Those reddit snippets should be fully ignored. The linked article answers your question incorrectly as well (while handwaving around a few unrelated true things here and there.)
The “cloud” is the answer to the question, “Where am I likely to find the electron?” It’s a probability distribution. While it’s mathematical form is a little messy, it’s no different from a Gaussian distribution or whatever in terms of how it functions. So your question becomes, “What sets the shape of the probability distribution for the location of the electron?” and/or “Why isn’t it an infinite spike at zero radius?”
Already at this stage it is required to take the leap from saying (classically), “A particle has a location,” to saying in QM, “A particle has a probability distribution for its location.” This is essential to QM. You can still measure the location of a particle, but the result will be random according to the probability distribution.
Classically, if you know the starting conditions for a particle and you can write down all the forces it feels from the world, you can predict it’s position for all time. You might start by writing Newton’s 2nd law (or an equivalent formulation of the physics) to relate forces (or potential energies) to changes in the particle’s momentum. The evolution of the momentum and, in turn, the particle’s position can then be extracted to provide a description of the position for all time.
In QM, if you know the starting conditions and the forces, you can predict the particle’s position’s probability distribution for all time. Newton’s 2nd law is not the relevant dynamical principle anymore. It needs to be something new that fundamentally works with probability distributions of positions or something related to those (in practice: the “wavefunction” – but we won’t need that detail here.) The relevant dynamical principle – the engine of the physics – is now the Schrödinger equation.
The math is a step up, but the principle is the same: you write down the forces (or potential energies) involved, take any needed starting conditions, and you can work out how the probability distributions behave.
Not any ol’ shape of probability distribution can result from the Schrödinger equation. For any specific force you impose, only certain shapes satisfy the equation. This is where the discrete nature of the physics comes to light. Analogy time:
Discrete answers to a spatial question shows up in many classical systems, for different physical reasons but equivalent mathematical reasons. A plucked guitar string is a common example. The vibrating string has a fundamental (lowest) frequency vibration that corresponds spatially to a one-humped vibration, as in the first figure on this wiki page. The next fundamental shape isn’t arbitrary; it’s the next pattern (two-humped) that satisfies the underlying dynamics while also respecting the fact that the ends of the string are pinned down and can’t move at all. This combination of having “constraints” on the system due to the forces and the boundaries and having a specific dynamical principle at play (here, Newton’s laws) is enough to yield this set of discrete behaviors for the string. These specific shapes of vibrations – plus combinations of them occurring simultaneously – fully describe the possibilities.
The shape of the probability distribution in QM for a particle subject to forces and constraints follows the same mathematical principle as the plucked string: certain shapes are fundamental. Further, different fundamental shapes for the probability distribution generally correspond to different amounts of energy in the system. These different “states” of the system are all that can exist (along with combinations of them).
Since the different states have different energies, the stable configuration is the one(s) with the lowest total energy. But even that lowest energy state has some specific shape for its probability distribution. For the electron in hydrogen, that lowest energy state has a probability distribution that is spherically symmetric and falls off exponentially with radius. That’s the dynamical “end of story” of this isolated system. There’s nothing further for it to evolve into. There is no more-squished probability distribution available because no viable shape with smaller energy exists given the specific forces involved and the dynamics embedded in the Schrödinger equation.
For a plucked string, the ends of the string being pinned down is like a “brick wall” constraint for the system. For the electron attracted to a proton, there isn’t a hard boundary like this. The proton provides not a brick wall defining a spherical enclosure but rather a spherical gradual “valley” that influences the electron. All the same, though, these forces lead to a suite of special shapes that underlie all resulting behavior. The electron system also has considerations of angular momentum that lead to much more complexity in the suite of shapes, and this is why your “hydrogen wavefunction” image has so much going on. (That image doesn’t actually include the lowest-energy state at (1,0,0), which is simplest in shape.) But the fundamental reasons for “Why doesn’t the electron fall in?” don’t require all that complexity.