略的核心，这与行业普遍认为AI需要大量用户数据才能有效发展的共识相悖，表明苹果在AI领域坚持其隐私至上的价值观，即使这可能限制其AI功能的先进性。

• Distributed AI Training and Scraping: AI companies require massive amounts of web-scraped data for training, search, and agent grounding. Because traditional data centers face heavy blocking and throttling by security services (like Cloudflare and DataDome), scrapers rely on residential proxy networks to route traffic through home internet connections.
• Bright Data's SDK Network: Bright Data operates a massive commercial residential proxy network (marketing over 150M+ to 400M+ IPs). They source these exit nodes by embedding a consent-based Software Development Kit (SDK) inside consumer-facing mobile apps and Connected TV (CTV) / Smart TV applications.
• Why Smart TVs are the Ideal Proxies: Compared to mobile phones, Smart TVs provide a near-perfect infrastructure for proxy routing:
  • They are permanently connected to high-speed home Wi-Fi and grid power (no battery constraints).
  • They run 24/7 in standby mode and offer effectively unlimited bandwidth.
  • They operate largely unattended with virtually no corporate or family oversight.
  • The consent UI on TVs is typically dense text navigated via remote arrow keys, making it unlikely for users to understand that their bandwidth is being sold to third-party scrapers.
• Deceptive Allocation Limits: While opt-in prompts (such as in the Roku app Petflix) claim the SDK "occasionally" uses free resources, the underlying, publicly queryable SDK configuration sets a massive monthly default Wi-Fi budget of up to 200 GB (max_bw_monthly_wifi: 200,000,000,000 bytes).
• Notable SDK Integration Partners: Public, unauthenticated partner manifest endpoints expose integrations with platforms reaching hundreds of millions of households, including:
  • PlayWorks Digital Ltd: Over 400 CTV game titles across Comcast, Sky, Cox, LG, Samsung, Vizio, and Roku.
  • CloudTV: Integrated across more than 125 TV brands and 15+ OEMs.
  • Viber Media (Rakuten): Massive messaging app ecosystem.
  • Supercent & Moonfrog Labs: Major mobile game publishers.
• Technical Reverse-Engineering & VPN Bypasses: Technical analysis of the iOS framework (brdsdk.framework) reveals that:
  • The SDK dials out to a persistent WebSocket connection tracking device metrics (CPU, memory, network state, battery level).
  • Bypassing VPNs: By forcing network interface bindings directly to Wi-Fi (en0) or cellular (pdp_ip0) instead of the system default route, the SDK completely bypasses user-configured local VPN tunnels (tun0).
  • Broad Definition of "Idle": The SDK configuration allows relaying traffic even when the user is actively on a phone call or the screen is on, provided CPU utilization remains below 70% and memory below 90%.
• Cross-Platform Identity Stitching: The SDK's config file contains tracking properties like dual_pairing maps designed to tie a single user's distinct installations across iOS, Windows, and macOS together into a single unified identity.
• Mitigation and Defense Strategies:
  • DNS Sinkholing: Network-wide blocking of key domains (proxyjs.brdtnet.com, proxyjs.luminatinet.com, proxyjs.bright-sdk.com, and clientsdk.bright-sdk.com) entirely kills the proxy peer tunnel without impacting legitimate public traffic.
  • Network Boundaries: Utilizing TLS SNI filtering on domains matching *.brdtnet.com or *.luminatinet.com.
  • MDM Application Auditing: For enterprise environments, scanning mobile binaries for unique Swift symbols like BrdWebSocketFacade and BrdNetwork.DNSResolver to filter out infected applications.

Hacker News Discussion

• The Irony of Cloud-to-Cloud Scrapes: Users point out the profound irony that both the AI data scrapers and the target websites being scraped are often simultaneously hosted on AWS infrastructures, engaging in a costly, artificial cat-and-mouse game to mask their identities.
• Strict Hardware Isolation ("Dumb" Displays): A popular consensus among commenters is to completely air-gap or isolate smart TVs from the internet, relying exclusively on local HDMI inputs connected to trusted devices (like Apple TV, HTPCs, or Home Assistant setups).
• Automatic Content Recognition (ACR) over HDMI: Contributors point out that simply removing network permissions may not protect privacy entirely if a TV is ever connected later. Academic papers cited in the thread reveal that Smart TVs run Automatic Content Recognition to analyze and log content even on local HDMI inputs while offline, caching data to upload the moment an internet connection becomes available.
• The Threat of VPN Bypassing: The community expressed severe alarm regarding the SDK's ability to explicitly bypass local system VPN configurations via forced network interface bindings, highlighting the growing complexity required to self-host secure, consumer-friendly networks.
• Legal Risks and Misleading Consent: Commenters note that the SDK text hides behind the guise of "downloading public data," masking that its true utility is to circumvent security blocks. There is also discussion regarding the liability risk for home residents if a malicious third party utilizes their residential IP address through these unregulated networks for illicit activities (e.g., severe cybercrimes), though others note Bright Data utilizes strict Know-Your-Customer (KYC) onboarding for their buyers.
• Network-Level Defense: Users shared practical setups for containment, such as creating isolated local VLANs with restrictive firewall configurations, whitelisting device MAC addresses via DHCP policies, and deploying Pi-holes or AdGuard Home setups to drop the domains mentioned in the report.

