Luckily almost all modern corporate tracking is done through javascript execution + cookies. The days of parsing actual webserver logs are over for the most part. After all, it's only the browsers that execute javascript code and provide profitable personal information about the human behind the browser that matter. People with JS off are not providing sellable information and therefore classified and treated as if they were bots.
Turning off JS by default and temp-whitelisting only mitigates most of this tracking.
The issue is, even with all the browser protections, you still create an account anywhere or buy something an input your name/email address/shipping address, your "hashed data" immediately gets sent to meta/google as a conversion with "this guy bought a cat toy", and you start getting ads for cat related stuff everywhere.
They don't even need to "track" you properly for this stuff to work and it seems there's no way to escape it.
I don't experience that though I have friends who use smartphones who describe it. So I think a lot of it is via javascript. I doubt every retailer, or even a significant fraction, has their backend sending that type of data to $megacorp. But maybe I'm just lucky or shop weird places or it's because I use a new email address @superkuh.com for every account sign up. Or maybe I'm just not seeing the targeted ads for my $superkuhprofile that do exist because I have almost all ads successfully blocked. Perfect is the enemy of good anyway, all mitigations help a bit. And blocking JS is a huge mitigation.
If those companies are using big SaaS companies for eCommerce and have not going "Don't Track" part of their admin panel to turn off tracking, a lot of those SaaS companies will just sell off the data.
So sure, cat toy small time retailer on Etsy won't but credit card processor or shipper might.
I think part of the issue is that these retailers are also customers of meta/google on the side of purchasing ads, and as a merchant you're highly encouraged to send as much data on your events as you can, or your conversion tracking can be "less accurate"and your campaigns are less efficient.
So it's less about "we're sending the data to $megacorp" and more about "I want the most bang for buck on my own campaigns" when the decision is made.
Using a different email certainly helps, though!
EDIT: highly encouraged by meta et. al! Whether this is a legitimate request to improve results or pure self-interest on the part of meta I don't know!
It really does. But also, having to do this points out a glaring flaw in the design of the fediverse websites. They're applications and not documents. They require executing complex code from unknown third parties just to show a bit of text and some multi-media. This isn't needed at all. And it wasn't like this till mastodon v3 when they broke it.
Despite requiring Javascript execution mastodon actually does have the post contents of a URL in the hidden meta-content HTML header on the page where it scolds you and blocks you for not executing their arbitrary code. All they'd have to do is put that same text in the HTML as actual <p> text. And it's not just mastodon instances, the other fediverse "applications' are just as silly in their intentional breaking of accessibility for no reason.
I have yet to find a social network which is actually accessible. The Google thing (circles?) was never actually useable, it was the biggest horror show of all. m.facebook.com was basically the only website that was ever really accessible. All the other players, including the "free and morally superior" alternatives couldn't give two fucks about people with disabilities, which reflects nicely on the fact that they are actually not an alternative, they are a playground for misguided developers...
Fact is, if you are launching a social network which is not accessible from the get go, you are part of the problem. You have no moral high ground, you're just playing around and leaving people behind.
These kind of write-ups all have an implicit premise that is unstated: they're talking about corporate AI run by corporations. They're not actually talking about the technology. Corporate AI will never be ethical or safe because corporate persons have different motivations and profit incentives driving them than human persons do. And most of the time they're quite nasty when viewed through the lens of human ethics.
It reminds me of the parable of the blind monks each feeling a different part of the elephant and arguing about it's shape. They're each not wrong, but they're also only talking about a limited subset of the elephant (AI).
Cory Doctorow is much more eloquent in his explaination of this important distinction in his reverse centaur metaphor.
AMD hasn't signaled in behavior or words that they're going to actually support ROCm on $specificdevice for more than 4-5 years after release. Sometimes it's as little as the high 3.x years for shrinks like the consumer AMD RX 580. And often the ROCm support for consumer devices isn't out until a year after release, further cutting into that window.
Meanwhile nvidia just dropped CUDA/driver support for 1xxx series cards from their most recent drivers this year.
Last year, AMD ran a GitHub poll for ROCm complaints and received more than 1,000 responses. Many were around supporting older hardware, which is today supported either by AMD or by the community, and one year on, all 1,000 complaints have been addressed, Elangovan said. AMD has a team going through GitHub complaints, but Elangovan continues to encourage developers to reach out on X where he’s always happy to listen.
Seems like they're making some effort in that direction at least. If you have specific concerns, maybe try hitting up Anush Elangovan on Twitter?
Is it really that short? This support matrix shows ROCm 7.2.1 supporting quite old generations of GPUs, going back at least five or six years. I consider longevity important, too, but if they're actively supporting stuff released in 2020 (CDNA), I can't fault them too much. With open drivers on Linux, where all the real AI work is happening, I feel like this is a better longevity story than nvidia...where you're dependent on nvidia for kernel drivers in addition to CUDA.
