24 ms·
I'd say what this article talks about hardly scratches the surface. I won't say more but I am surprised at the reaction here, all of it is legal, if you don't l
by _8j50 3y ago
I'd say what this article talks about hardly scratches the surface. I won't say more but I am surprised at the reaction here, all of it is legal, if you don't like it, talk about passing laws. And what would you penalize? Lying on the internet?
The 4th amendment is a restriction on the government not private companies. You have no right to expect people joining your discord server have honest intentions. Why are you using discord to begin with?
You know, there is a reason gang members require newbies to commit a heinous crime when they join.
How can I put it, trust establishment is not a technical problem, anyone can betray you, so find sociological methods to prevent that.
But if you are a criminal, if you justify your harmful means by some ideological ends or use whataboutism to justify stealing from or harming others, the people that gain your trust and betray you are heroes to me! How can you act outraged when they too justify their means by their ends? Don't get mad at someone beating you at your own game.
You all should know that an agent of these companies that engaged in hacking (getting past any security control without permission) is criminally (federal) liable. Conspiracy to committ most serious crimes itself is usually a crime. It's illegal to hack criminals. What is being talked about here is criminals that suck at vetting who joins their chatrooms.
This is what krebsonsecurity does except these companies sell it as a service and they have certain useful commercial relationships.
The FBI, if it was half as competent as people think it is (must be the suits?) would also legally be doing this if it paid people well. Undercover agents have always been a thing, they even get immunity for crimes while on the job! Unlike civilians doing threat intel.
Lastly, this is threar intelligence research & development. There's a whole sub-field in infosec on what to do with intel once you have it. But really, marketing aside, this is HUMINT mostly. The information is technically about threats (humans) but it is usually not about their techniques. More like "look out for this IP" or "you're about to get ransomwares" or "this person's creds at your company is pwned".
I hope I don't piss off anyone but if you really don't like this, make it commercially non-viable by requiring disclosure of compromised info or potential compromise to all affected parties.
Oh, and just to add: I said I hope I don't piss off anyone because like they mention in the article, they monitor you at a young age and will develop enough text from you that you can never type long text like this without being identified (stylometry). But I don't think I've said anything that controversial. Just keep in mind, those of you who dabble in criminal stuff, even at a young age, you're never escaping that stuff. Before it was just stylometry, but I have no idea how much LLMs have improved this technology in recent years.