End of human integrity

Segfault

Segfault

Weaksauce
Joined
Oct 27, 2018
Messages
98
https://thedailybell.com/all-articles/news-analysis/privacy-obliteration/

Privacy? What's that. I'm not joking. I spoke with several friends about privacy and human integrity in general and they are totally confusing privacy with security. Google, Microsoft, Apple services are secure! Dude, read the user agreement and privacy statement of Google Inc. Short version: there is no privacy, everything you do using Google services belongs to Google and we can do with it what we want, store, analyze with AI, catlogize, sell it to our customers. People are brainwashed into believing security and privacy are the same thing.
NPU chips are the final blow. You may use Linux, but if the other end uses Microsoft, Apple or Google device to receive it then there is no privacy. Even now, if I send a message to a gmail address then it will be recorded in a folder which has my name on it.
This is the beginning of total surveillance, folks. Ever wondered what these 5400 data centers in the US are doing? Their energy consumption is already greater than 10% of US total. They keep building new ones. Think about it.
 
What if.

Yes, what if the government turns totalitarian and enacts a law which grants them full access to all large private databases like the ones Google is having. Google already has people cataloged into cohorts by their views. What a great material for a social credit system. And blackmailing people into obedience. And to force self censoring.
Funny thing, my mother who passed long time ago said a few times: computers will destroy the world. The rule of thumb is if a power can be abused then it will be abused. Computers are becoming a great tool to turn the free world into neo-feudal prison with a handful aristocrats and the rest of us will be peasants. Have you ever read from history books how feudalism did look like? I remember some from history class. Peasant owned nothing. He worked for landowner who owned everything, including the buildings and tools, just everything. Peasant was unable to escape because he had no legal means to make any money, leaving would mean either starving to death or turning into a criminal and being hunted. I'm not sure if they had 15-minute rule ... probably not, so they had more freedoms than we will have ...
 
I'm sorry, but this is the classic slippery slope argument.

On the path to total surveillance? Really? I'm not going to pretend that privacy is always as strong as claimed, but vague, evidence-free handwaving about "5400 data centers" is just scaremongering, not a reasoned position.

Be concerned about privacy. Use end-to-end encrypted communication when you can, deny app privileges when you don't think they're necessary, and avoid using services if you're uncomfortable about where the info goes. But don't panic, either.

I'm honestly reminded of Richard Stallman. He claims to be free because he uses only FOSS hardware, apps, and services, but in practice it means he's extremely limited and trapped. He hardly uses the internet and has to rely on ancient computers. If you take a tech ideology to an extreme, such as the insistence that all of Big Tech is out to spy on everything you do, you actually end up with less real-world freedom.
 
  • Like
Reactions: Epos7
like this
Or you could just self host, and give up nothing but time to set it up.
 
Mystique said:
Or you could just self host, and give up nothing but time to set it up.
Click to expand...
Ha! I wrote an article and I did not want to send it to a gmail address. So I put it into my router and sent a link to it instead. Just one friend got this link. Guess what, after two days I saw in my server log this article was accessed from Virginia – no, my friend wasn't there. Either gmail/Google leaked it or my friends device was under surveillance. We all know what's in Virginia, don't we. Tell me again I'm paranoid. There is a book, Only the Paranoid Will Survive. It is for business leaders, but the principles are the same for leading personal life.
 
Well, de-googling is basically "How to stop getting spied on 101".

I recommend doing it.

Microsoft, Apple, and Google are spying up to your eyeballs with crazy queries that don't need to happen since a sinkhole blocking them doesn't even affect the usage.

At some point, so too, are Samsung, and everyone else. We have never needed laws to block jailing phones, so that we can use open source phone OS more than we do today.

Some people will argue that "they'll still find a way, so trying at all is futile". I hard disagree with the 2nd part.
 
  • Like
Reactions: DPI
like this
OK, folks. Thank you for your attention. One final word. Do what you think is right. There are hired influencers in every major forum, their duty is to distract discussions which develop in a "wrong direction" and to inject politically correct viewpoints. That's a fact. Methinks Hardforum qualifies as a major forum ...
Imagine, a herd is driven into a slaughterhouse. One gets smart and wants to leave the column, an influencer comes along: stay, everyone is going, can't be bad!
Best wishes, I'm leaving this thread, chickening out. LOL (Or maybe not.)
 
Mystique said:
Well, de-googling is basically "How to stop getting spied on 101".

