Noesion Radarnoesion.ai
Research2026-08-24

Cognitive scientists and AI researchers are studying why children learn language from a tiny fraction of the text used to train large language models, a difference known as the data efficiency gap. The BabyLM competition, now four years old, trains models on 100 million words or fewer, and one entry beat a far larger Meta model on a grammar benchmark. Frontier labs are largely not pursuing the approach, and models trained on children's headcam video learn only simple words so far.

What changed

Language models have improved mainly by training on ever larger text collections, with no established way to learn well from child-sized amounts of data.

  • 15 trillion tokens: Llama 3.1 pretraining
  • ~100M words heard by a preteen
  • BabyLM corpus: 100M words
  • 61 hours of headcam video

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.