The answer to that question is straightforward: When they are babbling about mathematical models that have nothing to do with physics. When they are bloviating on the nature of physically meaningless concepts like Singularities. Here is a textbook example from that Journal of Mathematicism, Quanta Magazine.

In the fashion of modern theoretical physics the existence of singularities is first attributed disingenuously to Einstein: “Singularities are predictions of Albert Einstein’s general theory of relativity.“

That statement is simply false. The Singularities characteristic of Black Holes are predictions of the Schwarzschild solution to the GR field equations, while the Singularity of the Expanding Universe model (Big Bang) is a prediction of the FLRW solution to the GR field equations. It should be noted that both Black Holes and the Expanding Universe are also predictions of the respective mathematical solutions.

In order to arrive at their results it was necessary for both Schwarzschild and FLRW to make a number of simplifying assumptions. Schwarzschild left out a critical prediction of GR, that the speed of light varies with position in a gravitational gradient; FLRW decided that it was reasonable rather than self-contradictory to solve GR for a non-relativistic universal framework. Both “solutions” produce Singularities that are a consequence of their simplifying assumptions. Singularities are mathematical artifacts that have no physical significance.

The article acknowledges that some physicists understand that Singularities are not physically meaningful: “… singularities are widely seen as “mathematical artifacts,” as Hong Liu, a physicist at the Massachusetts Institute of Technology, put it, not objects that “occur in any physical universe.”

But what fun is that? The article quickly pivots to a discussion of some additional arcane mathematical models in which the Singularities are also present and “are proving hard to erase“:

The British mathematical physicist Roger Penrose won the Nobel Prize in Physics for proving in the 1960s that singularities would inevitably occur in an empty universe made up entirely of space-time. More recent research has extended this insight into more realistic circumstances. One paper established that a universe with quantum particles would also feature singularities, although it only considered the case where the particles don’t bend the space-time fabric at all. Then, earlier this year, a physicist proved that these blemishes exist even in theoretical universes where quantum particles do slightly nudge space-time itself — that is, universes quite a bit like our own.

Note that all of these models treat space-time as a substantival entity despite the fact that there is no scientific evidence to support the common mathematicist belief that space-time is a physical, causally-interacting substance. This lack of evidence for an essential component of the models means that none of them have anything of scientific value to say about the nature of the Cosmos we actually observe.

It should also be noted that the inability of the modelers to “erase” the Singularities in their models is a mathematical problem that has nothing to do with physical reality which does not contain Singularities. It’s just some math, not physics and the people purveying these models as if they had some physical significance are not physicists – they are just mathematicists.