Resistance you only notice by its effects

In physical space, resistance determines efficiency. When airflow and energy stagnate at specific nodes, the occupants' decision-making and thoughts become obstructed. Feng Shui is not mysticism; it is about how to remove “invisible resistance” from a space, allowing the environment to unify with the rhythms of heaven and earth.

Picture a meeting room on the third floor at half past two in the afternoon. Eight people, door closed, blinds drawn against glare. The first twenty minutes go well. By the second hour the same questions are being asked twice, decisions are pushed to next week, and someone suggests coffee. Nobody blames the room. They blame the agenda, the lunch, the length of the meeting.

The room has stopped moving. The air in it is no longer being replaced at the rate eight occupied bodies need. Daylight has been shut out, and the lighting holds a fixed level chosen for a half-empty office in the morning. The resistance is real, and it sits at a node nobody is watching.

Every system in the building reported normal

What makes this scenario so common is that nothing failed. The thermostat saw a temperature inside its band and was satisfied. The VAV box serving the room took its cue from that temperature and stayed near minimum airflow. The lighting ran its schedule. The blinds were closed by hand and will stay closed until someone opens them tomorrow.

Each system did exactly what it was configured to do. The failure lives at the junction between them, where occupancy, air, light and time of day ought to meet and never do. Air conditioning answers to temperature. Lighting answers to a clock. Nobody answers to the eight people in the room.

This is what invisible resistance looks like in practice. It is rarely one broken component. It is the friction produced when several correct components each optimise their own variable, and the variable that matters to the occupant belongs to no one.

What Feng Shui was describing with stagnation and flow

Feng Shui has words for this. Stagnant corners, blocked circulation, a room where energy gathers and does not move. Set the cosmology aside and these describe things a careful observer could notice long before anyone could measure them. There is the corner where the air never seems to change, the corridor where people slow down and bunch up, the room that feels heavy by mid-afternoon.

The tradition kept these observations as rules because it had no other way to store them, and that was its weakness. A rule about keeping the space behind a desk open, or avoiding dead-end circulation, might encode a real pattern about airflow or attention. It might not. There was no instrument to tell the difference, so good observations and bad ones were handed down with equal authority.

The rhythms of heaven and earth become practical in the same way. A building sits under a sun that moves across its façades, surrounded by outdoor air that changes by the hour and the season. A space in step with those rhythms adjusts to them. A space running on a fixed schedule ignores them and slowly drifts out of phase with its own occupants.

Turning stagnation into something a system can read

Resistance becomes manageable once it has a signal. For the meeting room, three signals are enough to describe what went wrong: whether the room is occupied, which way the air quality is trending, and how much daylight is available at the window.

In an apporo deployment these signals come from different places, which is exactly why they were never connected. Sensors and wall controls typically sit on KNX, where group addresses live in the devices themselves and keep working if the controller goes offline. Air handling units and VAV boxes usually speak BACnet. Luminaires answer through DALI. Home Assistant brings all of them into one entity model, so the meeting room exists as a single thing with air, light and occupancy attached to it.

A Node-RED flow can then express the rule the room was missing. When the room is occupied and air quality is getting worse, raise ventilation to that zone, whatever the thermostat reports. When daylight at the window is strong and glare is low, bring the blinds up and dim the fixtures. Neither rule is clever. They connect variables that were never allowed to talk to each other.

The limits belong here too. A single air quality sensor near the return grille can report a healthy average while the far corner stays stale, so placement decides whether the data describes the room or only part of it. Sensors drift and need a maintenance schedule. None of this is expensive, but it has to be planned.

Most of the work is taking steps away

The second kind of resistance is operational. A room that needs five switches, a blind remote and a call to facilities before a presentation can start has resistance built into its use. People give up. They leave the lights at full and the blinds shut, and the room settles into its worst configuration because that is the easiest one to reach.

Removing that friction usually means fewer controls. One scene on a wall panel sets lighting, blinds and ventilation for the activity. Automations restore sensible defaults when the room empties, so the next group does not inherit the last group’s choices. The apporo GUI lets a facility team adjust those scenes without writing code, and version control with rollback means a change that makes things worse can be undone in minutes.

The last piece connects back to the tradition’s weakness. When a flow is tuned, the reason for the tuning should outlive the person who did it. apporo’s O&M SOP handover keeps the logic and its rationale together, so a new facility manager three years from now can see why the meeting room ventilates on air quality and not only on temperature. A rule that remembers its reason can be revised. A rule that has lost it can only be obeyed or ignored.

Who gets to ask where the resistance is

Everything above is ordinary building automation. None of it needs AI, and none of it should wait for AI. Sensors, a shared entity model and a few well-chosen flows remove most of the stagnation a building produces.

What a language model changes is who can investigate. The history already exists, kept in PostgreSQL through Supabase. Finding the rooms that went stale on weekday afternoons has always been possible for someone who can write the query. In most buildings nobody does, because that person is busy or was never hired.

With Pi Agent running on Home Assistant, the people who make decisions about the space can ask directly. A facility manager can ask in a sentence. An owner can check whether the rooms nobody likes to book are the rooms with the weakest ventilation response. An architect planning a refurbishment can ask which zones showed the most stagnation before committing to a layout.

The answers are only as good as the sensors behind them, and a model can phrase a thin answer with confidence, so the underlying data should always be one step away. Within those limits, the resistance the tradition could only sense becomes something anyone responsible for the building can locate, question and remove.

Where to go from here