Pure silicon barely conducts until you spike it with the wrong atoms
A pinch of phosphorus adds free electrons, and a pinch of boron adds holes, missing electrons that behave like positive charges.

The magic inside a computer chip starts with a material that is deliberately bad at carrying electricity.
Pure silicon is what engineers call a semiconductor, meaning it conducts better than insulators like rubber but worse than conductors like copper wire. The reason is in its atoms. Each silicon atom carries four electrons in its outer, or valence, shell, and each atom bonds with four neighbors to form a tetrahedral crystal, a repeating lattice with four partners around every atom.
Because almost every electron is busy holding that crystal together, very few break free to wander through the crystal and carry current. That scarcity of mobile charges is exactly what makes silicon semiconducting.
It would not be all that useful without its secret weapon, doping, in which silicon is mixed on the atomic level with a small amount of another element. To make one type, engineers slip in an element like phosphorus, which has five outer electrons instead of four. Phosphorus is similar enough to fit into the silicon lattice, but it brings one extra electron that has no bond to sit in. That spare electron can roam, so the crystal conducts much better.
To make the opposite type, they add an element like boron, which has only three outer electrons. Where boron sits in the crystal, one bond goes missing. Engineers call that missing electron a hole.
The hole still helps carry current, because a neighboring electron can hop over to fill it, leaving a new hole behind it. Although electrons are the things physically moving, there are far fewer holes, so it is simpler to track the holes as moving. Because a hole is a missing electron, it acts as a positive charge, so the current is carried by positive-acting holes.
Questions people ask
Does pure silicon conduct at all?
In a perfect crystal at room temperature, a very small fraction do gain enough heat energy to break free, which is why pure silicon conducts a tiny bit. Doping overwhelms that effect by adding millions to billions of times more free carriers, so engineers can set the conductivity precisely.
How much phosphorus or boron do you actually add?
Very little is needed, often roughly one dopant atom for every million silicon atoms. Because each dopant contributes one free electron or one hole, even that trace changes the conductivity dramatically without wrecking the crystal.
Why bother talking about moving holes instead of just moving electrons?
In copper, current is always carried by electrons. In boron-doped silicon, electrons are still the things physically shifting, but tracking the few empty spots is easier math, and the spots drift in the opposite direction under voltage, just as positive particles would.