Directory · C
New here? Start with the premise →
Cravings
2 min read · 372 words
A craving is the reward system lobbying for a specific input with escalating urgency.
The Appetite entry covers the wanting signal broadly. Cravings are the focused version — the system has identified a specific substance, behavior, or experience and is producing an increasingly loud request for it. The signal starts as a whisper and builds. The longer it’s unmet, the more bandwidth it occupies, until the craving dominates the available processing space and the one at the controls can barely attend to anything else.
This escalation is by design. The system learned that this specific input produces a reward signal, and the wanting circuit is applying pressure to secure it. The pressure is proportional to the system’s assessment of the input’s value — which was calibrated by previous reward, not by the input’s actual benefit to the organism.
The craving does not carry information about need. It carries information about the reward circuit’s learning. The system that craves sugar has a reward circuit that learned sugar produces dopamine. The system that craves distraction has a reward circuit that learned scrolling produces novelty-hits. The craving reports what was previously reinforced, not what the organism actually requires.
To work with a craving from the chair: let it arrive. The signal is real, the pressure is real, and attempting to suppress it by force typically strengthens the loop. Instead: note the craving. Name it. (The system is producing a sugar craving.) Then check: is this reporting a genuine need (fuel deficit, energy drop) or is this the reward circuit running a learned loop? If it’s a need, address the need. If it’s a loop, the signal will peak and pass — the hardware cannot maintain the craving at maximum intensity indefinitely.
The peak-and-pass cycle is the craving’s vulnerability. The system escalates, hits maximum, and if the input is not provided, the signal begins to decline. The decline is slower than the organism wants. But it happens. The one at the controls who can sit through the peak without engaging the behavior has ridden the wave — and each wave ridden without the behavior weakens the loop’s next cycle.