#Data Curation

Data Quality Is the Compute Multiplier Says Ari Morcos

Ari Morcos explains why high quality data acts as a compute multiplier for AI models, dramatically lowering training and inference costs.

Web-Scale LM Pretraining Poisoning Feasible

New research reveals public discussion interfaces enable web-scale language model pretraining poisoning, with 'HalfLife' analysis quantifying the threat.

OpenData Pipeline Elevates Agentic AI

The OpenThoughts-Agent project introduces an open data pipeline that significantly enhances generalization for agentic language models, outperforming existing benchmarks.

The Bitter Lesson: Why Data Curation is AI's Underinvested Frontier