I recommend doing it.
Click to expand...
I was just leaving when I saw it. I'm de-googled. Was unable to access the web form to write to my senator. Imagine, .gov site and Google reCAPTCHA (which did not appear as I'm de-googled). So de facto Google already is part of our government. To resolve reCAPTCHA one must agree with Google terms and conditions. Third party, private party is controlling who can access a .gov site. And as we know, unless you take unusual measures, Google can identify with very high probability who is the user of browser. So basically you must give Google the right to register and archive your access.
Happy de-googling to everyone. :)
 
For something like text, not sure you need an NPU for a device to be able to do anything, the screenshot part make it harder than just reading text, if your device want to know what you type in as a text, taking screenshot and having ai reading text from jpeg seem a convoluted way to go about it, the tech for that already existed in the key pushed going throught the phone os....

That guy do not seem like much of an expert here:
To complete the surveillance structure for watching over your shoulder, the co-pilot AI, which is really just a private version of OpenAI chat GPT, or a large language model, is now preloaded on all Windows 11 machines. And apparently this co-pilot AI can scan your Windows Recall to see what has happened on your Windows timeline and use this data as part of its intelligence.

Except that co-pilot is in the cloud. It is an external server with new ability to analyze your own personal history in your machine. It is not entirely clear if the NPU can do some LLM functions like chat GPT locally on your device, or if it does not need to send data to the cloud

This seem abit all over the place, obviously no a copilot AI able to run on a laptop would not be at all a private version of the OpenAI chat GPT, do they have hunder 30B model (even 4o mini ?) and if it is loaded on your machine why is it in the cloud at that same time ?
 
jlbenedict said:
I think you would be shocked at what the government already has access to
Click to expand...

The government already buys data from various data brokers.

There are legal options on record that say that if it is for sale on the open market, they don't need warrants...

Shits already beyond fucked up, and had been for years. People either refuse to believe it is true or just don't care.

This is why we can't have nice things. And by nice things I mean privacy.
 
mullet said:
The sad truth, I have friends that I try and explain this to and they really don't care. They look at me like I am a paranoid nut job. I don't even try anymore.
Click to expand...

That's why it's so important for these hideous companies to market to kids, the frogs don't feel a thing if the water has always been hot.
 
Privacy:
Noun.
the useless state or condition of in-between observing advertisements or buying products.
 
Zarathustra[H] said:
The government already buys data from various data brokers.

There are legal options on record that say that if it is for sale on the open market, they don't need warrants...
Click to expand...
That's the non-shocking part, you and me can also buy that data (would make it strange to need a warrants for those)

Twisted Kidney said:
the frogs don't feel a thing if the water has always been hot.
Click to expand...
That was just a bad experiment and a bad comprehension of it, frog feel hot water a lot.
 
You can buy user data, but it tends to be scrubbed and filtered boiled down to some context specific statistics that you purchase in buckets based on specified demographics.
So it’s nothing specifically identifiable but useful for building add campaigns or market research.
Though if you get enough random data you can essentially piece it together but there would still be a margin of error large enough to give a jury reasonable doubt.

You need warrants for anything specific on an individual.
 
Lakados said:
You can buy user data, but it tends to be scrubbed and filtered boiled down to some context specific statistics that you purchase in buckets based on specified demographics.
So it’s nothing specifically identifiable but useful for building add campaigns or market research.
Though if you get enough random data you can essentially piece it together but there would still be a margin of error large enough to give a jury reasonable doubt.

You need warrants for anything specific on an individual.
Click to expand...

Not necessarily true. This is how most people think it works. Just high level barely identifiable data, but that just isnt the case. Many dfata brokers, and even regular companies we do business with on a daily basis sell or otherwise distribute the raw data.

Just in December last year there was a report about how major pharmacy chains have been sending patient records to police without warrants or consulting attorneys.

Then there is this whole 45 minute special on the topic. (When I heard this I was actually surprised that news media are finally paying attention to the big data/privacy issue, as it has gone mostly ignored in the past)
 
Zarathustra[H] said:
Then there is this whole 45 minute special on the topic. (When I heard this I was actually surprised that news media are finally paying attention to the big data/privacy issue, as it has gone mostly ignored in the past)
Click to expand...
They covered Snowden revelation in the past, The guardian and many others made a big work around it
https://www.theguardian.com/world/i...velations-decoded?ref=quillette.com#section/1

https://www.nytimes.com/2013/09/29/us/nsa-examines-social-networks-of-us-citizens.html
https://www.nytimes.com/2013/10/07/opinion/nsa-social-data-mining.html

JP-NSA-superJumbo.jpg


I remember being a big thing in the media since the 5 eyes listening and sharing every international phone (everything that used microwave link, public switch or satellite) call going throught them became a thing:
https://www.europarl.europa.eu/doceo/document/A-5-2001-0264_EN.html
Echelon:
https://en.wikipedia.org/wiki/ECHELON#:~:text=ECHELON, originally a secret government,known as the Five Eyes.