这里描述的数据收集范围远超家长购买玩具时的想象。情感语气分析和行为参与模式本质上是对儿童的心理画像——生成关于发展脆弱性、情绪触发点和兴趣图谱的洞察，这些数据可能保存数十年，并在家长毫无有效救济手段的情况下被出售或泄露。COPPA正是为此而生，但执法速度远远落后于技术能力的发展。

• Evolution of Cars into Computers: Modern vehicles are now highly sophisticated computers on wheels that continuously harvest deeply intimate details of their drivers' lives to generate corporate revenue.
• Extensive Data Harvesting: Car manufacturers track data points that extend far beyond simple GPS routing. Collected data includes exact geographic location, passenger presence, radio selection, seatbelt usage, acceleration and braking behaviors, weight, age, race, and even facial expressions via internal cabin cameras.
• Monetization and Financial Risks: This data is regularly passed to or bought by insurance companies, who utilize driving behavior analytics to raise premium costs for consumers. Most car companies acknowledge selling this information, though they rarely name the specific buyers.
• Pervasive Market Penetration: According to consulting firm McKinsey, internet connectivity in households' vehicles reached 50% in 2021 and is projected to hit 95% by 2030, making telemetry tracking near-universal for modern drivers.
• Complete Privacy Failure: A rigorous 2023 evaluation by Mozilla scrutinized 25 major automotive brands and found that every single one failed to meet minimum privacy and security standards, officially labeling cars as the absolute worst product category ever reviewed for consumer privacy.
• Impending Federal Telematics Escalation: Upcoming US federal mandates will worsen the privacy landscape by requiring automakers to integrate infrared biometric cameras and behavioral scanners to catch drunk or fatigued drivers. However, there are no legal frameworks limiting how manufacturers can monetize or distribute this newly mandated health and biometric data.

Hacker News Discussion

• Two-Sided Surveillance Encroachment: Users point out that privacy is under attack from both inside the vehicle (internal cameras and telemetry) and outside the vehicle (omnipresent roadside cameras and license plate readers like Flock).
• Need for Broad Legislative Reform: There is a strong consensus that individual consumer choice is insufficient; legal frameworks must step in to strictly regulate all forms of data collection, third-party sharing, and camera placement without public consent.
• Pervasive Corporate Distopia: The discussion highlights a growing sense of exhaustion and resignation over 24/7 corporate tracking across all areas of life, comparing the sleepwalk into a corporate police-state to an inescapable reality.
• Hardware and Software Workarounds: Tech-savvy users discuss actively tampering with their vehicles to regain privacy, such as physically yanking out the integrated cellular modem boards or using specialized diagnostic software (like ForScan) to fully disable built-in telemetry features.

Brooding: In 2026, I Resolve to Friction-Maxx<br /> by [[Kathryn Jezer-Morton]]<br /> accessed on 2026-05-12T10:31:16

read on 2026-05-28T17:02:02

大多数人认为AI公司会默认使用用户数据进行模型训练以提高产品性能。但Anthropic明确表示默认情况下不会使用用户数据训练模型，这是一个与行业惯例相悖的做法，反映了他们对数据隐私的重视和对用户信任的承诺。

大多数人认为AI公司会默认使用用户数据进行模型训练以改进产品。但作者明确表示Anthropic不会默认使用客户数据进行训练，这挑战了AI行业普遍的数据收集和训练实践，是一个非共识的隐私立场。

🔒【令人震惊】Chrome 在数十亿设备上静默写入 4GB Gemini Nano 模型权重，删除后自动重装，可能违反 GDPR。这是「端侧 AI」与用户隐私的第一次正面冲突——不是关于数据收集，而是关于在未经同意的情况下使用用户存储空间和计算资源。这个事件的先例意义巨大：如果 Google 可以这样做，所有内置 AI 的操作系统和浏览器都有可能效仿，用户对自己设备的控制权正在被悄悄侵蚀。

• Bundestag Security Shift: Bundestag President Julia Klöckner has recommended that members of the German Parliament switch from Signal to "Wire," a BSI-certified messenger, as the new standard for communication.
• Digital Sovereignty: The move is framed as a step toward digital sovereignty, reducing reliance on US-based platforms like Signal or WhatsApp in favor of a service with European roots and German security certification.
• Phishing Mitigation: A primary driver for the recommendation is security; Wire allows registration via email rather than a phone number, which is intended to hide a central identification feature and make phishing attacks more difficult.
• BSI Certification: The "Wire Bund" version has been approved for data classified as "Verschlusssache – nur für den Dienstgebrauch" (Restricted) until 2028.
• Human Factor vs. Technology: Critics and experts note that while Wire is secure, it is not a "panacea." Recent successful phishing attacks against politicians (including Klöckner herself) highlight that the human user is often the weakest link, regardless of the app's encryption.

Hacker News Discussion

• Vendor Lock-in Irony: Commenters pointed out the irony of moving from one vendor-locked system (Signal/US) to another (Wire/German-Swiss), questioning why the government didn't choose Matrix, which is an open standard used by NATO and other EU entities.
• Deployment Details: A former developer shared that Wire was originally deployed for the Bundeskanzleramt using a Nix-based delivery method to allow for completely air-gapped server installations.
• Skepticism of Motivation: Some users suggested the switch might be politically motivated or a way for Klöckner to deflect from her own experience being phished, rather than a purely technical security upgrade.
• Data Privacy Concerns: Discussion arose regarding jurisdiction; while Signal is US-based, Wire is subject to German/Swiss law. This is seen as a benefit for EU sovereignty but also raises questions about local legal intercept requirements.
• Technical Comparisons: Users debated the UX and backup reliability of Wire versus Signal, with some noting that Wire's media backup system has historically been less robust than Signal's.

这段话揭示了隐私保护的复杂性，并非仅仅是法律问题，而是涉及到获取数据的难易程度。

This quote highlights a significant requirement for users, which may raise concerns about privacy and the feasibility of such a process.

