Writing

• May 16, 2026
  From Double Descent to Scaling Laws
  Reconciling 2010s deep-learning overfitting with 2020s LLM scaling laws.