The thing that is often missed in debates about entropy and Universe is that the classical notion on entropy is not compatible with General Relativity. Richard Tolman almost 100 years ago proposed an extension that was compatible.
One of the consequences of that extension was a possibility of a cyclic universe. On expansion one sees that classically defined entropy increases but then it will decrease on contraction.
These days that work is pretty much forgotten, but still it showed that with GR heat dearth of the universe was not the only option.
There is https://en.wikipedia.org/wiki/Conformal_cyclic_cosmology
If I had to bet money on it, I would say it's right, especially in light of things like this: https://phys.org/news/2025-03-ai-image-recognition-universe....
Heat death was never the only option in GR. The field equations always allowed for a big crunch or a big rip.
Yes, but that implies that in GR entropy or at least the value based on the classical definition can decrease.
So apparent increase in complexity can be attributed to gravity.
Sean Carroll today's go-to person for GR has been working at popularizing these ideas (for more than 10 years(!))
https://arxiv.org/abs/1405.6903
>For example, our universe lacked complex structures at the Big Bang and will also lack them after black holes evaporate and particles are dispersed.
See my comment below for link to Scott's preview.