You missed the note at the top "GPUs listed in the following table support compute workloads (no display information or graphics)". It doesn't mean that all CDNA or RDNA2 cards are supported.
That table is very is very misleading it's for enterprise compute cards only - AMD Instinct and AMD Radeon Pro series.
For actual consumer GPUs list is much worse https://rocm.docs.amd.com/projects/radeon-ryzen/en/latest/in... , more or less 9000 and select 7000 series. Not even all of the 7000 series.
I think that speaks to them not understanding at the time the opportunity they were missing out on by not shipping a CUDA-like thing to everyone, including consumer tech. The question is what'll it look like in a few years now that they do understand AI is the biggest part of the GPU industry.
I suspect, given AMDs relative openness vs. nvidia, even consumer-level stuff released today will end up with a longer useful life than current nvidia stuff.
I could be wrong, of course. I've taken the gamble...the last nvidia GPU I bought was a 3070 several years ago. Everything recent has been AMD. It's half the price for nearly competitive performance and VRAM. If that bet turns out wrong, I'll just upgrade a little sooner and still probably end up ahead. But, I think/hope openness will win.
Also, nvidia graphics drivers on Linux are a pain in the ass that I didn't want to keep dealing with. I decided it wasn't worth the hassle, even if they're better on some metrics. I've been able to run everything I've tried on an AMD Strix Halo and an old Radeon Pro V620 (not great, but cheap, compared to other 32GB GPUs and still supported by current ROCm).
ROCm is open source and TheRock is community maintained, and in a minute the first Linux distro will have native in-tree builds. It will be supported for the foreseeable future due to AMDs open development approach.
It is Nvidia that has the track record of closed drivers and insisting on doing all software dev without community improvements to expected results.
> And the worst privacy, transparency, and FOSS integration due to their insistence on a heavily proprietary stack.
The market doesn't care about any of that. The consumer market doesn't care, and the commercial market definitely does not. The consumer market wants the most Fortnite frames per second per dollar. The commercial market cares about how much compute they can do per watt, per slot.
> there exist plenty like me that demand high freedom and transparency and will pay double for it if we have to.
The four percent share of the datacenter market and five percent of the desktop GPU market say (very strongly) otherwise.
I have a 100% AMD system in front of me so I'm hardly an NVIDIA fanboy, but you thinking you represent the market is pretty nuts.
I did not claim to represent the market as a whole, but I feel I likely represent a significant enough segment of it that AMD is going to be just fine.
I think local power efficient LLMs are going to make those datacenter numbers less relevant in the long run.
I was thinking to get 2x r9700 for a home workstation (mostly inference). It is much cheaper than a similar nvidia build. But still not sure if good value or more trouble.
I own a single R9700 for the same reason you mentioned, looking into getting a second one. Was a lot of fiddling to get working on arch but RDNA4 and ROCm have come a long way. Every once in a while arch package updates break things but that’s not exclusive to ROCm.
LLM’s run great on it, it’s happily running gemma4 31b at the moment and I’m quite impressed. For the amount of VRAM you get it’s hard to beat, apart from the Intel cards maybe. But the driver support doesn’t seem to be that great there either.
Had some trouble with running comfyui, but it’s not my main use case, so I did not spent a lot of time figuring that out yet
I have a dual R9700 machine, with both cards on PCIe gen4 x8 slots. The 256bit GDDR6 memory bandwidth is the main limiting factor and makes dense models above 9b fairly slow.
The model that is currently loaded full time for all workloads on this machine is Unsloth's Q3_K_M quant of Qwen 3.5 122b, which has 10b active parameters. With almost no context usage it will generate 59 tok/sec. At 10,000 input tokens it will prefill at about 1500 tok/sec and generate at 51 tok/sec. At 110,000 input tokens it will prefill at about 950 tok/sec and generate at 30 tok/sec.
Smaller MoE models with 3b active will push 70 tok/sec at 10,000 context. Dense models like Qwen 3.5 27b and Devstral Small 2 at 24b will only generate at around 13 - 15 tok/sec with 10,000 context.
This is all on llama.cpp with the Vulkan backend. I didn't get to far in testing / using anything that requires ROCm because there is an outstanding ROCm bug where the GPU clock stays at 100% (and drawing like 60 watts) even when the model is not processing anything. The issue is now closed but multiple commenters indicate it is still a problem. Using the Vulkan backend my per-card idle draw is between 1 and 2 watts with the display outputs shut down and no kernel frame buffer.
Talking to friends who have fought more homelab battles than I ever will, my sense is that (1) AMD has done a better job with RDNA4 than the past generations, and (2) it seems very workload-dependent whether AMD consumer gear is "good value", "more trouble", or both at the same time.
