Agency in the AI Era - What Matters When Execution Gets Cheap
AI is making execution dramatically cheaper. That changes what makes someone valuable. In this episode, I explore the difference between productivity and agency : the ability to decide what should happen next, act with incomplete information, and take responsibility for the outcome. We get into: Why AI compresses time, skill, capability, and organizational effort Why doing the wrong thing 10x faster is still waste How AI is narrowing some skill gaps and expanding what individuals can execute Why judgment and direction become more important as output gets easier The productivity trap of generating endless plausible, unnecessary work Why high agency also means knowing when to stop, kill a project, or ignore the trend Generative AI and professional writing ChatGPT reduced completion time by 40% and increased output quality by 18% in an MIT study of professional writing tasks. Generative AI in customer support AI assistance increased productivity by 14% overall and by 34% among novice and lower-skilled workers. AI task-completion horizon METR research finds that the length of software tasks frontier AI agents can complete with 50% reliability has historically doubled roughly every seven months. Personal initiative and agency A randomized trial in Togo found personal initiative training increased small-business profits by 30%, versus a statistically insignificant 11% from traditional business training. 00:00 Why agency matters more when execution gets cheaper 01:49 What agency actually means 03:08 The four kinds of AI compression 04:50 The expanding AI task horizon 05:56 Productivity vs. agency 07:02 The AI productivity trap 07:55 Can agency be trained? 09:21 Restraint as a form of agency 10:20 End “Productivity improves the speed of the vehicle. Agency chooses where the vehicle goes.” “As AI makes action cheaper, restraint becomes more valuable too.” Noy & Zhang — Experimental evidence on the productivity effects of generative AI (Science) Brynjolfsson, Li & Raymond — Generative AI at Work (NBER) METR — Task-Completion Time Horizons of Frontier AI Models Campos et al. — Personal initiative training in Togo (Science/PubMed) Brand Runner is for creative leaders and operators who want sharper thinking on creative productivity, creative operations, AI-enabled workflows, and the systems behind better marketing work. Subscribe: https://brandrunner.substack.com/ LinkedIn: https://www.linkedin.com/in/fabien-hameline/ Work: https://fabienhameline.com/ Music licensed through Soundstripe. Code: YVH8LJJIBF2ALQBJ, 2QEXTYLSD6YTPJPV, BHCMGFYBTJXRGAXE, MUIP3GXFUWVOVIRV, 76EX3ND0QS49IVKI Research referencedChaptersQuotesLinksFollow Brand Runner