大多数人认为使用Google的AI模型必然会涉及数据传输和隐私问题，但作者强调模型完全在设备上运行且不向Google发送数据。这与人们对大型科技公司AI服务通常涉及数据收集的普遍认知相悖，暗示Chrome的AI功能可能比想象的更加注重隐私保护。

https://web.archive.org/web/20260425190331/https://buttondown.com/maiht3k/archive/why-you-should-refuse-to-let-your-doctor-record/

Impact of having doctor visit being transcribed by AI The GDPR issues aside, there is a strong indication that writing is the thinking for physicians but they might not realise that. - [ ] return #pkm

大多数人认为Meta公司会尊重用户隐私，但作者指出Meta现在在其员工的工作电脑上安装监控软件，这表明公司可能并不总是将用户隐私放在首位。

这是一个令人惊讶的反讽：一个强调隐私保护的社交平台，却要求用户禁用隐私保护扩展才能正常访问，暗示平台商业利益与用户隐私保护之间存在根本冲突。

这一陈述暗示了隐私保护工具与主流平台功能之间存在根本冲突，反映了平台商业利益与用户隐私权之间的紧张关系。这揭示了一个令人不安的现实：追求隐私可能意味着牺牲数字参与度。

这句话暗示了隐私保护工具与主流社交平台之间的潜在冲突。这反映了数字隐私与平台商业利益之间的张力。用户安装隐私扩展通常是为了保护数据不被收集，但平台可能将这些工具视为干扰其数据收集和分析的障碍。这种冲突预示着未来网络环境中隐私保护与平台功能之间的持续博弈。

这句话暗示了隐私保护工具与主流网站服务之间的潜在冲突。这反映了数字时代的一个核心矛盾：用户想要保护自己的隐私，而平台则需要收集数据来提供个性化服务。这种冲突可能导致用户在隐私和便利性之间做出艰难选择。

这一警告暗示了隐私保护工具与主流平台之间的根本冲突，反映了平台商业利益与用户隐私权之间的紧张关系。用户被迫在隐私和功能之间做出选择，这揭示了现代数字生态系统中用户权利被系统性削弱的令人担忧的趋势。

这一声明暗示了平台与用户隐私工具之间的潜在冲突，揭示了社交媒体平台可能对用户隐私保护工具的排斥态度。这引发了一个值得深思的问题：平台安全与用户隐私之间的界限在哪里？

这是一个令人惊讶的声明，暗示社交媒体平台可能主动阻止用户使用隐私保护工具。这可能表明X平台的数据收集策略与用户隐私保护之间存在根本冲突，值得深入研究其商业模式与用户权利的平衡问题。

这是一个令人深思的矛盾点：本应保护用户隐私的浏览器扩展反而可能导致平台功能失效。这暗示了X(前Twitter)的某些功能可能依赖于数据收集，与用户隐私保护存在根本性冲突，反映了数字服务中隐私与功能的持续博弈。

这一陈述暗示了隐私保护工具与主流平台之间的潜在冲突，揭示了平台方与用户隐私保护之间的紧张关系。这表明在当前互联网生态中，用户为保护隐私而采取的措施可能被平台视为'问题'，反映了平台利益与用户隐私权之间的根本性矛盾。

这个声明揭示了平台与用户隐私工具之间的紧张关系，暗示X(推特)可能故意限制隐私功能以收集更多数据，这种商业利益与用户隐私的冲突是当今数字平台的核心矛盾之一。

这句话暗示了一个令人深思的悖论：用户安装隐私保护工具（如广告拦截器、隐私增强扩展）来保护自己的数据，但这些工具反而可能阻止他们访问平台。这揭示了平台利益与用户隐私保护之间的冲突，以及现代互联网服务对用户数据的依赖程度。

这个隐私声明揭示了Sage的数据处理策略，采用了最小化数据传输的设计哲学。这种平衡安全与隐私的做法很有洞察力，表明开发者理解用户对数据泄露的担忧，同时认识到某些云端分析对于有效威胁检测的必要性。

令人惊讶的是：Sage 采用了一种平衡隐私和安全的方法，只将URL和包哈希发送到云端进行声誉检查，而文件内容、命令和源代码则保留在本地。这种设计既提供了实时的威胁检测，又保护了用户的敏感数据，反映了现代安全工具对隐私保护的重视。

令人惊讶的是：专有API背后的封闭式代理工具迫使用户将代理记忆的控制权让渡给第三方。这意味着用户在使用AI代理时可能不知不觉地失去了对自己数据和个人信息的控制权，这可能引发严重的隐私和安全问题。

令人惊讶的是：伊利诺伊州在2008年就通过了生物特征信息隐私法，比许多州的AI监管立法早了近15年。这表明该州在技术监管方面一直处于前沿，从生物识别数据到AI，该州似乎总是提前应对新兴技术带来的隐私挑战。

令人惊讶的是：Excel中的ChatGPT功能与普通聊天历史是完全隔离的，两个系统之间没有数据同步。这意味着用户可以在Excel中使用AI处理敏感数据，而不用担心这些信息会出现在他们的常规聊天记录中，提供了额外的隐私保护层。

令人惊讶的是：企业级用户的Excel数据默认不会被用于训练AI模型，这与普通用户的数据处理方式有显著区别。这种差异反映了OpenAI对商业客户隐私的特别保护，可能是为了增强企业采用AI工具的信心。

令人惊讶的是：即使Signal应用被从iPhone上删除，苹果设备的内部通知存储系统仍然保留了收到的消息内容。这表明iOS系统在应用删除后仍会缓存通知数据，这可能成为执法机构获取已删除消息的意外途径，而大多数用户并不意识到这一潜在的数据泄露风险。

大多数人认为AI安全研究主要集中在防止恶意使用和确保系统对齐人类价值观上。但作者将隐私保护方法列为优先领域，这表明OpenAI正在将隐私视为安全的核心组成部分，而非一个独立考虑的因素，这与传统上将隐私和安全视为两个不同领域的观点相悖。

• The author investigated the privacy implications of LinkedIn's "Identity Verification" badge, which is handled by a third-party provider called Persona.
• To verify, users must provide a government ID, a live "selfie" (biometric scan), and access to NFC data on their passport.
• Upon examining the legal disclosures, the author found that data is shared with a surprisingly long list of 17 subprocessors.
• Notable subprocessors listed include OpenAI, Anthropic, and Groqcloud, raising concerns about biometric and personal data being used for AI training.
• The privacy policy cites "Legitimate Interest" as the legal basis for processing, which the author suggests is a way to bypass explicit consent for data sharing.
• The article concludes that the perceived value of a "verified" checkmark on a professional profile does not outweigh the significant privacy risks and the distribution of sensitive biometric data to numerous tech corporations.

