Can LLMs give us AGI if they are bad at arithmetic? (wesmckinney.com) 4 points by Terretta 7mo ago ↗ HN
[–] jqpabc123 7mo ago ↗ I found that the frontier models surprisingly struggled to do basic arithmeticLLM = Large *Language* Model.They are designed to do language prediction, not logical deduction.Attempting to extract truth or accuracy or facts or intelligence from a probabilistic algorithm is an exercise in frustration.
2 comments
[ 4.6 ms ] story [ 22.2 ms ] threadLLM = Large *Language* Model.
They are designed to do language prediction, not logical deduction.
Attempting to extract truth or accuracy or facts or intelligence from a probabilistic algorithm is an exercise in frustration.