"salaries of top AI researchers employed in industry roles, which have more than tripled from $595,999 to almost $2mn between 2001 and 2021" [$595,999 to almost $2mn]
"when researchers move from academia to industry they become less likely to publish new research papers and more likely to file patents" [less likely to publish]
The article treats this as a structural shift in knowledge production, not a temporary labor market fluctuation. The authors explicitly note that researchers moving to industry become less likely to publish and more likely to file patents—a shift from open to proprietary knowledge. This dynamic extends beyond AI researchers to other elite knowledge workers, suggesting a broader reorganization of intellectual capital concentration in private firms.
The article frames this as a significant shift in the mode of knowledge production itself, not merely a change in employment. The authors note that this exodus into 'larger companies with strict intellectual property regimes' represents a move away from 'open, shared knowledge production to closed, proprietary work.' This has implications beyond AI: if the same dynamic applies to other knowledge work as the article suggests, it signals a broader reorganization of how intellectual capital is managed and monetized in the economy.