Use something like Backblaze or Hetzner storage boxes for off-site backups. There are a number of tools for making this painless, so pick your favorite. If you have the means, I recommend doing a disaster recovery scenario every so often (i.e. disconnect existing drives, reinstall the OS, and load everything from remote backup).

Generally speaking, follow the 3-2-1 rule:

• 3 copies of everything on
• 2 different types of media with
• 1 copy off site (at least)

For your situation, this could be:

• 3 copies - your computer (NVMe?), TrueNas (HDD?), off-site backup; ideally have a third local device (second computer?)
• 2 media - NVMe and HDD
• 1 copy off site - Backblaze, Hetzner, etc

You could rent a cloud server, but it’ll be a lot more expensive vs just renting storage.

All of them? Maybe an international consortium that pays devs in their home currency.

I disagree that anything you describe could actually be both commercially viable and deployable without authoritarian involvement

You haven’t heard of Ring cameras? Commercial security systems? They do basically what I’m describing, just not as well because they don’t have as much of an incentive. Are end users willing to pay for these more advanced models? No, so consumer grade cameras stick to object detection like deer vs racoon instead of specific individual detection (e.g. scanning eyes).

Governments, however, are willing to pay that amount. Why? Because they think it’ll help them detect criminals, and they think that helps keep people safe. It’s an extension of the HOA idea, just with government-scale funds backed up with law enforcement to go after threats. That, in itself, isn’t authoritarian, but setting up such a system opens the door for authoritarians to take control and misuse it.

I’d go so far as to say that the people in your theoretical HOA are analogous to supporters of a authoritarian regime.

Analogous, sure, but the HOA has no enforcement arm for non-residents, so all they can do is ask the police to intervene. That’s the difference with a city, it has a police force it can order to intervene using information from that system. It’s the mixing of enforcement and surveillance that makes it authoritarian.

So a surveillance system is not itself authoritarian, it’s only authoritarian of there’s some enforcement arm to enforce obedience or punish disobedience.

If it is nearly impossible to meaningfully use apolitically, then it is not apolitical.

Again, I disagree. Something is only political when used for political ends.

Similarly, even if HOAs could deploy a system like that, that’d make them authoritarian.

That really depends how the system is used. If it explicitly doesn’t record regular residents and people who have signed up officially as visitors (and homeowners can review footage), I don’t think the camera system itself would really be authoritarian. Yeah, the system would be capable of violating privacy, but as long as the system is transparent and reviewable by the residents, I think it can be privacy-respecting. Basically, it would be like a home security system, but across a neighborhood, and it can even be self-hosted to not let third parties access the data (and police requests would go through the HOA board, which consists of residents).

That’s my point. If the system itself can be used in a privacy-respecting way (and the vast majority can), even if it’s typically not used that way, the system itself cannot be authoritarian. If an institution uses it in an authoritarian way, then the institution is authoritarian.

In short:

1. cameras are not authoritarian
2. databases are not authoritarian
3. license plate and face recognition software isn’t authoritarian
4. connecting 1-3 together in a searchable way isn’t authoritarian (would be a fun hobby project)
5. Sharing info from 4 isn’t authoritarian (again, could be a fun hobby with friends)
6. An institution (gov’t, business, HOA, etc) using 4 and/or 5 to enforce policy on citizens/employees/residents/etc is authoritarian

I have friends that use home cameras to do object classification as a hobby, mostly to identify and fee record wildlife. I’ve also heard of people doing this to identify package deliveries and catch package thiefs. Sharing those models with others on the internet is largely the same idea as what flock is doing, and with enough data, similar solutions to what Palantir is doing could be done entirely by hobbyists.

The products Flock and Palantir aren’t authoritarian in and of themselves, it becomes authoritarian when those products are used to enforce policy.

I would argue that such a product would be by its nature political, because it’s only practical use case was the furtherance of a political goal.

Again, I disagree. Surveillance has a lot of use cases outside of government, and a huge use case is keeping the government in check. Palantir could have sold its services to non-profits like the ACLU as a check on local, state, and law enforcement agencies. They could have sold it to HOAs and neighborhood watch associations as an early warning system for repeat offenders.

The government skirting the 4th amendment (and a few others) doesn’t automatically make its sub-contractor’s products “authoritarian,” it makes its use of those products authoritarian.

So a system that does so (like the ones sold to the govt) is a political software product.

I disagree with that conclusion. The use by the government is authoritarian, but that doesn’t make the product authoritarian.