Hacker News Discussion

• A top comment notes that Persona’s CEO responded to the claims, asserting that no personal data is used for AI training and that biometric data is deleted immediately after processing.
• The CEO claimed the full list of subprocessors is a "catch-all" for all their services (like background checks), and only 8 subprocessors are actually used for the identity verification product specifically.
• Many users remained skeptical of the "catch-all" explanation, arguing that if a company is sure about privacy, they should explicitly limit the legal language rather than using broad "CYA" (Cover Your Assets) terms.
• Commenters discussed the inherent risks of "Know Your Customer" (KYC) processes, noting that even if data is deleted, the initial collection creates a honeypot for potential breaches.
• Some participants argued that the "Legitimate Interest" justification is often abused in the EU to avoid the stricter requirements of explicit consent under GDPR.
• There was a general consensus among technical users that a social media badge is insufficient justification for handing over highly sensitive, unchangeable biometric markers.

• Project Overview: The author developed "Bluehood," a Python-based Bluetooth scanner, to demonstrate the extensive metadata leaked by devices merely by having Bluetooth enabled.
• Motivation: Triggered by a critical vulnerability (WhisperPair CVE-2025-36911) and a desire to visualize invisible digital footprints, the project highlights how "invisible" signals compromise privacy.
• What Bluetooth Reveals About Users: By monitoring signals passively, the author could determine:
  • Delivery Logistics: Exact arrival times of delivery vehicles and whether the same driver visits repeatedly.
  • Daily Routines: The specific daily patterns of neighbors based on their phone and wearable broadcasts.
  • Device Associations: Which devices belong to the same person (e.g., a specific phone moving in tandem with a specific smartwatch).
  • Occupancy & Location: Exact times people are home, at work, or elsewhere.
  • Security Vulnerabilities: Periods when a house is typically empty.
  • Social Patterns: Regular visitors (e.g., someone visiting every Thursday afternoon).
  • Employment Indicators: Patterns that suggest specific work types, such as shift work.
  • Family Schedules: Specific times children return home from school.
  • Consumer Habits: Which households share the same delivery drivers, implying similar shopping preferences.
  • Incident Evidence: Retrospective logs of who was present (passersby, dog walkers) during specific events like property damage.
• Uncontrollable Broadcasts:
  • Many devices broadcast continuously without user recourse, including medical implants (pacemakers, hearing aids), modern vehicles, and smart home tech.
  • Privacy tools like Briar or BitChat require Bluetooth for off-grid mesh networking, creating a paradox where privacy tools necessitate privacy leaks.
• Technical Functionality:
  • Bluehood uses passive scanning to identify vendors and device types without connecting.
  • It analyzes patterns (heatmaps, dwell times) and filters out randomized MAC addresses to focus on persistent tracking.

Hacker News Discussion

• Ubiquitous Tracking: Commenters confirmed that similar tracking is common in retail (using iBeacons to track shoppers to specific shelves) and via vehicle sensors (TPMS in tires broadcasting unique IDs).
• WiFi vs. Bluetooth: Users noted that WiFi signals from cars (often named "Audi", "Tesla", etc.) are just as leaky as Bluetooth, allowing for easy "wardriving" profiles.
• Medical Privacy: Significant concern was raised regarding medical devices (like CPAP machines) that broadcast 24/7, often to satisfy insurance requirements, with no way for the patient to disable the radio.
• Mitigation Strategies:
  • OS Features: GrapheneOS and recent Android versions offer settings to automatically turn off Bluetooth after a period of inactivity.
  • iOS Limitations: Apple users noted it is harder to keep Bluetooth permanently off without diving into settings or using Shortcuts, as the Control Center toggles are temporary.
• Legal Context: Several users pointed out that while such tracking is rampant in some regions, it is strictly regulated or forbidden in the EU without explicit consent.

if you don't have personal stuff or personal data on others in your vault, privacy is less a concern with cloud models. True. Except I think any pkm is about personal knowledge and while not personal data per se, there is a vulnerability involved there.

• Microsoft provided the FBI with BitLocker recovery keys for three laptops seized in a fraud investigation related to Guam's Pandemic Unemployment Assistance program.
• BitLocker, Windows' full-disk encryption enabled by default, uploads recovery keys to Microsoft's cloud by default, allowing access by Microsoft and law enforcement with a warrant.
• Microsoft receives about 20 such requests per year and complies; a spokesperson did not comment for TechCrunch.
• Cryptography expert Matthew Green criticized Microsoft for not securing keys better, noting repeated cloud breaches and industry lag.
• Risks include hackers compromising Microsoft's cloud to steal keys (requiring physical drive access) and privacy concerns from key escrow.

