323
LLMs can unmask pseudonymous users at scale with surprising accuracy
(arstechnica.com)
This is a most excellent place for technology news and articles.
At the very worst, the administration would put out a very confusing statement not to trust AI.
That would be fun.