Double Slit:

The word “measure” or “observe” carries a lot if confusion when people talk about the double slit experiment, as if it wasn’t one of the most important aspects in physics.
Imagine that at first you’re kicking a ball at a wall at random. Now you put a goalkeeper between you and the wall, he catches the ball and looks at it and then throws it at where he thinks it would have gone if he wouldn’t have caught it. Those are two entirely different scenarios. Same thing goes for the double slit experiment. A non-measuring and measuring setup are completely different, it just seems like magic because you can jump between them with a flip of a switch.
At the atomic scale it is impossible to measure something without the probe interacting and altering the object you are measuring. Imagine if instead of weighing a car to get it’s mass, your have to smash another car into it and calculate the mass from the impact energy. I know it dumb, but that’s essentially the best we can do when it comes to particle physics.
Also in a world where cameras are omnipresent, we are used to take snapshots of reality. But that’s now something you can simply do at atomic scale. You can’t take a photograph of a photon mid air.
Additionally, imagine if I give you two sets of images of a car going along a road.
The first set shows it has travel fairly far between the first and second image taken. You also know the time that passed between the images. Now it is trivial to calculate the average velocity of the car. But if I ask you about the position of the car some time between the first and second image, the best you could do is guess, since you don’t know whether the car might have speed up or slowed down during that time. The second set of images shows the car at almost the same spot, the time between the images is also much much smaller. So small in fact that even the precision of your stopwatch starts to matter. Now you could count the pixel differences and still figure out how much the cat has moved and its average velocity, but it will be much less precise than the first set. But, if I where to ask you about the cars position between the images, you could tell me very precisely where the car was, since it’s velocity couldn’t fluctuate a lot in such a small time frame and distance. That’s in essence the uncertainty principle, you either now the position or the velocity of a particle precisely. Of course you still can’t take a snapshot of a particle. But to figure out the cars position you could also force it to stop briefly or put off-ramps along the road, and by checkout which one it took, you can figure out where it was at that point in time. This is essentially what we do with particles. We confine the particle in it’s position/path to figure out whether it went through one slit or the other and therefore give it a defined trajectory through our method of measurement. That’s really all the magic, you can’t check without interacting and intercating changes the outcome at those scales.

Quantum Computing:
Before we had our normal bit based computer we had analog computers. First mechanical systems then electrical. They operated on either speeds, forces, positions, friction, etc for mechanical ones and voltages, currents and charges for electrical. Calculations hat to be represented with physical components, where each module was a calculation step. The good thing is, the calculation was essentially instantaneous, the bad thing was they were hard to manufacture, had to be calibrated and once setup for compute, they they were hard to alter. But when it comes to things like derivatives and integrals they are hard to beat. For example current is just the time derivative of the charge. So you if set up a computer to measure charge at the input and current the output, you can instantly calculate derivatives of very complicated formulae.

Then, along came Turing and showed that not only can numbers be represented as a binary number sytems, but also that addition, subtraction, multiplication and division operations can be executed through a series of not-and (NAND) operations on those binary numbers. This was good enough for 99% of our problems since we now had a component that was easy to manufacture en masse and almost every mathematical computation could be done to an acceptable level of precision given enough time. But one thing that classical computer suck at is randomness, since whole point of bit based computers was that you don’t have to deal with random fluctuations in your electrical components like with analog computers. We don’t want to measure 1 - 5 ± 0.1 Volts, we want a clear threshold for a 0 and a 1 for our systems to work reliably.
That’s where quantum computers come in. Their bits have an undefined state which can be resented mathematical and is especially useful in calculations where probability is involved. Then, just like with the double slit experiment you “measure” or interact with them to lock in their properties and get your result. Its hard to imagine but there are simply areas of mathematics where working with probabilistic states instead of numbers is necessary.
Think of it as 3 separate areas of mathematics:

Analog computing - > Algebra and Analysis Bit based computing - > Discrete Mathematics Quantum computing - > Stochastic and Crytograpy

Stochastic is generally a hard to understand topic, so quantum computing isn’t far removed in comprehension. That’s why quantum computers also won’t replace classic bit based computing, since they do not do discrete mathematics any better, it’s just that classic computers really are that much worse at probability calculations hence the claims how much faster quantum computing is (in those very specific areas)