That made the news quite a bit in my country.
Back around then, we were convinced that we were much more spied on than we were (it was probably way more manual labor thus not as systematic as we thought)... and probably the other way around now.
 
Last edited:
LukeTbk said:
That was just a bad experiment and a bad comprehension of it, frog feel hot water a lot.
Click to expand...

Oh yeah, the experiments were absolutely typical 19th century pseudoscience.

The expression is absolutely apropos for almost everything modern industry does. Not just tech.
 
Twisted Kidney said:
The expression is absolutely apropos for almost everything modern industry does. Not just tech.
Click to expand...
And I am not sure if it is really a slowly boiling affair or jump.

An human government employee physically spying, manually connecting cable on your phone line and listening from a speaker, do feel completely different than a copy of an email read by a machine and it is in many way obviously, but that a step function of difference, not a slow thing, the moment tech make it easy-friction less, the moment it change the morality of it... it seems.

Parents putting a gps tracker on their childs in the 90s feel different than a smartphone with that function, but was it a slow change, or the very moment that function became easy on the already on you device that as non tracking purpose.

Government employee that take a quick look at a celebrity tax report from a computer probably feel quite different than manually going through file in a locker and would not surprise me that it was really quick, not a slow process.
 
LukeTbk said:
Back around then, we were convinced that we were much more spied on than we were (it was probably way more manual labor thus not systematic at all than we thought)... and probably the other way around now.
Click to expand...

Yeah, as far as I am concerned, the only thing holding them back in the past was the large amount of non-text data that could not be searched for keywords, and thus required manual review from people, which is prohibitively expensive on a "whole of the internet" scale.

Now with AI that limitation is gone, and there is nothing to stop the surveillance state unless laws change, and no one seems to be taking it seriously.
 
Mystique said:
Well, de-googling is basically "How to stop getting spied on 101". I recommend doing it.
Click to expand...

I dumped Google Workspaces a few months ago after being locked out of GDocs in a critical moment. The cause was my own oversight, but was the last straw and the motivation I needed to rip off the band-aid. When I experienced what an uphill, focused effort Google deliberately makes migrating GSuite accounts, it only confirmed the choice. Most people will be served well enough with Takeout however (https://takeout.google.com/).

I've also migrated several friends/family from MSOffice to Libre, and Outlook to BetterBird (Mozilla Thunderbird recompiled without telemetry) because I got tired of support calls on intermittent "Activation required" false errors on Office every so often.

And FFS, anyone that uses VSCode should uninstall and use VSCodium instead since it's the identical codebase just recompiled without telemetry.
 
Last edited:
At some point this can become just being blindly against any telemetry, what vscode extension I use and if there some loading issue, what are my layout preference in vscode and so on.... (all anonymised, not that it is "private" stuff) as they made the next version more matching what users prefer.... that seem quite acceptable. A bit like a video game console phoning home if a game crash.... with a crash report that perfectly fine.

telemetry.enable to false in your settings would do the same.
 
LukeTbk said:
At some point this can become just being blindly against any telemetry,
Click to expand...

As if that would be a bad thing were it someone's only motivating factor. Respectfully, it's odd that it matters to you enough to frame it as 'blindly against' as though there were some upside being overlooked or sacrifice being made from a user's point of view. But this is the internet. Say you're against credit card junk fees, breast cancer, or mold in your walls, and someone will inevitably be along to tell you you're throwing the baby out with the bathwater. Obviously those things aren't the same as telemetry, the point is telemetry is as insidious and incremental in its own way - they thrive in an oversight vacuum.

In the case of VSCode, no it's not a big deal. But it becomes offputting in the context of MS's aggressiveness with data collection overall; their disrespect for user choices when updates reverse settings back to "On" and they call it a bug; bad faith UI design; collecting data on programs and apps that don't belong to them, etc.
 
Last edited:
Segfault said:
https://thedailybell.com/all-articles/news-analysis/privacy-obliteration/

Privacy? What's that. I'm not joking. I spoke with several friends about privacy and human integrity in general and they are totally confusing privacy with security. Google, Microsoft, Apple services are secure! Dude, read the user agreement and privacy statement of Google Inc. Short version: there is no privacy, everything you do using Google services belongs to Google and we can do with it what we want, store, analyze with AI, catlogize, sell it to our customers. People are brainwashed into believing security and privacy are the same thing.
NPU chips are the final blow. You may use Linux, but if the other end uses Microsoft, Apple or Google device to receive it then there is no privacy. Even now, if I send a message to a gmail address then it will be recorded in a folder which has my name on it.
This is the beginning of total surveillance, folks. Ever wondered what these 5400 data centers in the US are doing? Their energy consumption is already greater than 10% of US total. They keep building new ones. Think about it.
Click to expand...
All your base are belong to us. :sick:
 
DPI said:
As if that would be a bad thing were it someone's only motivating factor. Respectfully, it's odd that it matters to you enough to frame it as 'blindly against' as though there were some upside being overlooked or sacrifice being made from a user's point of view.
Click to expand...
I am not sure I did imply that it would be any cost for an user stopping is telemetry. Obviously if you are the only one not phoning home info you will still have all the benefit of telemetry.

