The case against treating AI hallucinations as intractable
One poster argues that views of hallucinations should catch up with changes over the past six months.
TLDR
A poster argues that people who see hallucinations as an intractable problem for current AI models should reconsider, pointing to changes over the past six months. A second poster quotes that argument and says factuality keeps improving.
Combined views
2.5K
1 Source, first seen ago
