You've identified yourself as a system and seen how the feedback loop of notifications keeps you trapped. But control theory isn't just about analyzing existing misery; it's also about designing systems to achieve a desired state. This desired state is called the setpoint.
A setpoint is simply the target value you want a system's output or state to maintain. For a thermostat, the setpoint is the desired room temperature. For a self-driving car, it's the target speed or distance from other vehicles. For you, the system, the setpoint could be:
Defining a clear setpoint is the first crucial step in attempting to exert control over your digital habits. Without a target, your system (you) will simply drift, driven by external inputs and internal feedback loops towards whatever state the algorithms optimize for (usually maximum engagement).
Algorithms have their own setpoints for you – metrics like daily active users, time spent on platform, and conversion rates. These are the targets they are controlling you towards. To take back control, you need to define your own setpoints and design a way to steer your system towards them.
The visualization below shows some hypothetical data points (perhaps your daily screen time) and a defined setpoint – the target you are trying to reach or maintain.
The setpoint, often denoted as $r(t)$ (for reference signal) or $y_{sp}$ (for setpoint output), is the desired value for the system's output or state.
For a constant target, the setpoint is a fixed value, e.g., $y_{sp} = 2$ hours of screen time.
In the visualization, the dashed red line represents a constant setpoint value. It's the target level you are aiming for.
The difference between the setpoint and the actual system output is called the error signal, denoted as $e(t)$.
$e(t) = r(t) - y(t)$
Where $r(t)$ is the setpoint and $y(t)$ is the actual output (e.g., your daily screen time). The error signal tells you how far off you are from your target.
In the visualization, the vertical distance between each blue data point and the red setpoint line represents the error for that specific day.
The primary goal of a control system is to minimize this error signal, ideally driving it to zero ($e(t) \to 0$). This means getting the system's output ($y(t)$) to match the setpoint ($r(t)$).
Achieving this requires a controller (which we'll discuss in the next module) that uses the error signal to adjust the system's inputs and steer its behavior towards the setpoint.
Defining your setpoint is the necessary first step. It's stating your intention to deviate from the algorithmic default and aim for a state that serves you, not just the platforms. The challenge lies in implementing a controller (your own strategies and discipline) capable of counteracting the powerful forces pushing you away from your desired state.