Anon
My feedback
236 results found
-
618 votes
Anon
supported this idea
·
-
494 votes
Anon
supported this idea
·
-
978 votes
Anon
supported this idea
·
-
6 votes
An error occurred while saving the comment
Anon
shared this idea
·
-
212 votes
Anon
supported this idea
·
-
180 votes
Anon
supported this idea
·
-
1,179 votes
Anon
supported this idea
·
-
847 votes
Anon
supported this idea
·
-
16 votes
Anon
supported this idea
·
-
21 votes
Anon
supported this idea
·
-
28 votes
Anon
supported this idea
·
-
26 votes
Anon
supported this idea
·
-
65 votes
Anon
supported this idea
·
-
58 votes
Anon
supported this idea
·
-
25 votes
Anon
supported this idea
·
-
21 votes
Anon
supported this idea
·
-
9 votes
Anon
supported this idea
·
-
5 votes
Anon
supported this idea
·
-
9 votes
Anon
supported this idea
·
-
12 votes
Anon
supported this idea
·
Links are vital, dangerous, and impossible to verify anymore with data brokers and AI spear phishing.
Recent research shows with a little development, modern LLMs can be used to fully automate spear-phishing campaigns-- from reconnaissance to email and evaluation.
These atracks work as well as human hacking experts with 54% success.
This also makes it way more profitable.
Average non-QubesOS users need virtualization like disposable VMs in QubesOS.
Heiding, F., & Lermen, S. (2025). Evaluating large language models' capability to launch fully automated spear phishing campaigns. In ICML 2025 Workshop on Reliable and Responsible Foundation Models. https://openreview.net/forum?id=f0uFpuea1s