Link / Augment

You and a support component work interactively, and neither alone carries the whole task.

Stuck on a proof in the statistics course you are taking, you have two bad options within reach: give up and look at the full solution, which teaches you nothing, or grind alone for another hour, which is producing nothing. What you actually want is a third thing: something that nudges you past the specific wall while you still do the climbing.

LINK / AUGMENT is the relationship mode where you and a support component work interactively and neither alone carries the whole task. Unlike delegation, nothing is handed off; unlike working solo, you are not unaided. You remain in the loop doing the core reasoning while the partner, a tutor, a study partner, an AI, supplies targeted pieces: a graduated hint, a check on a step, a question that exposes your wrong assumption. The cue is a task where full outsourcing would destroy the value (here, your learning) but zero support means stalling. The operation: keep the attempt in your hands, request the smallest useful contribution ("tell me if my second step is valid, not the answer"), incorporate it, and continue, alternating between your work and its input. What changes is momentum without forfeited ownership: you finish the proof and you could do the next one. The boundary: if the partner's share keeps growing until you are mostly assembling its output, you have drifted into delegation while billing it to yourself as learning.