CorrectPrediction
By end of 2024: no robust solution to hallucinations
“No robust solution to hallucinations”
Context: One of seven predictions "By end of 2024" posted on X (Mar 10, 2024). The quote is a short bullet from that list.
How we judge it
Our interpretation: correct if by 31 Dec 2024 frontier LLMs still hallucinate at rates that prevent unsupervised use in high-stakes settings and no lab claims a general fix.
Resolution criteria are our interpretation of the claim, written before grading. Methodology
- Said
- Resolve by
- Verdict
- Correct ()
- Stated probability
- None stated
- Specificity
- 2 of 3 (dated, some interpretation)
- Score
- 1.0 points × weight 2
Evidence & grader’s note
- LessWrong "2024 in AI predictions": the list is "approximately correct regarding non-agent models"
- Gary Marcus on X grading himself 7/7 (Dec 30, 2024)
Hallucination remained an acknowledged open problem at all major labs through 2024 and beyond.
