People deploying these systems are just hoping that Prompt injection attacks won’t happen.

They could design systems that would be resistant, but the only thing that matters now is deploying new software… not creating actual security or sustainable systems.

Unless you’re (not you PhoenixDog specifically, the general You referring to the reader) blocking your browser fingerprinting, using Ad blocker, DNS filtering from trusted sources, not using social media app, not installing apps that require Google Play Services and not using your credit card for digital payments then simply changing the IP address that the server sees doesn’t do very much.

VPNs can be part of a solution but if you are not doing all of the other things then they do not hide who you are because of the plethora of other avenues that you can be tied to an existing advertising identity.

The fact that giant companies keep your data and don’t delete it when you tell them to has been true since the beginning of social media. Your things are not deleted, they’re simply marked as deleted so you don’t see them. The actual binary data never goes anywhere.

The rule of thumb is that if the data leaves your possession then assume someone has a copy of it. If it is encrypted and you don’t control your keys then it isn’t encrypted. (See: Bitlocker keys and Microsoft)

They know their algorithms promote anger and hate, they even tried changing them to promote wholesome content but that change caused ad sales to fall so they decided to sacrifice civilization instead.

Someone appoint this person to the Federal Reserve

I like money

Mike ‘The Prophet’ Judge has forseen this

nor have the training/licensing required to cary/display them.

Exactly.

Recommending that random untrained people go around armed is a recipe for disaster (see: ICE’s random untrained people going around armed). It’s irresponsible, at best.

I read these people as if they’re calling for violence but trying to skirt the line in a way that doesn’t get them banned.

You can do it with any 3D printer, as long as you orient the letters so they’re at the top of the print, you can insert a pause in the print job just before it starts the layers with the letters. Then you do a manual filament swap and resume the print.

Another option is to print a QR code on them with a link to instructions/the organization.

Same.

If someone isn’t willing to take a minimum amount of effort to stay in contact with me, then I’m not losing much if I don’t ever talk to them again. They don’t have to quit Discord, they just have to find me elsewhere.

Sure, sure.

The problem is with who gets to decide who ‘the homeland’ consists of.

I want to be shocked, but gestures to the world, that these transparent attempts to control the market by limited consumer choices were not immediately responded to by the responsible regulatory systems that we have setup.

Maybe once the Fascists are gone we can replace the smoking hole that is the US government with something more akin to the fairy tales that we were told in class.

jk, new dark ages coming right up

The person posting about a RAT is either unwell or trolling, dumping paragraphs of nonsense and screenshots that don’t actually show anything.

The second that they claimed to be able to detect data exfiltration via wireguard is what lost me (even script kiddies use encryption, advance attacks would exfiltrate data in DNS requests or some other exotic method). That and they were not describing a malware infection but an active attack by a person/people who were able to determine what steps that OP was taking and react.

Also, if you think your system is compromised the first thing you do is remove power from the infected machines, you don’t use them to try to determine what is wrong (when the attacker could have just corrupted your tools, or replaced the kernel with a kernel who lies to sys calls., etc)

Anti-viruses’ popularity is a bit paradoxical.

If something is good then it becomes popular, and then once it is popular malware developers add specific testing for that specific AV software so it becomes less good over time.

In my testing, by copying the claimed ‘prompt’ from the article into Google Translate, it simply translated the command. You can try it yourself.

So, the source of everything that kicked off the entire article, is ‘Some guy on Tumblr’ vouching for an experiment, which we can all easily try and fail to replicate.

Seems like a huge waste of everyone’s time. If someone is interested in LLMs, then consuming content like in the OP feels like knowledge but it often isn’t grounded in reality or is framed in a very misleading manner.

On social media, AI is a topic that is heavily loaded with misinformation. Any claims that you read on social media about the topic should be treated with skepticism.

If you want to keep up on the topic, then read the academia. It’s okay to read those papers even if if you don’t understand all of it. If you want to deepen your knowledge on the subject, you could also watch some nice videos like 3Blue1Brown’s playlist on Neural Networks: https://www.youtube.com/watch?v=aircAruvnKk&list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi. Or brush up on your math with places like Khan Academy (3Blue1Brown also has a good series on Linear Algebra if you want more concepts than calculations).

There’s good knowledge out there, just not on Tumblr

Budget-Chicken-2425, waking up one day to find out the entire world is reading your FBI domestic terrorist dossier:

That’s enough to radicalize a chicken.

A bit flip, but this reads like people discovering that there is a hammer built specifically for NASA with specific metallurgical properties at the cost of $10,000 each where only 5 will ever be forged, because they were all intended to sit in a space ship in orbit around the Moon.

Then someone comes along and posts and article about a person who posted on Tumblr about how they were surprised that one was used to smash out a car window to steal a Door Dash order.

LLMs will always be vulnerable to prompt injection because of how they function. Maybe, at some point in the future, we’ll understand enough about how LLMs represent knowledge internally so that we can craft specific subsystems to mitigate prompt injection… however, in 2026, that is just science fiction.

There are actual academic projects which are studying the boundaries of the prompt-injection vulnerabilities if you read in the machine learning/AI journals. These studies systemically study the problem, gather data and demonstrate their hypothesis.

One of the ways you can tell real Science from ‘hey, I heard’ science is that real science articles don’t start with ‘Person on social media posted that they found…’

This is a very interesting topic and if you’re interested you can find the actual science by starting here: https://www.nature.com/natmachintell/.

I use virtual desktop for GUI tools. I put the volume mixer, OBS and a terminal with ncmpcpp for music. So if I need to screen record or swap inputs or change music it is all on the same key.

remove is -R

so

sudo pacman -R vulcan-mesa-implicit-layers

You can force it to ignore dependencies (be careful doing this as it is bypassing an important safety check in the package manager), for example if you’re trying to remove a package that you’re going to replace with another one then you can ignore the dependency warning when you try to remove it.

Use -Rdd to remove and ignore dependency checking.