Biology is not a simple linear physical equation. We have buffers. There's a lot of balancing mechanisms inside and altering your food supply can change the state your body operates in. Or if I may try this naive comparison, when I drive my car and use brakes, I slow down up until a point the engine starts to reaccelerate. The ECU adapts to reallocate more energy because going too slow would "kill" the engine. For a computer, fasting would be like reaching memory pressure so high, your OS garbage collects/flush caches and then programs suddenly runs better.
ps: not that I like all of Levin's work, I do admit that his tendency to abstract into grand ideas is not helping, but I think the morphogenesis/gap-junctions part of his work is still very interesting and straightforward.
Yes, biology is not as simple. With mTOR inhibition fasting is not a full story. mTOR is a sensor of food (amino acids) and a master regulator of growth signals. mTOR is inhibited naturally by fasting and artificially by drugs like rapamycin. And rapamycin is currently a winner in life/health extension across species (Matt Kaerbelin, Rich Miller from theDrive podcast). So with rapamycin inhibiting mTOR (or other fasting mimetic drugs) it's enough to only _send_a_signal_ of low energy/food to the biosystem to obtain gains in health and reverse biological time. So your analogy fits best here, as you are only using signals.
While with fasting it is a reduction of actual food/fuel. Garbage collection can be comparable to fasting-induced autophagy, but in fasting it is also an absence of growth signals like IGF-1 that provides gains in health function (though Greg Fahy research on thymus regeneration in humans with a mix including a growth hormone may contradict this).
The majority of Levin's work in developmental biology done in embryos, which are naturally a highly flexible tissue and naturally express high levels of Yamanaka genes, which are now used for epigenetic reprogramming (OSK). Growing eyes on the tadpole's tail is cool, but I want to grow a tooth for myself now.
Sinclair is now using epigenetic regprogramming with OSK (developmental) genes to effectively regrow the severed optic nerve in primates and proceeding to humans. Levin has mouse paws cut off in a similar way and regrows them in a wearable bioreactor with a mix of drugs, which include not only "electroceuticals" (ion channel modulators), but probably also some (growth) hormones or hormone precursors (as he mentioned in a podcast they may use some estrogen precursor).
Yea, they regenerate ~70% of the whole paw with multiple tissues, but the end result seems the same if it would be obtained by OSK reprogramming.
In a recent podcast #300 Attia was again mentioning the rather reliable benefits of caloric restriction. So thermodynamically it still seems ridiculous that by reducing the inflow of energy (at least the one of the most well-known type) the biological (and presumable computational) efficiency is increased at literally no cost.
ps: not that I like all of Levin's work, I do admit that his tendency to abstract into grand ideas is not helping, but I think the morphogenesis/gap-junctions part of his work is still very interesting and straightforward.