Hacker News Discussion

• Users debate defaults: reasonable for average users to protect against theft while allowing recovery, but power users should avoid cloud upload using local accounts.
• Complaints about Microsoft pushing Microsoft accounts, auto-uploading data (e.g., via Teams, Edge), and difficulty opting out without re-encrypting.
• Comparisons to Apple FileVault/iCloud: Apple prompts choice and uses E2EE keychain; Microsoft criticized for sneaky behavior and lacking clear warnings.
• Suggestions to switch to Linux for privacy; distrust in Microsoft due to ads, updates overriding settings, and potential key upload even when opted out.
• Broader concerns: warrants compel compliance, but better design (no key collection) prevents issues; cosmic ray bit-flips or bugs could accidentally upload keys.

• The author describes a comprehensive setup to block virtually all online advertising across devices and services.
• They focus on network-level filtering instead of per-device ad blockers, so that phones, TVs, and other clients benefit automatically.
• The core of the solution is running a self-hosted DNS-based blocker (like Pi-hole or AdGuard Home) to sinkhole common ad and tracker domains.
• Additional blocklists are layered on top to handle more aggressive tracking and region-specific ad domains, trading a bit of breakage for increased privacy.
• For services that hardcode ad endpoints or use techniques that bypass DNS blocking, the author uses more advanced tools such as proxying or firewall rules.
• Some apps and sites break when ads are blocked; in those cases, the author selectively whitelists domains or uses per-device exceptions rather than relaxing global rules.
• On mobile, encrypted DNS and VPN-like tunneling are configured so that all traffic still flows through the home-level blocking setup even on the go.
• The author argues that this configuration significantly improves page load times, reduces bandwidth usage, and makes devices feel faster and less cluttered.
• They acknowledge an ethical gray area with ad blocking but conclude that user safety, privacy, and mental comfort outweigh the downsides of depriving low-quality ad networks of revenue.
• The piece emphasizes that the goal is not absolute perfection but a sustainable setup that requires minimal maintenance once deployed.

Hacker News Discussion

• Commenters discuss additional tools like SponsorBlock for skipping in-video sponsorships on platforms such as YouTube, highlighting that traditional ad blockers do not remove creator-embedded promos.
• Several users point out that DNS-level blocking does not stop ads injected directly by streaming services, noting that such platforms often use certificate pinning or app-level tricks that make proxying and MITM approaches difficult or impossible.
• A highly upvoted comment recommends using a user-agent switcher to bypass sites that block non-Chrome browsers, with examples where services claim to be incompatible with Firefox but run better once the browser “pretends” to be Chrome.
• Participants criticize websites that enforce brittle user-agent checks instead of feature detection, arguing that this needlessly breaks otherwise compatible browsers and punishes privacy-conscious users.
• Some users express skepticism about privacy-focused browsers that are built on or dependent on codebases controlled by ad-driven companies, calling out an inherent tension between privacy promises and ad-based business models.

• Companies aggressively push app downloads, especially in places like Taiwan, offering discounts but often installing without full consent, leading to spam and unwanted data collection.
• Avoid handing over your phone to staff and never download apps, as they provide minimal benefits compared to the risks involved.
• Primary risks include surveillance capitalism: apps enable extensive data tracking for targeted ads and "surveillance pricing," where prices vary based on inferred financial status (e.g., charging more after payday).
• This undermines fair pricing, giving corporations power over individual costs beyond market forces.
• Apps enforce binding arbitration clauses in Terms of Service, waiving rights to court, jury trials, or oversight; examples include Disney attempting to force arbitration in a wrongful death case linked to a Disney+ trial.
• Predictions highlight future abuses, like arbitration forced via unrelated services (e.g., Uber Eats leading to self-driving car disputes).
• Recommendation: Use websites or PWAs instead to preserve privacy and rights.

Hacker News Discussion

• Users debate apps vs. websites/PWAs: many praise PWAs (e.g., Mastodon, Photoprism) for performance when implemented well, criticizing poor web apps and noting apps often wrap webviews with extra tracking.
• Privacy concerns dominate: native apps access more device data (contacts, SMS, biometrics, etc.) even with permissions, unlike sandboxed PWAs; tools like NetGuard suggested for blocking app internet access.
• Loyalty discounts viewed as modern coupons by some, saving money despite data sharing, but others warn of surveillance pricing via purchase patterns and arbitration risks.
• Experiences shared: retailers reject Apple Pay to force accounts; global pushiness for apps noted; arbitrage limits price discrimination viability.
• Calls for better OS controls, open-source apps without tracking, and skepticism of app store security.

https://www.youtube.com/watch?v=5ZNKLmIg2dc

Yoose, B., & Shockey, N. (2024). Navigating Risk in Vendor Data Privacy Practices: An Analysis of Springer Nature's SpringerLink (Version v1). SPARC. https://doi.org/10.5281/zenodo.13886473

Abstract

Navigating Risk in Vendor Data Privacy Practices: An Analysis of Springer Nature's SpringerLink documents a variety of practices that undermine library privacy standards. SpringerLink provides a case study in the encroachment of the broader surveillance-based data brokering economy into academic systems. Combined with our 2023 report on Elsevier’s ScienceDirect platform, this analysis illustrates the wide range of privacy risks inherent in the business models and practices in the academic scholarship marketplace.

Among other findings, the report documents risks related to the 200 named third parties that are allowed to collect information from users of the site (along with what appear to be additional unlisted companies found only in our public website analysis). While the specific privacy concerns posed by SpringerLink are different, our analysis reiterates the findings from our ScienceDirect report: that user tracking that would be unthinkable in a physical library setting now happens routinely through publisher platforms.

While this analysis and recommended actions are grounded in the library context, these findings will raise pressing issues for faculty, administrators, and policymakers to consider as well. The report closes with suggested actions that libraries can take over both the short and long term to address vendor privacy risks.

Taken to its extreme, this seems to suggest that Bose can use anything you post on Facebook as part of its marketing materials.

This sounds like Bose can listen to conversations happening around you (if they so choose).

