michelejoseph

3 ms·

1. 1.▲
   Contradish – AI Policy Consistency Checker(contradish.com)

   1 points by michelejoseph 1mo ago|0 comments
2. 2.▲
   Semantic invariance testing for AI (Contradish)(contradish.com)

   1 points by michelejoseph 2mo ago|0 comments
3. 3.▲
   by michelejoseph 7mo ago

   contradish catches when 2 users get different answers to the same question
4. 4.▲
   by michelejoseph 1y ago

   This theory suggests hallucinations in LLMs aren't glitches but signals of narrative strain just like human emotion, memory distortion, or rationalization under stress. It introduces Compression-Aware Intelligence as a possible missing