Edit: I misread the "2x r9700" as "2 rx9700" which differs from the topic of this comment (about RNDA4 consumer SKUs). I'll keep my comment up, but anyone looking to get Radeon PRO cards can (should?) disregard.
I have 2 of them. I would advise against if you want to run things like vllm. I have had the cards for months and I still have not been able to create a uv env with trl and vllm. For vllm, it’s works fine in docker for some models. With one gpu, gpt-oss 20b decoding at a cumulative 600-800tps with 32 concurrent requests depending on context length but I was getting trash performance out of qwen3.5 and Gemma4
If I were to do it again, I’d probably just get a dgx spark. I don’t think it’s been worth the hassle.
Automatic registration means young adults will not have the consciously confront the possibility. This will certainly decrease the number of people establishing the paper trail that they are contentious objectors.
The overwhelming majority of registrations are automatic through drivers licensing. It's well proven that a significant portion of men who are registered don't even know there is selective service
There are plenty of local LLMs out there run by humans that play nice. It's not the LLMs that are the problem. It's the corporations. That's the commonality. Human people aren't doing this. These corporate legal persons are a much more dangerous and capable form of non-human intelligence with non-human motives than LLMs (which are not doing the scraping or even calling the tools which are sending the HTTP requests). And they have lobbied their way to legal immunity to most of their crimes.
Imagine if this was, say, at a bar. A person holds up a sheet of paper on a clipboard and says anyone can write about anyone else in the bar. Would the police be called? Would legal threats be made? People freaking out like this is not consistent with social norms. It's magical thinking about the internet somehow being different or unique.
Good luck dealing with these small minded administrators.
Tell me you never been to a bar without telling me you never been to a bar. Bars are usually a huge hot or not (well, parties in general). People are talking and gossiping about each other the whole time. At worst you only be talked about as well.
Okay. Now imagine he's doing it out on the public sidewalk and not using bar property. Would it be justified to use physical force and steal the clipboard owner's property?
Also the guy with a clipboard is only showing those notes to a couple people at a time after they journey to his location, which makes a pretty big difference for the amount of disruption. There's no magical thinking about the internet being special. The social norms are different because the situation is so different.
They stole his phone using physical violence. It's in the article. I'll quote it for you.
>So i was like chill out bro, ill just delete the video, but the dean said, "no confiscate his phone".
so one of the guards just snatched the phone from my hand
the dean said wipe everything, and dont give him his phone back. I was like wtf is happening? bro I have my private photos on there dont do this. like you cannot do this. I tried reaching for my phone but one of the security guys just held me. and started being rough with me. like pushing me around and shit.
That wasn't really because of the website though. The meeting could have been about basically anything and grabbing the phone for recording could have happened. It was bad but let's not mix up the two situations.
Bars probably don't have a directory of everyone that attends the bar, that you scraped and published without permission.
The fact that a person is a student at the school can be very sensitive information. The classic example is someone who leaves an abusive spouse/family and does not want to be found. Now their name and picture is out there, and their timetable and therefore whereabouts could be partially inferred from the school calendar by someone who knows their interests.
> The fact that a person is a student at the school can be very sensitive information.
But they were already in the directory? That's much more "out there" than the gossip site.
I'm really skeptical of this line of logic. It feels like motivated reasoning based on not liking the site, because a privacy issue like that is easier to attack (if it's real). I think the meaningful criticisms are based on the actual functionality, the commenting.
I understand that opinion, but the opposite view is now conventional. Corporate/college directories are usually not available in public, but only with a local auth. Even if the scraping site restricted signups to local email addresses, the college is responsible for the distribution of its directory PII so could not allow this.
Leaking PII like this would be illegal in Canada for example.
I don't understand how the word "leaking" would apply here. Unless there was an unmentioned login wall for the directory he scraped, the site is mirroring the names and faces off of a much larger and already public site that nobody has said a single word in complaint of.
Yeah, but he also took students' PII and put it on his website. If someone did that in a bar, there's a good chance they'd get their face punched by others in the bar.
Yep. I still develop Gtk2 applications today. It's a very snappy and low resource usage toolkit aimed entirely at desktop computers. None of that "mobile" convergence. I suppose you could put Gtk2 applications into containers of some sort but since Gtk2 has (luckily) been left alone by GNOME for decades it's a stable target (like NES or N64 is a stable target) and there's no need for it.
Most of the bloat these days is from containers and Canonical's approach to Ubuntu since ~2014 has been very heavy on using upstream containers so they don't have to actually support their software ecosystem themselves. This has lead to severe bloat and bad graphical theming and file system access.
