Scale AI lost its focus on product, says Mercor's CEO (www.businessinsider.com)

Mercor CEO Brendan Foody publicly criticized Scale AI on the 20VC podcast, saying the company "lost the focus on product, on scaling quality," even while praising former CEO Alexandr Wang’s distribution and sales skills. Foody contrasted Mercor’s annotation strategy—paying contractors an average of about $95/hour and recruiting elite annotators such as IMO medalists and Ph.D. students—with rivals’ lower rates (he said Scale and Surge pay roughly $30/hour). Scale’s public pay bands list STEM experts at $30–$50/hour and generalists at $15–$30/hour. A Scale spokesperson pushed back, saying data quality metrics are “at record highs.” The dispute comes amid turbulence at Scale: a $14.3 billion investment from Meta, Business Insider reporting on exposed Google Docs and other security holes, and layoffs of roughly 200 full-time staff and 500 contractors. For the AI/ML community, the episode underscores that labeled-data businesses compete not just on distribution but on annotator quality, compensation, and operational security—factors that materially affect model performance, data privacy, and customer trust. It also highlights how rapid scaling and organizational complexity can create bureaucracy and quality regressions, making annotation economics and governance a strategic concern for companies that depend on high-integrity training data.