Swing a ball on a string in a horizontal circle, then let go. The ball does not fly outward from the centre; it flies off along the tangent, in a straight line, in whatever direction it was moving at that instant. That shows what the string was doing all along: pulling the ball inward, bending a path that would otherwise be straight. At every moment the velocity points along the circle, and at every moment the pull turns it a little toward the centre.
Turning a velocity is a change in velocity, so it is an acceleration even when the speed stays fixed. It points toward the centre, at right angles to the motion, and its size is a=v2/r. Speed counts twice: going faster means a bigger velocity to turn and less time to turn it. A tighter circle, smaller r, also means more acceleration, which is why a hairpin bend feels sharper than a gentle sweep at the same speed. The time for one lap is the period T, and v=2πr/T.
The outward push you feel on a spinning ride is not a force pulling you out. Your body keeps trying to go straight, the seat or wall pushes you inward to stop it, and you feel that push as being thrown outward. Also, not every curved path is circular motion: a thrown ball curves too, but its acceleration is always straight down, not toward a fixed centre, so it is projectile motion. In circular motion the acceleration keeps rotating so that it always aims at the centre.
Think of it like this. Riding a bike around a roundabout, you keep pedalling at the same pace but must keep turning the handlebars, every instant, to stay on the circle; stop steering and you head off in a straight line along the tangent. That constant steering plays the role of the inward acceleration. Tighter circles and higher speeds both need sharper steering, just as v2/r predicts.
Core idea. Going round a circle at constant speed means the velocity keeps changing direction, so there is an acceleration of size v2/r always aimed at the centre.