Even when your AI chats are encrypted, someone watching the network can still guess what you’re talking about.
They call it "Whisper Leak" side-channel attack.
And in tests, models like OpenAI and Mistral gave away topics with 98% accuracy.
There’s a particular Nostr client who charges for nip05 checkmarks. Sometime ago I found how someone was subverting their payment for a nip05 by being able to upload a json file through their media hosting and then linking to it. Apparently, it’s a non-issue for them, and the dismissal was written so passive aggressively. I guess they like loosing money and supporting scammers at the same time 🤷
This sums up what it’s been like trying to work with Apple so I can publish the NYM app. Calls me her coworker’s name and then proceeds to post what I can only describe to be AI gen’d explanations of a 504 error, something I was trying to report on their end, and all the links are to google searches smfh.
Rolled out two quick updates to #NYM the #bitchat client and more...
Hotfix: improved publishing message and reaction speed
Hotfix: corrected shop item flair not correctly parsing html and broken slap with trout