DigiNews

Tech Watch by Johan Denoyer

← Back to articles

Regression to the Mean: on LLMs and the quiet death of the new

Quality: 8/10 Relevance: 9/10

Summary

The piece analyzes how large language models tend to output the most probable, average continuations due to training on vast corpora. It argues that true novelty often lies in deviations from the mean, and cautions that over-reliance on consensus can erode originality. The essay offers a framework for recognizing when to seek out the tail instead of the center.

🚀 Service construit par Johan Denoyer