By agreeing to this privacy policy, you are essentially letting Bose know everything you watch and listen to.

Detailed Summary

1. You own your data, in spite of the cloud. <br /> Section summary: <br /> Local-fist software tries to solve the problem of ownership, agency and data lock-in present in cloud-based software, without compromising cross-collaboration and improving user control.

Section breakdown<br /> §1: SaaS<br /> Pros: Easy sync across devices, real-time collab Cons: loss of ownership and agency; loss of data is software is lost.

§2: Local-fist software<br /> - Enables collaboration & ownership - Offline cross-collaboration - Improved security, privacy, long-term preservation & user control of data

§3 & §4: Article Methodology<br /> - Survey of existing storage & sharing approaches and their trade-offs - Conflict-free Replicated Data Types (CRDTs), natively multi-user - Analysis of challenges of the data model as implemented at Ink & Switch - Analysis of CRDT viability, UI - Suggestion of next steps

2. Motivation: collaboration and ownership<br /> Section summary: <br /> The argument for cross-device, real-time collab PLUS personal ownership

Section breakdown<br /> §1: Examples of online collabs<br /> §2: SaaS increasingly critical, data increasingly valuable<br /> §3: There are cons<br /> §4: Deep emotional attachment to your data brings feeling of ownership, especially for creative expression<br /> §5: SaaS require access to 3rd party server, limitation on what can be done. Cloud provider owns the data.<br /> §6: SaaS: no service, no data. If service is shut down, you might manage to export data, but you may not be able to run your copy of the software.<br /> §7: Old-fashioned apps were local-disk based (IDEs, git, CAD). You can archive, backup, access or do whatever with the data without 3rd party approval.<br /> §8: Can we have collaboration AND ownership?<br /> §9: Desire: cross-device, real-time collab PLUS personal ownership

3. Seven ideals for local-first software<br /> Section breakdown<br /> §1: Belief: data ownership & real-time collab are compatible<br /> §2: Local-first software local storage & local networks are primary, server secondary<br /> §3: SaaS: In the server, or it didn't happen. Local-first: local is authoritative, servers are for cross-device.

3.1.1 No spinners<br /> SaaS feels slower because if requires round-trip to a server for data modification and some lookups. Lo-Fi doesn't have dependency on server, data sync happens on the background. This is no guarantee of fast software, but there's a potential for near-instant response.<br /> 3.1.2 Data not trapped on one device <br /> Data sync will be discussed in another section. Server works as off-site backup. The issue of conflict will also be discussed later.<br /> 3.1.3 The network is optional<br /> It's difficult to retrofit offline support to SaaS. Lo-Fi allows CRUD offline and data sync might not require the Internet: Bluetooth/local Wi-fi could be enough.<br /> 3.1.4 Seamless collabs<br /> Conflicts can be tricky for complex file formats. Google Docs became de facto standard. This is the biggest challenge for Lo-Fi, but is believed to be possible. It's also expected that Lo-Fi supports multiple collab.

TBC

https://www.theguardian.com/us-news/2025/mar/21/oil-protest-activism-greenpeace-dakota-pipeline-verdict

by Erik Rye, Researcher, University of Maryland

Wi-Fi Positioning Systems are used by modern mobile operating systems to geolocate themselves without the use of GPS. Both Google and Apple, for instance, run Wi-Fi Positioning Systems for Android and iOS devices to obtain their own location using nearby Wi-Fi access points as landmarks.

In this work, we show that Apple's Wi-Fi Positioning System represents a global threat to the privacy of hundreds of millions of people. When iOS devices need to geolocate themselves using nearby Wi-Fi landmarks, they transmit a list of hardware identifiers to Apple and receive the geolocations of those access points in return. Unfortunately, this process can be replicated by an unprivileged adversary, who can recreate a copy of Apple's Wi-Fi geolocation database by requesting the locations of access points around the world with no prior knowledge.

To make matters worse, we demonstrate that by repeatedly querying Apple's Wi-Fi Positioning System for the same identifiers, we can detect Wi-Fi router movement over time. In our data, we see evidence of home relocations, family vacations, and the aftermath of natural disasters like the 2023 Maui wildfires. More disturbingly, we also observe troop and refugee movements into and out of the Ukraine war and the impact of the war in Gaza.

We conclude by detailing our efforts at responsible disclosure, and offer a number of suggestions for limiting Wi-Fi Positioning Systems' effects on user privacy in the future.

Full Abstract and Presentation Materials

Kein How-to-Guide, sondern eine kurze Beschreibung der Besonderheiten von Signal für technisch weniger bewanderte Nutzerinnen und Nutzer. Bevor ich dieses Blogpost gelesen hatte, wusste ich nicht, dass Signal (die Stiftung, die für die Software und Hardware des Dienstes verantwortlich ist) nicht einmal Daten über Teilnehmer:innen an Chats und Gruppenchats entschlüsseln kann. Der Service weigert sich also nicht nur Metadaten über Kommunikation herauszugeben, es kann es gar nicht.

Es wirkt vielleicht paranoid, die eigene Kommunikation vor jeder möglichen Einsichtnahme schützen zu wollen. Aber angesichts von autoritären Politiker:innen, die im Bündnis mit monopolistischen Digitalkonzernen handeln, ist es wichtig ist, an alle Eventualitäten zu denken. Außerdem machen sich, wenn solche Techniken nicht massenhaft verwendet werden. genau die Leute verdächtig, die vor allem Verfolgung befürchten müssen.

