Wouldn’t it be more useful to link to: https://www.fsf.org/bulletin/2026/summer/ensuring-freedom-one-blob-at-a-time? Interesting project though.

Anonymity might actually be a reasonable fit here: if data couldn’t, in any way, shape or form and without measurable confidence, be attributed to a person (or context related to them: a car, or a specific shirt worn that day, being examples); then details about one’s private life couldn’t be inferred, based on publicly available information.

If for example: a surveillance system logged a person frequently moving from one house to another, while confident in the identity of the person departing from home, and a track record of the persons present in the house of destination on arrival, it can be inferred the person in question, has some relationship with the other persons.

So because this person has no choice, other than to be subjected to public surveillance, it is physically impossible to keep such details to oneself (private if you will). Similarly, if every other started wearing spywearables (“smart” glasses, pendant, pin, etc.) or (unknowingly) bugging their houses (a home surveillance system, robot vacuum, or any other “smart” appliance with listening devices); how are you going to keep anything private, and when is in no context “expectation of privacy” “reasonable”?

Got to keep shoveling that minimum viable product slop, but heaven forbid a label which might serve as an indication for it.

I think it’s clear I can’t convince you, you can’t convince me, and that the only thing left is to respectfully disagree.

We can keep falling over these minor differences, but as far as I’m concerned it’s a distinction without a difference, practically speaking. You seem to disagree, which is completely fine too. Have a good one! :)

And this is why you use No-AI DuckDuckGo

The discussion was about “effective privacy” in public: your every move not meticulously being monitored in it, and therefore effectively remaining unknown; despite technically falling under public “knowledge”. I understand how “privacy” is typically interpreted in context of legal cases and law, and how it doesn’t, or inadequately accounts for advances in technology: leading to formerly hidden public activities, possibly becoming uncovered as a result. If anything, I think we’ve clearly demonstrated the usefulness of a catch-all term, that generally doesn’t discern between context, in which personal data is being collected; which was ultimately the intent of this post.

Ah, I see. Then I’d argue that which applies to masking, similarly applies to tag obfuscation. And ideally the plate would remain readable in case of disputes (as would recognition of a face), but mass-surveillance systems have taken that nuance from the equation.

If overall collection of personal data isn’t a concern (which it is for me), and the focus is solely on sabotaging automated systems: clothing with confusing patterns may throw off classifications, IR glasses might impair facial recognition, and dirt-like noise atop license plates might hinder character recognition; but the irony is, that one likely draws more attention to themselves in doing so (an officer being alerted upon failure, and likely required to manually review: potentially leading to an actual investigation).

I don’t mind what most ordinary people know about me, unless they’d systematically collate those activities: especially for the purposes of exposing underlying patterns. If the persons involved or activities carried out weren’t memorable, they are likely to be forgotten. And generally speaking: observations by an untrained person are hardly reliable, and they certainly lack the resolution of digital counterparts.

But that lack of anonymity directly translates to a potential privacy-implication. You may also argue that nothing is private, because you could put a gun on a person’s head, and force them to give up any secret they have. Surveillance is equally involuntary, and similarly exposes details about one’s private life, which where traditionally very unlikely to be extracted.

Yes, exactly. If there’s no systematic record-keeping of my activities by actual human beings, I have little reason for concern. But with advances in technology: increased coverage, data quality, retention periods, automated detection/tagging/notifying/reporting(/actioning) and real-time or low-friction sharing through networking; it has become such an undeniable vector for (future) abuse. There’s nothing which can reasonably assure me, these systems aren’t going be abused, because there’s a track-record of abuse already, and the incentives for abuse are only gaining in strength.

Relatively speaking. These patterns were extremely unlikely to be uncovered (unless, again, you were a worthy enough target to allocate resources to); and therefore such information might as well be considered “effectively” private.

What you’re upset about it how much more easily people can connect those pieces of information and access them.

Yes. The pieces were either not available, or would be way too resource-intensive to piece together: simply to generate the type of insights outlined in your scenario.

The cat can’t be put back in the bag, what you gotta do is be one of those psychopaths who use tag obfuscation and wear face masks around.

A defeatist mindset can only result in defeat, while we’re still amidst the rollout. I’m not quite sure what you mean with “tag obfuscation”, but there’s nothing “psychopathic” about wearing masks, solely in an effort of protecting one’s personal information, from potentially being used in a malicious manner. What is psychopathic however, is subjecting innocent persons to that sort of tyranny, and victim-blaming them whenever they resist.

I feel the same way. But the question is rather: what would be the term that would describe the quality, of not being surveilled or harassed for that matter, which applies in any context (private or public)?

Just mandate an opt-in default. If the “3–10% of people [that] want this”, wish to get out of their way and explicitly enable spyware, I believe this is a much more reasonable overall baseline (with respect to the other 90–97%). And considering most web-users rarely clear cookies, the preferences are effectively persistent already: which really appears to be the lobby’s primary interest. If people were conditioned by dark patterns to accept, while holding onto cookies, excessive data collection happens without user-awareness and potential for reevaluation.

