The Three Horizons of Adaptation

We can map three horizons for human adaptation:

Augmentation (3-5 years): The Augmentation Phase for AI will amplify human capability but requires significant human steering. Success means becoming fluent in human-AI collaboration, developing taste for what AI can and cannot do, and building evaluation skills that keep pace with generation capabilities. Learning to better use existing (and rapidly developing) AI systems is where we are today and what this website is here to help with.

Orchestration (5-10 years): In the Orchestration Phase AI systems increasingly coordinate with each other, requiring humans to shift from direct manipulation to system design. Success means thinking in workflows not tasks, optimizing for emergence not control, and developing meta-skills that remain valuable as specific capabilities get automated. Human value will be in how we compose and supervise systems that can do practical human tasks autonomously.

Cultivation (beyond 10 years): As we move into the Cultivation Phase AI systems will meaningfully improve themselves, requiring humans to focus on values, boundaries, and purposes rather than mechanisms. Success means clarity about human flourishing, wisdom about unintended consequences, and courage to make choices about what we want rather than just what we can achieve. We become gardeners of intelligence rather than architects of solutions.

The evidence suggests we’re climbing a curve that is becoming progressively steeper. The objective in maximizing human potential in a world of intelligence machines is to develop the skills and wisdom to maximize collaboration with machine capabilities and create possibilities neither could imagine alone.

We have to move beyond a fear of this technology or denial that it’s rapidly becoming more capable. Human potential will be maximized in discovering what becomes possible when invention itself is no longer the bottleneck. And unleashed from the mechanisms and constraints of biological survival we might finally have to decide what we actually want to do with intelligence. The answer to that question is the real work ahead.