Stop Waiting for New AI Models: Here’s How to Use Data to Make AI at Work, WORK

Source: Nate’s Newsletter Date: 2025-05-14 URL: https://natesnewsletter.substack.com/p/stop-waiting-for-new-ai-models-heres

Summary

Organizations obsess over getting the latest model when the real productivity ceiling is data quality — “the best model in careless hands is worse than useless.” Nate argues that redirecting energy from tracking model releases toward data infrastructure could double AI returns, while incremental model improvements (10% better emails) don’t meaningfully advance business outcomes without foundational data work.

Implications

Enterprise adoption thread. The data-first argument is a recurring theme across the practitioner newsletter space, and its persistence suggests it’s addressing a genuine failure pattern in enterprise AI deployments. Organizations that treat model upgrades as the primary lever keep underperforming expectations.

AI economics thread. The “unsexy reality” framing is strategically useful: if data infrastructure is the bottleneck, the economic value in AI stacks accrues to whoever controls data preparation and quality — not necessarily to model vendors. This shifts the competitive map.

Watch: Whether enterprise AI buyers start prioritizing data infrastructure vendors over model vendors in budget allocation — and whether dedicated data-for-AI products (beyond generic data engineering tools) emerge as a distinct category.