Every making endeavor runs on six activities. Each activity connects two of three systems, Need, Resources, and Solution, using one as instrument to modify the other. When you rank these activities by how much attention they naturally draw from you, the top two aren't random. They share structural connections in the network, which means they concentrate your perception in specific parts of it. And the parts they don't reach get starved.
This is overcurrent. Not just "I prefer these two activities." Something more mechanical. Your top two activities create a circuit in the making network that sustains itself. Each one gives the other something to work on, or both converge on the same target, or both read from the same source. The circuit stays energized. The rest of the network goes quiet.
The geometry of a pair
Any two activities must share at least one system. That's arithmetic: six activities spanning three systems means every pair overlaps somewhere. But how they overlap determines what kind of overcurrent you get. There are exactly four possible structural relationships.
The pipeline. One activity modifies a system that the other picks up and uses as its instrument. Attention flows through a shared node in sequence. Take Provisioning (Need drives Resources) and Delivery (Resources produces Solution): Provisioning builds the capacity that Delivery then uses. The flow is Need to Resources to Solution, a directed chain. What gets missed is everything running in the other direction. Feedback from Solution back to Need never enters the picture. The pipeline is efficient and narrow.
The double reshaping. Both activities target the same system from different angles. Take Scoping (Resources constrains Need) and Validation (Solution tests Need): both modify your understanding of what's needed, one from the capacity side, one from the evidence side. Need gets refined thoroughly from two directions. Meanwhile the other two systems sit still. Thorough in one-third of the picture. Blind to the rest.
The launchpad. Both activities use the same system as their starting point without modifying it. Take Validation (Solution tests Need) and Verification (Solution improves Resources): both read Solution for what it reveals, but aim their findings in different directions. The existing artifact becomes the lens for everything. You interrogate it from two angles without producing more of it or questioning whether it was the right thing to produce. What exists gets treated as more settled than it may be.
The closed loop. Both activities connect the same two systems in opposite directions, forming a direct feedback circuit. This is the tightest lock. Take Specification (Need determines Solution) and Validation (Solution tests Need): each activity's output feeds the other's input directly. Need and Solution chase each other in tight iteration while Resources sits entirely outside. The circuit feels complete from inside. The third system becomes invisible.
The squeeze
Here's the structural insight most people miss. The activities at the bottom of your ranking aren't just "less preferred." They're squeezed by the pattern at the top.
When your top two both target the same system, the systems they don't touch get neglected from both directions simultaneously. If your top pair both reshape Solution, say Specification defining it from the Need side and Delivery building it from the Resources side, then Solution gets enormous attention while Need goes untested and Resources goes uninvested. Two-thirds of the network's corrective mechanisms go dark, and the maker feels enormously productive because the artifact is growing.
When your top two form a closed loop, the excluded system isn't merely deprioritized. It's structurally invisible. The loop's internal coherence makes its external blind spot disappear. A person looping between production and process improvement can get measurably better at building things and never once question whether any of it serves the actual need. Need sits outside the circuit. And the circuit feels whole. That's the danger: the completeness is felt, not real.
The bottom two activities compound the problem. They're not just individually neglected. They're neglected as a pair, and that pair has its own structural relationship. If both target the same system, it's getting starved from two directions with no corrective mechanism. If they form a broken loop, two systems lose their direct feedback relationship and misalignment accumulates silently.
Why this doesn't feel like a problem
Overcurrent is self-sustaining. The top-two circuit generates visible progress. Real output. Tangible work. Things are happening.
What the bottom produces is a different kind of absence: defined by what doesn't happen. Questions that don't get asked. Corrections that don't get made. Gaps that widen silently because the mechanism that would detect them isn't running.
The person inside the overcurrent can't see the cost because their top-two activities are working. The outputs are real. The reinforcement is real. What's missing is information that only the starved activities would have generated. And since those activities aren't running, that information doesn't exist, which means the need for it stays invisible.
Your top two create the conditions that squeeze your bottom two. The shape of what you over-attend determines the shape of what you miss. And the things you miss are precisely the things your overcurrent prevents you from noticing you're missing.
Steering, not flattening
The response isn't to stop doing what you're drawn to. Your top pair isn't a defect. The pipeline, the loop, the launchpad, the double reshaping: these produce real value. They're where your making has energy and coherence.
The response is to know the shape of your circuit. Which system sits outside it. Which feedback relationship has gone quiet. Which system gets reshaped endlessly while the others drift. And to build structures that compensate: collaborators who attend where you don't, checkpoints that force the questions your perception skips, processes that route attention toward what your overcurrent starves.
You can't flatten your attention profile. You can learn its geometry and build around it.