I’m no fan of browser signaling, because to me it seems like yet another unique identifier: to potentially, and ironically be (ab)used to improve browser-fingerprinting.

I generally don’t really mind clusters, although the clusters on Atlas Obscura don’t seem to dissolve and instead remain visible, which did throw me off initially. But I do strongly dislike some implementations, where (de)clustering happens very frantically, when switching between zoom-levels. However, rendering all points does looks quite clever, but effectively also creates clusters in high-density areas; which is ironic. Perhaps that in such areas you might as well cluster, to create some sort of middle-ground: with either a count or heat signature to more clearly discern density.

I don’t know, if OP is considering the machine not connecting to the internet, and getting files on and off by USB transfers, it might fit OP’s threat model. And if OP has an older device (on Intel, rather than Apple silicon), OP could keep using the same device.

You could perhaps have TailsOS on a USB drive (with persistent storage enabled on the drive: for your writings), and after use remove the drive from the device and keep it well hidden. That seems like the most attainable method as far as I’m aware.

Management has lost all moral compass”: Android’s head of security walks out on Google Military AI, surveillance of Europeans, ethical principles cast aside: why he walked away

By Aymeric Geoffre-Rouland Published on 06/12/26 at 3:33 p.m.

René Mayrhofer has been protecting the security of your Android smartphone for nine years. He has just resigned from Google for a reason that directly concerns you: the company has signed an agreement authorizing the Pentagon to use its AI for classified operations, and the man who secured your smartphone believes these tools will “likely be used against” European citizens. “Management has lost its moral compass”: Android’s head of security slams the door on Google

He is the man who protected the security of your Android phone. René Mayrhofer left Google after nine years, in a resignation letter published on his blog titled “Google Management Has Lost Its Moral Compass.

”The trigger: a contract signed in late April between Google and the U.S. Department of Defense authorizing the Pentagon to use the company’s AI models for classified work, including military operations planning and intelligence.

“My decision has become inevitable,” Mayrhofer writes. “I am a pacifist. Actively contributing to harming human beings is not something I can or will condone.”

From the 2018 open letter to the 2026 resignation

Mayrhofer’s career at Google has mirrored the company’s ethical principles. In 2018, following the internal revolt against Project Maven (a drone image analysis program for the Pentagon), Google had published clear commitments: no AI for weapons, no AI for surveillance. Mayrhofer had signed the employee petition that year. Google subsequently withdrew from the contract.

This deal implies that Google products will likely be used directly against me and my loved ones. In this context, I don’t see how I could not resign.

• René Mayrhofer, Director of Android Platform Security (in his resignation letter)

Seven years later, the shift is complete. In February 2025, Google quietly removed from its AI principles the section listing prohibited uses: weapons, surveillance, and technologies likely to cause harm.

The change, announced by Demis Hassabis (CEO of DeepMind) and James Manyika, was not communicated internally, according to Mayrhofer. “None of this is discussed or communicated within the company. Decisions are made at the top, period.” But you can still find the PDF here, on the Wayback Machine.

A European academic speaks out against a loophole-ridden clause

Mayrhofer is not just a Silicon Valley executive. As a full professor at Johannes Kepler University in Linz, Austria, he views the Google-Pentagon contract through a European lens. What alarms him is that the agreement authorizes the U.S. military to use Google’s AI for “any lawful purpose,” with no restrictions on scope. In other words, the only limit is what the U.S. government itself considers lawful.

However, according to Mayrhofer, this administration “has already violated international law on several occasions.” He cites a warning from KU Leuven University in Belgium, which advised its staff against collaborating with U.S. institutions. His fear is that this elastic definition could encompass the surveillance of European citizens.

In 2018, Google committed to never using AI for weapons or surveillance (page 4 of the original AI principles). In February 2025, this section was entirely removed. © Screenshot taken from the Internet Archive (Wayback Machine), version dated January 30, 2025, five days before Google removed this section

He is not alone. In April, Andreas Kirsch, a researcher at Google DeepMind, told Business Insider that he was “incredibly ashamed” of Google’s decision. But Mayrhofer is the most senior executive to have taken this step publicly.

“I am aware that, as a tenured academic in the EU, I am quite privileged. Many others do not have this freedom,” he acknowledges. His notice period runs through August. He announced that he would immediately step away from any work on AI systems that might fall under the military agreement. “I desperately hope that Google’s leadership will rediscover its moral compass,” he concludes.

In particular, yes. But people seeking to secretly record others are also creeps in my book.

So prying off the protective glass, drilling out the LED, and filling the void with resin, like this creep demonstrates. With electrical tape covering the LED, it seems creeps are required to block the camera sensor, for roughly one second after initiating video recording: as the photodiode value next to the LED is compared to brightness of the camera sensor; so by temporarily matching the two, Meta’s plausible deniability mechanism can be bypassed.

One memo reportedly described releasing it during a “dynamic political environment,” when privacy and civil liberties advocates would be distracted.

This line says it all…

That shouldn’t be acceptable either.