Let’s assume that the primary innovation between eukaryotes and prokaryotes is that eukaryotes have the ability to store and compare state information, rather than simply respond to it.
Perhaps it’s even appropriate to state that a defining property of eukaryotes is “memory”.
In prokaryotes, external influences act directly upon internal processes, e.g. RNA translation is directly modified by auto-inducers in quorum sensing behavior.
In eukaryotes, the cytoplasmic space provides a buffer between nuclear processes, organelles and provide a mechanism to modify both incoming and outgoing messaging insulated from the external environment, and allows all kinds of fun flexibility with expression without breaking core nuclear processes.
Mitochondrion specifically are responsible for providing the peptide/protein maps which are expressed via “receptors” on the cellular surface. Instead of thinking of mitochondrion as creating “energy” products, what if we assume those products (e.g. glycogen) are equal weight information messengers, along with every other “neurotransmitter” type, and represent bayesian stop/go messaging for whatever peptide/protein expression that particular mitochondrion is “responsible for”.
An interesting mechanic that would appear in this conceit is that external processes are no longer directly lethal to the cell as a whole, as the cell itself can now “up regulate” and “down regulate” processes as necessary.
If we assume the nucleus represents the “core state” of a cell, and the mitochondria represents “current state”, and the goal of the cell is to maintain homeostasis, then the difference between nuclear and mitochondrial states give us a mechanical way to generate “more” or “less” instructions.
Switching scope, work like this which suggests neuronal mitochondrial density is driven by astrocytic interactions, and we assume that what that astrocyte is doing is encoding information to neurons, then in effect mitochondrial diversity is an artifact not of “energy” requirements, but of receptor density which drives the state comparison mechanics.
Considering how intensely mitochondrial dysfunction is associated with dementias, does mitochondrial disruption represent a loss of ability in differentiating/producing the metabolic responses that ultimately generate behavior (and “external memory”)?