To me where it gets tricky is when private entities grow to government-sized proportions, and begin to use these same tools for similar purposes

A private entity can do authoritarian things, like spying on its employees or customers. Authoritarianism isn’t strictly tied to governments, but anything that acts like a government. Here’s the first definition I found:

Characterized by or favoring absolute obedience to authority, as against individual freedom.

Software can’t really favor obedience to authority, it can’t really deny you your freedoms, it’s just software. Likewise for a camera system. The only way those things can be authoritarian is if paired with some form of enforcement arm, like corporate security or law enforcement. So that combined system is authoritarian, the cameras or software on their own cannot be authoritarian.

That’s my point.

Fortunately they use kph. Dodged a bullet there…

What is the actual likelihood of the following:

• China remotely disabling it
• hackers remotely disabling it
• company grifting you

IMO, that list goes from least likely to most likely, and this story is likely just FUD to reward some local manufacturer or other interest.

I’m trying to avoid using the term fascist, because it means something specific but nobody can really agree what that thing is. For the purposes of this discussion, I’d prefer to say “authoritarian”

It’s more that people probably know what it means, but choose to misuse it to smear their political enemies, and then other people who don’t know what it means repeat it.

Here’s a clear definition in case you or anyone else that reads this isn’t clear on it (or pick your favorite dictionary, it’ll be similar):

A system of government marked by centralization of authority under a dictator, a capitalist economy subject to stringent governmental controls, violent suppression of the opposition, and typically a policy of belligerent nationalism and racism.

Is a network of cameras with facial recognition fascist according to that definition? No. Is it useful to people pushing for such a government? Yes. Is it useful to other authoritarian systems of government? Yes. Is it useful to non-authoritarian systems of government and non-government entities, including private citizens? Also yes.

I wouldn’t call traffic cameras invasive because they’re only at (some) intersections.

What if they’re at every intersection, stop signs included?

If the only thing that turns something into an authoritarian system is scale, then it’s not the system that’s authoritarian, but the way they’re used that is authoritarian.

I oppose red light cameras not because they’re authoritarian in and of themselves, but because they can be used by authoritarians to screw people. I oppose Ring doorbells not because they’re authoritarian, but because the corporation has control and can hand that data over to authoritarians without consent from the owner (or be compelled by authoritarians).

“Authoritarian” is an adjective that describes people, governments, or policies, not inanimate objects or software systems.

A private citizen recording people in public and the government doing so are fundamentally different

Exactly! The capability to record the public isn’t authoritarian, the government policy of recording the public is authoritarian.

This may sound like a pedantic point, but I think it’s an important one. If cameras are authoritarian, then ban cameras and the problem goes away right? The government will just use radar, track financial transactions, or something else entirely, and you have the same problem.

The real problem isn’t cameras or facial recognition, but that the government tracks people. To solve that problem, we shouldn’t ban the various ways the government can track people, we should ban the government from tracking people. Don’t b regulate the tools, regulate the people using the tools.

What do you mean by “image”? As in, everything currently loaded into Python? Wouldn’t that just be execing the current architecture[0] again with the same ENV vars and args?

And that’s actually what I want much of the time, since I’d rather pay that cost on startup than throughout a running program’s life, esp. for heavy dependencies.

So a keyword is great here, since I can balance startup time and hitching during operation.

Why put the caveat that it’s not from your country? If it has remote access, it can be remotely disabled. Maybe it’s from a designed feature, or maybe it gets hacked.

I think that privacy invading surveillance systems are primarily useful for furthering authoritarian goals, by intention or not.

And I disagree. I think this all started when we allowed things like traffic light cameras, speed cameras, and toll cameras to automatically bill based on license plate. I don’t think most would consider those to be “primarily useful for furthering authoritarian goals,” they’re merely there for routine law enforcement with specific goals.

Flock cameras are basically that same exact system, but instead of only being used when something tangible is triggered (red light, radar, or toll booth motion sensor), they passively collect information. Flock is a private company that sells its surveillance services to cities (and private orgs) to assist with tracking down license plates or alerting when there’s a gunshot detection. This is allegedly legal because you don’t have any expectation of privacy when you’re in public (hence why Ring doorbells are legal), and private companies don’t have to follow the same rules as law enforcement. I personally don’t think Flock’s founders are fascist, they seem to genuinely want to help reduce crime. I worked for a similar company that mostly did perimeter security (i.e. generally only operated on private property), and the founder was absolutely not fascist, but they did want to help reduce crime.