Sure, one is connmapperl. It is a server/client application where the server is a GUI map of the world that shows all the various clients collected IP established connections via geoip lookup (local). It stores everything in sqlite db and has a bunch of config/filtering options; http://superkuh.com/connmapperl.html Technically a fork of X11 connmap I made because I coulnd't get it to run on my old X11, but with many, many more features (like offline whois from raw RIR dumps, the db, the hilbert mapping, the replays of connection history, etc).
Another one is memgaze, a program to vizualize linux process virtual memory spaces as RGB images and explore them using various binary visualization and sonification tools. Ie, you can just click a hilbert map of all processes then in the new window click around inside the image of that particular process' virtual ram and then listen to it interpreted as an 8bit wav, or find an extract images, for example. Or search for strings, run digraph analysis, etc. http://superkuh.com/memgaze-page.html
Or feeed.pl, my very quick and low resource usage feed reader for 1000+ feeds written in Perl/Gtk2 that is text only (no html, no images, etc). It is really handy for loading .opml files and finding and fixing broken feeds using the heuteristics I hard coded in to find feed urls. http://superkuh.com/blog/2025-09-13-2.html
These are a few I made 2025-26 that other people might care to use. But I have a lot more that just scratch my own particular itches. Like a Perl/Gtk2 version of MS Paint that interprets arbitrary loaded and painted images as sound, or the things that I use to monitor my ISP uptime/speed, etc.
>I'll put it like this - what you're doing with that comment above is a lot like blaming smokers for feeding the tobacco companies. Despite all the lies and ads and manipulation, despite all the dirty tricks, despite the hard-core science used to get people hooked from every possible angle.
I have never used Facebook and I never will. What they have done is immoral and unethical and deserves regulation.
What I fear is that regulation will be informed from the false and dangerous equivalence you've made there comparing addictive drugs to looking at an audio-visual screen. Drugs literally can make you want without there being any enjoyment. Screens are just a medium, like, a radio (which can also be used for random internal operant conditioning), the screens and the radio are not the problem and they are NOT LIKE DRUGS. You actually have to enjoy the experience and repeat it. And that's just normal learning. That drug comparison will lead to government's treating computers' like drugs which means heavy regulation of end users and violence against them. A far more dangerous scenario than the issues were facing from the corporations now.
We need regulation of the corporations intentionally doing random interval operant conditioning. Not regulation of the medium they do it over and the people enjoying using that medium.
> What I fear is that regulation will be informed from the false and dangerous equivalence you've made there comparing addictive drugs to looking at an audio-visual screen.
Let's be extremely clear - I'm not the one who first made that comparison. That would be the tech bros, who hire all manner of addiction and gambling specialists and scientists in order to make their products as addictive as possible.
> the screens and the radio are not the problem and they are NOT LIKE DRUGS
For a fully competent adult, you can make that argument. Kinda.
To an unsupervised 9yo? An 89 yo? Facebook is a lot like drugs, only with the mind-altering effects much easier to direct. No, that's not the screens fault (or the radio), and no one said it was.
> That drug comparison will lead to government's treating computers' like drugs which means heavy regulation of end users and violence against them.
If I really believed that avoiding such a comparison would prevent government from over-regulation and violence toward social media users, then I'd avoid it. But I don't.
Also, using the insanity and violence of the drug war to self-censor obvious comparisons is certainly a choice.
> We need regulation of the corporations intentionally doing random interval operant conditioning. Not regulation of the medium they do it over and the people enjoying using that medium.
No one anywhere was arguing for regulating your screen or the internet - except maybe the government which you insist on doing the regulation, and the corporations who are large enough to own politicians. If you got that impression purely from the tobacco analogy (which you then morphed into the drug war somehow) I'd encourage you to try and reinterpret the point.
It doesn't look like it in the full sense of "free". But part of how one pays these services is by running a permissive modern browser which allows the corporation to spy on you even when you already paid in currency. In a sense by depriving them of the ability to easily spy on your this workaround is closer to "free".
>My best guess is -- ChatGPT is running something in your browser to try to determine the best things to send down to the model API
There's no way this is worth it unless the models are absolutely tiny, in which case any benefits from offloading to the client is marginal and probably isn't worth the engineering effort.
It’s free as a loss leader. The trick is to upsell later. Unfortunately for OpenAI there are plenty of competitors with fungible products, so it might be hard to pull a classic monopoly rug-pull.
They see everything your doing because you send the text. But this is talking about everything about your computer system. You would not normally be sending this to them or having it involved at all. This workaround allows you to not involve unneeded information about your computer setup. It is not about avoiding sending prompt text.
And as for "but chatgpt isn't paid" (another commenter), well, then yes, that's even closer to free by removing this spying on your computer setup. But they spy on the paid users too.
Turning off JS by default and temp-whitelisting only mitigates most of this tracking.
reply