this post was submitted on 02 Apr 2026
1 points (66.7% liked)

Privacy

4325 readers
90 users here now

Icon base by Lorc under CC BY 3.0 with modifications to add a gradient

founded 2 years ago
MODERATORS
 

LLM-generated passwords (generated directly by the LLM, rather than by an agent using a tool) appear strong, but are fundamentally insecure, because LLMs are designed to predict tokens โ€“ the opposite of securely and uniformly sampling random characters.

you are viewing a single comment's thread
view the rest of the comments
[โ€“] mr_anny@sopuli.xyz 1 points 3 days ago

Not again this horrible web design/engine whatever laggin/stuttering the way it almost induce a epileptic seizure in me.