I personally don’t consider either of those systems fascist by nature, but they can be used to achieve fascist goals. Tracking burglars across neighborhoods doesn’t sound especially fascist to me, but tracking protestors certainly does. These are very dangerous technologies that can easily be used for fascist purposes, so I think we shouldn’t allow them to be used at all, not because they are fascist, but because they can easily be used for fascist ends just by changing conventions around its use.

I don’t think we need to label a system as authoritarian or fascist to oppose them, we just need to point out how easily they can be misused.

The intention can be fascist, sure, but that doesn’t mean the solution is fascist.

For example, I think it’s pretty clear that Lemmy was designed by tankies to create a safe space for tankies (why would the instances the main devs maintain be overly protective of China and Russia if it weren’t?), but that doesn’t make Lemmy “tankie,” it’s a software project that can be used by fascists, tankies, commies, anarchists, statists, etc, because it’s just a software program.

Likewise, a surveillance system can be used by a fascist government, private company to protect company secrets, government agency like the Pentagon for internal use, or even private individuals to ID who is at the door. It’s only fascist of it’s used to further fascist goals, like identifying minorities or protestors. But then, it’s still not the software that’s fascist, but the whole system, meaning how people use it and the policies in place.

The chance of a given piece of software being “fascist” is incredibly low, since it would need to act in a fascist way and only a fascist way, or only be useful for fascist ends. Like the fascist LLM example I gave, or a training simulator that is hard-coded to only present fascist ideology.

I think those are important hairs to split.

Let’s say there’s a camera system built due to a direct public vote and rolled out by a political party all agree defends democracy. The stated goal is catching red light violations and speeders, and it’s a popular system. As part of the functionality it reads license plates, and that is verified by a human every time, and no footage is stored if there’s no violation.

Is that system fascist? Most would say no, and it exists in many states, like California and Washington.

Then the next election, a fascist is elected, and one of the first moves is to repurpose that system to track undesirables, and now it stores a ton of footage.

Is that system now fascist? It’s the same exact system as in the previous example, it’s just being used for fascist ends, such as tracking vehicles with certain plates (e.g. Illegal immigrants, minorities, etc) Nothing has changed in the capabilities or programming of the system, the only change was when to capture footage, what people use it for, and how long to store it.

Yes, it’s theoretically possible to design a fascist system, such as an LLM that only gives fascist answers, but that’s an incredibly narrow set of possible systems.

I haven’t used Immich, so no. I might get around to it sometime in the next couple months.

Yeah, ideally something like Immich is just some metadata pointing at existing files. An album should be some combination of synced directories (i.e. new files added outside Immich should be detected and added, if selected) and individual files.

That doesn’t sound that complex. Maybe I’ll look into Immich and see about adding that. I don’t currently use it, but I don’t have an image solution, so I’ll give it a shot.

Their argument was that software can, in itself, be fascist, and that’s what we went around and around on. The example given was facial recognition software that can determine race (and later, country of origin).

Essentially, I said exactly what you’re saying, while they argued the opposite. I wish I quoted them, but I did only directly address their claims, if you’ll take my word for it.

I don’t want the government to have and use facial recognition software (their example) and extensive security camera systems (my example, such as Flock), not because those solutions are fascist in and of themselves, but that they can be used by fascists to accomplish their goals. Even if the current regime uses them purely for good (i.e. completely opt in facial recognition, cameras inaccessible to police until there’s a warrant with no passive collection) the next regime may not.

If anyone is refusing to engage, it’s you. You provided no argument for your position, whereas I’ve explained as best I can in detail, with examples of similar things. Me not agreeing with you isn’t “refusing to engage,” it’s a good faith debate.

If there’s some point you’ve made that I’ve failed to address, I apologize, I tried to be thorough to not waste any time going back and forth.

Again, it’s not the software itself that’s fascist, it’s what it’s being used for that’s fascist. Facial recognition for determining citizenship could absolutely be used for non-fascist purposes, like simplifying border crossings to not require documentation (i.e. completely opt-in). Likewise, surveillance systems can also not be used until there’s an actual warrant (i.e. no passive recording), which can help in catching dangerous criminals.

The technology itself isn’t fascist, it’s how it’s applied that’s fascist. The mass data collection is fascist, the tools used to collect that data isn’t fascist in the same way that guns and tanks aren’t fascist, but they can certainly be used by fascists.

The fault lies with the makers and users of the softeware. Software doesn’t have political opinions, it’s software.

It’s like saying Panzer tanks were fascist because they were built by the Nazis. Tanks cannot be fascist, they’re tanks. So despite being made and used by fascists, they’re not fascist, they’re tanks.

That’s the same exact thing here. Facial recognition software can be used by fascists, but that doesn’t make the software itself fascist.