Once you get to that level of science, you can't expect the things they're talking about to physically exist or be actual descriptions of something. They're just models of how something is behaving.
I tried explaining this to someone using the atom as an example. It's too small to actually see what it looks like beyond a blob. The initial model was that it looked like a cake, with the particles all lumped together. Then we switched to the one that'll be in your minds, of a sphere with electrons flying around it in a single orbit. In reality, a more accurate description is of a sphere surrounded by various stacked shells of energy, as the electrons are moving so quickly they behave more like a layer, where charge couples into paired regions.
Then you have light, that can either act like a wave of energy or a solid particle (photons). Just models to describe the activity.
Superstring theory came about because, unlike the two models on their own (which work perfectly when separate from each other), trying to marry quantum physics with things like gravity doesn't work. Primarily because our way of understanding space time and thinking it can be chopped up into infinitely tiny pieces; e.g. a meter will go to centimeters, then millimeters, then microns, then nano, then picos... etc. That meant that the equations trying to join the two kept having infinity in them (because it could go on forever), and wouldn't work. So superstrings have been mathematically proposed to set a finite limit on the chopping down of space time. The multitude of dimensions they need to model mathematically may not actually exist as places, it's just how the math needs adapting to describe the behaviors at those sizes, energies, velocities, densities etc.
I also described to this someone how the theory is being proven correct. For example, the man who designed the periodic table not only managed to predict properties from the order of the table, he managed to predict elements that hadn't been isolated and left spaces for them. We've since found those elements.
In the same story, the physicists who have devised superstring theory have not only found certain particles, they've managed to extrapolate from the math which others should exist, which ones can't and what groups they should appear in. Then found them. Demonstrating their predictions are not based on wild guesses but predictable math. The problem is, the math involved in that work is waaaaaaaay above most standards (some of it doesn't even work using normal math) and there's no hope of most people understanding it. Most people can't remember Pythagoras' rules, let alone start talking about three dimensional vector based differential algebra, which is still miles from the level they're working at. But those seemingly wacky ideas they come up, like the numbers of dimensions, are based on predictable, demonstrable logic. They can predict the existence of something and then see it's effects manifest up to higher levels in particle accelerator experiments.
The scientist even know themselves how much these models are based on abstract math. For instance, when it got to quarks, they started giving them arbitrary names like up, down, strange, charm, top and bottom. Because they don't resemble anything humans are used to contacting with their tactile senses.
It may not be perfect or composed of actual strings but the nay sayers no not of what they discuss and just how much it is starting to permeate science they are coming to rely on. E.g. the memory in your phone uses quantum tunneling. Effectively the teleportation of solid particles through solid barriers. Something most people would still laugh at the idea of.
