CEOs Push Questionable Tools Despite Lower Productivity

Charles Stross, the science fiction author and blogger behind antipope.org, coined the term Barnum's Law of CEOs in a December 2025 post, arguing that C-suite executives routinely mandate AI tool adoption for employees despite research evidence these tools can reduce rather than improve productivity.

The post cites a widely-discussed productivity study in which software developers believed AI coding tools made them roughly 20% faster, but time-trial measurements showed they actually worked about 19% slower (antipope.org). Stross frames this pattern as a satirical law - named by analogy to showman P.T. Barnum - applying to executives who push adoption of tools that sound transformational but deliver mixed results in practice. The 20%/19% finding aligns with the METR research group's 2025 controlled study on experienced open-source developers using AI coding tools.

Stross reports a sharp divergence by seniority: approximately 87% of executives report using AI on the job versus around 57% of managers and 27% of frontline employees, per data cited in the post (antipope.org). The inversion - where those least directly exposed to tool performance adopt most readily while those who do the work adopt least - is central to the post's argument.

Stross argues the primary driver for top-down AI mandates is labor cost optionality rather than demonstrated productivity gain. For many companies, labor is the largest budget line item, and mandating AI adoption signals efficiency ambitions to investors and boards whether or not output metrics yet confirm them.

This is one analyst's perspective, not a peer-reviewed study. The figures cited are drawn from sources referenced in the post. The core tension - a gap between executive adoption incentives and worker-level productivity outcomes - is documented in multiple practitioner surveys and industry research, giving the framing broader evidential grounding than a single blog post.