Major interpretations address the measurement problem differently: Copenhagen (subjective ψ), many-worlds (all branches), objective collapse (physical collapse), Bohmian (deterministic trajectories), relational (relative properties). All reproduce empirical predictions.
You've encountered the measurement problem: quantum mechanics assigns a wavefunction that evolves deterministically via the Schrödinger equation, producing superpositions — yet every measurement yields a single definite outcome, not a superposition. Something not contained in the Schrödinger equation must be happening during measurement. Every interpretation of quantum mechanics is an attempt to make sense of this apparent collapse, and they differ fundamentally on whether collapse is real, what the wavefunction represents, and whether quantum mechanics is complete.
The Copenhagen interpretation, the default in most textbooks, takes a deliberately agnostic stance: the wavefunction is a tool for predicting probabilities, not a description of an objective physical reality. Collapse is simply the update of probabilities upon getting a result. The quantum-classical boundary is drawn pragmatically — the measuring apparatus is treated classically. This works perfectly for calculations but leaves open what is "really happening" before measurement, which many physicists consider an unacceptable silence about fundamental ontology.
Many-worlds (Everett) takes the wavefunction at full face value: it is real, and the Schrödinger equation is exact and universal. When a measurement occurs, the universe branches — all outcomes happen, in different branches of a vast universal wavefunction. There is no collapse and no special role for observers. The difficulty is explaining why we experience probability at all: why do we observe outcomes with Born rule frequencies rather than just finding ourselves in arbitrary branches? This is the "preferred basis problem" and the "probability problem," both active areas of debate. Objective collapse theories (GRW, CSL) modify the Schrödinger equation itself by adding small random collapse terms that are negligible for single particles but cumulative for large systems, causing macroscopic superpositions to collapse in microseconds. These are empirically distinguishable in principle, though extremely difficult to test in practice.
Bohmian mechanics (pilot-wave theory) restores determinism entirely: particles always have definite positions, guided by a real pilot wave that satisfies the Schrödinger equation. The apparent randomness of quantum mechanics is purely epistemic — we don't know the exact initial positions. This interpretation is empirically equivalent to standard quantum mechanics, but the guiding equation is nonlocal: the velocity of a particle depends instantaneously on the configuration of all other particles, no matter how distant. Relational quantum mechanics takes yet another path: quantum states are not absolute properties of systems but are relative to observers, and different observers may legitimately assign different wavefunctions to the same system. Choosing among these interpretations is currently a philosophical question, not an empirical one — and knowing that all of them reproduce the same predictions should give you healthy skepticism toward anyone who claims the question is already settled.