Systems
Technology should extend human capability
The central question is not how much work a system can do alone, but how much more a capable person can achieve with it.
Discussions about artificial intelligence too often begin with the question of whom it can replace. The question is easy to measure, fits neatly into a presentation, and leads to a remarkably poor vision of technology.
I am interested in a different question: how much more can a capable person achieve when given a system designed around their knowledge, responsibility, and way of working?
A tool is not a moral agent
A specialist understands context, remembers exceptions, and lives with the consequences of decisions. A business owner knows which tensions can be accepted and which will destroy the venture. A leader recognises the difference between an answer that is technically valid and a decision that is right for the organisation.
Technology can extend their reach. It can preserve continuity, surface relevant information, prepare alternatives, enforce agreed rules, and perform a substantial amount of repetitive work. It should not pretend that responsibility has disappeared merely because an answer was generated quickly.
Design for agency
A system that extends a person should leave them able to understand, challenge, and stop its action. It should expose the limits of its confidence, preserve the provenance of decisions, and distinguish a recommendation from an authority.
This is not a brake on innovation. It is a condition for using innovation where outcomes matter: in business, medicine, finance, infrastructure, and teams responsible for other people.
The best technology does not make a person redundant. It lets their experience operate further, faster, and with greater precision.