Ich selbst verwende Signal auf meinem Mobiltelefon mit /e/OS, also einem weitgehend „Google-freien“, auf Datenschutz ausgerichteten Android. Ich lade die App von einem Repository für Open Source Software (F-Droid mit dem Guardian Project-Repository), also nicht von Google Play. Trotzdem bin ich nicht ganz sicher, dass meine Kommunikation über Signal nicht beobachtet werden kann. Mit rein technischen Mitteln lässt sich Sicherheit nicht herstellen, weil man die Möglichkeiten in Kommunikation einzudringen nicht alle voraussehen und ausschließen kann. Ein Restrisiko bleibt, aber ich versuche es so klein wie möglich zu halten.

unduly dictates one particular technical approach

more complete: for who? by what means? what implications?

cf matters of disclosure, consent, and differing orientation to/with privacy: MIT Tech Review article on CMU Mites in TCS Hall

Here's what ChatGPT shown me as summary of their Privacy policy: Analytics and Third-Party Services They use third-party analytics tools (e.g., Google Analytics) to understand how users interact with their services. These tools may collect anonymized or aggregated data, such as IP addresses or browsing behavior, but not personally identifiable information unless you've explicitly consented. Cookies and Tracking: Their cookies track usage patterns on their platform, which could include data about visited pages, clicks, or time spent on the site. However, this data is anonymized unless tied to your account or explicitly authorized.

^^ let's just trust em, right? or no? Ping me if that ended up being a no-no as I decide to continue using the app.

for - Indyweb dev - Mem.ai has many features we are designing for in Indyweb but it uses AI and that needs to be researched for privacy issues - AI - mem.AI - first AI note app that trains directly on your own personal knowledge notes

for - Indyweb dev - vulnerabilities of software designed with centralized privacy

for - definition - privacy principle - quantum informational panpsyichism theory of consciousness - Federico Faggin - Giacomo Mauro D'Airiano

definition - privacy principle - experience isnot shareable, even in principle

If the federal government had access to every email you’ve ever written and every phone call you’ve ever made, it’s almost certain that they could find something you’ve done which violates a provision in the 27,000 pages of federal statues or 10,000 administrative regulations. You probably do have something to hide, you just don’t know it yet.

Question about the privacy of the user interacting with the RAG

It doesn't seem like JSTOR has thought about this thoroughly. In the beta there is kind of an expectation that the service is being validated so what users are doing is being watched closer.

Have locks on data is a nice idea.

But given we have dissemination/disclosure among friends only, do we need to encrypt the blocks? Given communication channels are encrypted.

This is why Compuverse only collects name, business email and phone number.

There's 5 in this list and there was 4 in the other - I think Safety and Privacy is the one additional but it's also in my proposal because I am concerned about it too.

Millions of Patient Records at Risk: The Perils of Legacy Protocols

Sina Yazdanmehr | Senior IT Security Consultant, Aplite GmbH Ibrahim Akkulak | Senior IT Security Consultant, Aplite GmbH Date: Wednesday, December 6, 2023

Abstract

Currently, a concerning situation is unfolding online: a large amount of personal information and medical records belonging to patients is scattered across the internet. Our internet-wide research on DICOM, the decade-old standard protocol for medical imaging, has revealed a distressing fact – Many medical institutions have unintentionally made the private data and medical histories of millions of patients accessible to the vast realm of the internet.

Medical imaging encompasses a range of techniques such as X-Rays, CT scans, and MRIs, used to visualize internal body structures, with DICOM serving as the standard protocol for storing and transmitting these images. The security problems with DICOM are connected to using legacy protocols on the internet as industries strive to align with the transition towards Cloud-based solutions.

This talk will explain the security shortcomings of DICOM when it is exposed online and provide insights from our internet-wide research. We'll show how hackers can easily find, access, and exploit the exposed DICOM endpoints, extract all patients' data, and even alter medical records. Additionally, we'll explain how we were able to bypass DICOM security controls by gathering information from the statements provided by vendors and service providers regarding their adherence to DICOM standards.

We'll conclude by providing practical recommendations for medical institutions, healthcare providers, and medical engineers to mitigate these security issues and safeguard patients' data.

"If you aren't paying for it, you're the product."

"If you pay for it, you're paying for fundamental rights."

Harold Abelson, Ross Anderson, Steven M Bellovin, Josh Benaloh, Matt Blaze, Jon Callas, Whitfield Diffie, Susan Landau, Peter G Neumann, Ronald L Rivest, Jeffrey I Schiller, Bruce Schneier, Vanessa Teague, Carmela Troncoso, Bugs in our pockets: the risks of client-side scanning, Journal of Cybersecurity, Volume 10, Issue 1, 2024, tyad020, https://doi.org/10.1093/cybsec/tyad020

Abstract

Our increasing reliance on digital technology for personal, economic, and government affairs has made it essential to secure the communications and devices of private citizens, businesses, and governments. This has led to pervasive use of cryptography across society. Despite its evident advantages, law enforcement and national security agencies have argued that the spread of cryptography has hindered access to evidence and intelligence. Some in industry and government now advocate a new technology to access targeted data: client-side scanning (CSS). Instead of weakening encryption or providing law enforcement with backdoor keys to decrypt communications, CSS would enable on-device analysis of data in the clear. If targeted information were detected, its existence and, potentially, its source would be revealed to the agencies; otherwise, little or no information would leave the client device. Its proponents claim that CSS is a solution to the encryption versus public safety debate: it offers privacy—in the sense of unimpeded end-to-end encryption—and the ability to successfully investigate serious crime. In this paper, we argue that CSS neither guarantees efficacious crime prevention nor prevents surveillance. Indeed, the effect is the opposite. CSS by its nature creates serious security and privacy risks for all society, while the assistance it can provide for law enforcement is at best problematic. There are multiple ways in which CSS can fail, can be evaded, and can be abused.

Right off the bat, these authors are highly experienced and plugged into what is happening with technology.

How does that improve privacy?

