Gemini models increasingly stucking in thinking loop
ai
Google's Gemini models are increasingly getting stuck in thinking loops, according to testing shared on Hacker News. In one hundred trials, the three point five Flash version froze twenty-three times, while the three point one Pro got stuck sixteen times. The frequency of these failures appears to be growing.
Source: https://news.ycombinator.com/item?id=48642229
Listen to this story
Hear this and more stories in a personalized audio briefing.
Open The Chonkerton