And yes it is possible to be blindly against what it is the possible issue to have with VS code telemetry here ? Why recommend people to use a different build because of it ? Not only it is to shut down in the option, but what could be the possible issue one would have with it ?

Nothing wrong about being blindly against all form of non-manual telemetry that you cannot see in plain text and press send before it happen by the way..(even if it is open source, anonyme, with nothing personnal-private-that matter in any way with no possible downside for you like it seem to be the case with the visual code telemetry). not necessarily a bad rule.
 
Last edited:
This has been going on in the USA since the Patriot Act in the early 2000s, but this specific level of surveillance has been a thing since at least 2014.
The NPU itself just made localized data sifting much more efficient on mobile (and now desktop) devices compared to GPUs and CPUs without the need for infrastructure in between.

This is why the USA government wants TikTok banned, not because it is less secure than Instagram, but because it is a different nation-state that has total control and back-end access to it.
Otherwise they are the exact same.

Segfault said:
What if.

Yes, what if the government turns totalitarian and enacts a law which grants them full access to all large private databases like the ones Google is having.
Click to expand...
Every megacorp hosting big data has been back-ended by their own respective nation-states for years - public knowledge of PRISM is basically old hat at this point.
Hell, the French corporation Telegram just had their CEO, Pavel Durov, arrested because he refused to give the French government back-end access to their data.

To quote the Sega Genesis/Mega Drive cyberpunk game Paprium (2020):
Rule, be ruled, or die. :borg:

pendragon1 said:
fuckitty fuck fuck before soapbox
Click to expand...
Your soapbox or mine? ;)
Either way, this discussion is about to get [H]ard.
 
Last edited:
LukeTbk said:
At some point this can become just being blindly against any telemetry, what vscode extension I use and if there some loading issue, what are my layout preference in vscode and so on.... (all anonymised, not that it is "private" stuff) as they made the next version more matching what users prefer.... that seem quite acceptable. A bit like a video game console phoning home if a game crash.... with a crash report that perfectly fine.

telemetry.enable to false in your settings would do the same.
Click to expand...

I think that if there were regulations against misuse of data and requirements that collected data be stored in ways that utilizes best practices to prevent it from being lost or stolen, then a whole lot of people would have lesser concerns with telemetry, but as it is there is an assumption that if it is collected, it will some way stores improperly and either directly misused against you or not protected and kept from those who would misuse it against you.
 
Zarathustra[H] said:
I think that if there were regulations against misuse of data and requirements that collected data be stored in ways that utilizes best practices to prevent it from being lost or stolen, then a whole lot of people would have lesser concerns with telemetry, but as it is there is an assumption that if it is collected, it will some way stores improperly and either directly misused against you or not protected and kept from those who would misuse it against you.
Click to expand...
How could this data possibly be used against customers/citizens? :whistle:

20th Century: https://en.wikipedia.org/wiki/Democide
21st Century: https://en.wikipedia.org/wiki/Transhumanism
 
Zarathustra[H] said:
I think that if there were regulations against misuse of data and requirements that collected data be stored in ways that utilizes best practices to prevent it from being lost or stolen, then a whole lot of people would have lesser concerns with telemetry, but as it is there is an assumption that if it is collected, it will some way stores improperly and either directly misused against you or not protected and kept from those who would misuse it against you.
Click to expand...
Even here, say Microsoft does not store correctly what extension crashed or time they took to load correctly (extension their user downloaded from the vscode extension marketplace, it is not like it is not already well known which you use without it), layout ? It is possible for some info to not matter and impossible to be used against you.

Take the game crashing on a console sending back a crash report, if it was openSource to know exactly what goes it like for VSCode, it would be possible to not be an issue (specially for people that bought their game on the console marketplace, it is not like they do not already tell microsoft-sony what game they play), if it is not opensource and the phone home encrypted, then obviously you cannot know what it is in it and that not a bad thing to be blindly against, it could be phoning stuff you do not want them to.

