‹ BackHN Continuity

Thread

AI chatbots give wrong answers to financial queries 'most of the time'

157 points · 89 comments · 1vuio0pswjnm7

  1. 01100011 · · focus · HN ↗
    Single shot or with reasoning enabled? My experience is that reasoning dramatically reduces hallucinations and improves output quality. I don't trust models without it.
    1. wonnage · · focus · HN ↗
      > Overall, the best-performing model was Claude Opus 5 on “reasoning” mode, which still made mistakes in 39 per cent of answers.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.