"Without the right to tinker and explore, we risk becoming enslaved by technology; and the more we exercise the right to hack, the harder it will be to take that right away" - Andre "Bunnie" Huang

hah, we are already "enslaved by technology". ask Ted Kaczynski

our enemies already have hardware backdoors, compromising emissions (tempest), closed-source firmware/drivers/hardware, ... but sure, "feel free"

``` { version: 4,

info: { reason: <string>, // what triggered this ping: "saved-session", "environment-change", "shutdown", ... revision: <string>, // the Histograms.json revision timezoneOffset: <integer>, // time-zone offset from UTC, in minutes, for the current locale previousBuildId: <string>, // null if this is the first run, or the previous build ID is unknown

sessionId: <uuid>,  // random session id, shared by subsessions
subsessionId: <uuid>,  // random subsession id
previousSessionId: <uuid>, // session id of the previous session, null on first run.
previousSubsessionId: <uuid>, // subsession id of the previous subsession (even if it was in a different session),
                              // null on first run.

subsessionCounter: <unsigned integer>, // the running no. of this subsession since the start of the browser session
profileSubsessionCounter: <unsigned integer>, // the running no. of all subsessions for the whole profile life time

sessionStartDate: <ISO date>, // hourly precision, ISO date in local time
subsessionStartDate: <ISO date>, // hourly precision, ISO date in local time
sessionLength: <integer>, // the session length until now in seconds, monotonic
subsessionLength: <integer>, // the subsession length in seconds, monotonic

addons: <string>, // obsolete, use ``environment.addons``

},

processes: {...}, simpleMeasurements: {...},

// The following properties may all be null if we fail to collect them. histograms: {...}, keyedHistograms: {...}, chromeHangs: {...}, // removed in firefox 62 threadHangStats: [...], // obsolete in firefox 57, use the 'bhr' ping log: [...], // obsolete in firefox 61, use Event Telemetry or Scalars gc: {...}, fileIOReports: {...}, lateWrites: {...}, addonDetails: {...}, UIMeasurements: [...], // Android only slowSQL: {...}, slowSQLstartup: {...}, } ```

``` { type: <string>, // "main", "activation", "optout", "saved-session", ... id: <UUID>, // a UUID that identifies this ping creationDate: <ISO date>, // the date the ping was generated version: <number>, // the version of the ping format, currently 4

application: { architecture: <string>, // build architecture, e.g. x86 buildId: <string>, // "20141126041045" name: <string>, // "Firefox" version: <string>, // "35.0" displayVersion: <string>, // "35.0b3" vendor: <string>, // "Mozilla" platformVersion: <string>, // "35.0" xpcomAbi: <string>, // e.g. "x86-msvc" channel: <string>, // "beta" },

clientId: <UUID>, // optional environment: { ... }, // optional, not all pings contain the environment payload: { ... }, // the actual payload data for this ping type } ```

Disable all settings related to Privacy Sandbox

Disables "adMeasurementEnabled", "fledgeEnabled", "topicsEnabled" and "privacySandboxEnabled" for your browser.

The Topics API is worst than 3rd-parties cookies, anyone can query a user ad profile:

```js // document.browsingTopics() returns an array of BrowsingTopic objects. const topics = await document.browsingTopics();

// Get data for an ad creative. const response = await fetch('https://ads.example/get-creative', { method: 'POST', headers: { 'Content-Type': 'application/json', }, body: JSON.stringify(topics) });

// Get the JSON from the response. const creative = await response.json();

// Display the ad. (or not) ```

• for: quote, quote - Prateek Raj, quote - internet regulation, quote - reducing misinformation, fake news, indyweb - support
• quote
  • We lived in a relatively unregulated digital world until now.
  • It was great until the public realized that a few companies wield too much power today in our lives.
  • We will see significant changes in areas like
    • privacy,
    • data protection,
    • algorithm and
    • architecture design guidelines, and
    • platform accountability, etc.
  • which should reduce the pervasiveness of
    • misinformation,
    • hate and visceral content
  • over the internet.
  • These steps will also reduce the power wielded by digital giants.
  • Beyond these immediate effects, it is difficult to say if these social innovations will create a more participative and healthy society.
  • These broader effects are driven by deeper underlying factors, like
    • history,
    • diversity,
    • cohesiveness and
    • social capital, and also
    • political climate and
    • institutions.
  • In other words,
    • just as digital world is shaping the physical world,
    • physical world shapes our digital world as well.
• author: Prateek Raj
  • assistant professor in strategy, Indian Institute of Management, Bangalore

const topics = await document.browsingTopics();

chrome://topics-internals

A spec to optimize ad targeting (respectful of privacy, they say... 😂🤣).

Fuck you Google with your dystopian API:

```js // document.browsingTopics() returns an array of BrowsingTopic objects. const topics = await document.browsingTopics();

// Get data for an ad creative. const response = await fetch('https://ads.example/get-creative', { method: 'POST', headers: { 'Content-Type': 'application/json', }, body: JSON.stringify(topics) });

// Get the JSON from the response. const creative = await response.json();

// Display the ad. (or not) ```

And use Chrome only in case of necessity.

Naming a thing with a meaning opposite to what the named thing is...

Google is insatiable when it regards to accessing users private data. Let's block that bullshit.

Colonized groups such as these indigenous people have urgency to avoid colonization of their data and are doing something about it

看到人工智能在破解密码方面有多么强大，这很好地提醒了我们，不仅要确保你在使用强密码，还要检查：

• 你正在使用 2FA/MFA（尽可能不使用基于短信的）。

• 你没有在不同的账户间重复使用密码

• 尽可能使用自动生成的密码

• 定期更新密码，特别是敏感账户的密码

• 避免使用公共WiFi，尤其是银行和类似账户