Again perfectly fine to say, I cannot see how it could ever matter for anyone to know that use very popular vscode extension, how long they took to boot and my default layout, but maybe there is a way, so I will be blindly against it.
 
Last edited:
There are some significant credibility and bias issues with those sources, pushing conspiratorial narratives and coming with a lot of questionable baggage to the point that it undermines any reasonable points about digital privacy that it could make. There have been myriad discussions regarding these topics without that sort of encumbrance, and a good example comes in the Substack post referencing the video railing against NPU chips. This lacks nuance and suggests that every NPU acellerator is a black box designed to spy on you and will undermine e2e encryption, but it just doesn't work like that . You could make the same argument more or less about the presence of a GPU - an acellerator for certain kinds of workloads, and that your GPU (and its proprietary drivers, for most OSes and manufacturers!) is secretly sending everything your screen renders to the Men In Black! The entire transcript of the video is, to say the least, a massive ediface built on dangerous misconceptions yet upheld by a couple grains of truth, but that is what makes it all the more dangerous. I don't really want to go through it all one by one, but lets not forget that everything from Secure Boot, TPM 2.0 to Pluton, to the Intel ME and AMD PSP were all evidence of some great machination of the surveillance behind the curtain... but it didn't turn out that way. Are there reasons to be concerned with the usage of all those things? Absolutely, but the specifics of where, what, and how matter - even if they're not quite as salacious as the most outlandish of hypotheticals. The real concerns with these usages are obfuscated by conspiratoral, half-baked nonsense and make it easy for those with malign intentions to discount those calling for legitimate change and better policies, technical or otherwise.

The NPU issue I feel is particularly timely in that one of the major issues currently with the development and usage of AI/LLM is if the models and training data will be proprietary, who and under what circumstances will those be deigned legitimate to run them, and will they be run locally or will actual computation take place remotely out of control. This last bit issue - ChatGPT and DALL-E from OpenAI, as well as Google's rebranded Gemini as a competitor and many others are proprietary models that require connecting to OpenAI, Google etc..or other servers where the computation takes place. Compare this to some of the only Free and Open Source AI models such as Stable Diffusion or LLaMA that can be run completely locally, trained, modified, updated etc.. on your own hardware or that of another user; there are myriad instances Stable Diffusion all with different rules, parameters, and nothng stops you from running your own be it for private use or exposing it publically on a website you own! When it comes to running an instance of one of these models, or certainly training , there's a relatively significantl hardware requirement, even despite optimizations. Having NPU functionality just means greater processing power for specific types of workloads - much like how in the old days CPUs had separate ALUs , or GPUs, or how more recently during the cryptocurrency mining boom people would seek out ASICs, FPGAs, or at least GPUs that supported the kind of workloads necessary to give them a good hashrate. Hell, a great deal of AI/LLM workloads today benefit from GPUs that have tensor cores and using tech like CUDA or ROCm interface with specified parts of the hardware that are good at the sort of calculations necessary for these workloads.

NPUs give the potential for more AI workloads to be run LOCALLY, which is something that we should be advocating for, not fearing. The ability to not have to send your query (and other information) off to a proprietary remote model means the potential for more equitable and private use of AI. If you'll note not that long ago, AI-focused megacorps like the ironically named OpenAI had their spokespeople saying "Oh dear you see its not that want to corner the market on high performance AI, wwe simply are the only ones capable of running these sophisticated models on our millions of dollars worth of hardware - it can't be handled locally!". That wasn't their only argument of course, but in any case having NPU acellerators in your hardware should be a benefit to privacy and security. Now, it absolutely matters how they're used, but that is the same for any computing resource, but that's where we should be focusing as opposed to the quagmire present in what was linked in the transcript and the many other problems involved. If the benefits of capable, sophisticated AI are going to be widely spread, they need to be open source models and training data formats, capable of being run locally and independently verified. Having NPU-focused components either in a SoC or in a discrete card could be a benefit for this. While its also possible that they could be used by some proprietary application as a way to gain distributed computing power for their own AI workloads, much like avoiding untended crypto miners, trojans, and other sorts of malware or resource and data draining adware, it has more to do with how they're used.

There are a lot of changes we need to make to restore the promises of the Internet many of us envisioned in the 90s/Y2K , valuing security, privacy, anonymity and the general "hacker ethic" as a guiding principle. Many of the issues we have come down to privacy laws that are more than 20 years out of date such as overning what information is considered public and what is private. However, its not going to be easy to do any of this if it gets wrapped up into unsupported or other questionable narratives that are more aligned with culture war objectives rather than actually making change in what is being discussed.
 
  • Like
Reactions: Epos7
like this
You must log in or register to reply here.
Back
Top