Message on X (Agency Over Intelligence)
By John Carmack · 2026
A speculation on how AI assistance could shift the relative importance of intelligence, agency, trust, and decision-making in human achievement.
Carmack considers how artificial intelligence might change the relationship between intelligence and personal agency. He argues that modern society has strongly rewarded people who combine high cognitive ability with a strong capacity to act. As AI systems increasingly perform tasks associated with intelligence, he suggests that exceptional agency could become more important, including for people who would previously have been limited by lower cognitive ability.
The argument depends on a willingness to accept AI guidance without allowing pride or ego to interfere. Carmack illustrates this with an intentionally extreme hypothetical example: a highly determined criminal who completely trusts advice from an always-available AI assistant. The example is not presented as approval of criminal behaviour, but as a way of isolating the broader idea. If AI systems can reliably supply judgement, planning and information, human effectiveness may depend increasingly on motivation, decisiveness and the willingness to follow competent external guidance.