Sample efficiency remains a critical bottleneck for current AI systems: humans learn comparable adult competency from reading a small fraction of the words that large language models require during pre-training, suggesting AI needs significantly more data to achieve human-level capabilities.
factualpending
Speaker
Nick BostromEvidence Quote
“sample efficiency seems to still be lagging behind humans... human child... small fraction of the number of words that an LLM has read... current AIs seem to need more data to achieve comparable levels of performance”
Created: 8/12/2026, 6:20:34 PM
Your notes
Loading notes...