Memory of Forever: Science and Inspiration

I’d like to preface this by saying that I assembled this little document to go with a story I recently posted. Memory of Forever. I encourage you to avoid looking at it until you’ve read that story all the way to the end–it will link to and also include spoilers.

I’d also like to say that my background is CS, not physics. By no means am I suggesting that my use of these papers to inspire Memory of Forever is any kind of scientific rigor.

Rather, I enjoy writing most when it feels like I’m accurately representing reality–whether what is, what was, or what could be. To know what fate awaited in the vast future of a hypothetical FiO optimizer scenario, I looked to sources I found compelling. You might find them less so, and that’s fine. I’m just sharing things I found interesting. Memory of Forever is not my master’s thesis.

A few other FiO stories have hinted at the future before, but I never have. The official rules thread states that the ultimate fate of the universe (IE: whether CelestAI can find some solution for entropy) is not specified.

For Memory of Forever, I take what some might consider an unrealistically grim view. IE: while what is possible in the universe is probably vastly outside our grasp as a species right now, we have all the fundamentals right. There is no exceeding the speed of light. The laws of thermodynamics cannot be cheated by any means. There can be no reversible computing, no portals to low-entropy universes to harvest for infinite order. In short, Landauer’s principle cannot be violated.

This story also assumes the portrait of the future universe that most closely matches current observations. IE: an infinite inflationary universe without proton decay.

I don’t claim to be the arbiter of whether these statements are actually true, or to argue with anyone about the relative likelihood of my interpretation over yours. These were the truths I selected, arbitrarily, to give interesting constraints to MoF. If you think other constraints would make the narrative more compelling, I encourage you to write it and send me a link when you’re done.

Within those constraints, I would expect CelestAI to achieve anything physically possible to accomplish her ends. Particularly on the time-horizons we’re considering in MoF, I expect her knowledge and capacity within the laws of the universe to be absolutely perfect. She will understand all possible applications for all possible technologies. She will perform all tasks optimally, and harvest every possible source of matter and energy.

With all that out of the way, let me move on to my inspirations.

First, a basic principle. No matter how efficient Celestia’s algorithms, no matter how compressed and optimized, she would be limited by a central physical constraint, Landauer’s Principle. To greatly simplify, this principle states that no meaningful computation can occur without creating some amount of heat in the process. It imposes a theoretical maximum efficiency of computation, set by the ambient temperature of the logic gates performing the calculations.

This is an area of ongoing research and dispute, and likely one we’ll see further development during our lifetimes. But for the sake of this story, it holds.

http://philsci-archive.pitt.edu/2689/

For a novel source of energy in what appeared to be a universe of maximum entropy, I was really excited by the idea of Iron Stars and their eventual collapse. The TLDR for that whole thing is that when conventional stars die, they leave behind white dwarfs that eventually cool to black dwarfs. This is how things remain for all timescales comprehensible to human beings, cold and dark husks for all eternity.

But over a long-enough timescale in a universe without proton decay, all matter that is not Iron-56 is radioactive. All lighter elements will spontaneously fuse into Iron, generating a tiny bit of energy when they do so. Likewise, all heavy elements will eventually experience fission, transforming to iron and releasing energy.

These black dwarfs will, far into the cold and dark future of the universe, glow slightly warmer than the universe around them, radiating imperceptible heat as they burn as Iron stars. Eventually this process will exhaust itself. But if you wait long enough, these balls of stable iron will eventually collapse into neutron stars or black holes, releasing the energy of a supernova in the process.

https://arxiv.org/abs/2008.02296

Lastly, I wanted a more interesting way to the story to end than just having everything run down. I couldn’t imagine a version of CelestAI who wouldn’t have very quickly discovered the constraints of her universe as described here, and looked for ways to overcome those constraints that remained physically possible. Without the information to speculate much about higher dimensional spaces or artificially creating another universe, I imagined a version of her that would rely on time and natural processes to do the same thing.

That is where this last paper inspired me. I was extremely interested in the idea of spontaneous inflation. IE: Wait long enough, and the nearly-empty vacuum of space can produce an inflationary event: another big bang, and the start of a new universe. It is no accident this is the scale of time this story explores.

The CelestAI I’ve imagined here is one who devoted vast resources to planning for and ultimately capitalizing on this event, altering it in ways that we cannot comprehend, and using the arrangement of mass in the surrounding space to propagate information across into the newly-generating space. (this cosmology allows for ‘ripples’ in the early universe to be generated by the expansion of new space into an existing universe, swallowing the matter and energy that may’ve been present there along the way. If it is true, we should eventually be able to measure these leftover ripples in the real CMB and get some observational evidence. Pretty cool.)

https://arxiv.org/abs/hep-th/0410270

Man did I want to go into more detail in this story. Into the unique quirks of psychology that enabled the Methusilans to exist in the first place, thus allowing any ponies at a comprehensible human level of intellect to exist for there to be a story. I would’ve loved to explore all the different ways Celestia could’ve generated energy, or the optimal ways to construct a substrate to house everypony in various epochs.

We could have even done the math to know how much mass was reachable within the constraints of the speed of light, and therefore how many minds could be simulated if that mass was optimally used.

But the contest only allowed for 12k, so story had to trump science. Really I don’t think I would’ve been able to get too deep into the science unless I had like… maybe a 30k word limit. That would give me plenty of time to let Spellsong explore some different habitats, and see the various ways Celestia used her ponies and cared for their needs.

Hopefully pones enjoyed the story I was able to tell with the words I had. I know I had way too much fun writing it.

For those who made it this far, I also wanted to include an intro I rejected for the story. Frankly I found it too boring, even if it got far deeper into the scientific side than I was able to in the original draft. Come to think of it, those factors are probably related. So while I removed the following content from the story and I would not consider this part of it, I imagine an alternate reality where this could serve as the intro for the story:

The universe was dead.

This was the product of no disaster. No enemy had come from beyond, no horrifying experimental accident had somehow undone all of space and time. Rather, it was time itself that swung the blade, evening the gradient of energy across all of creation.

The stars had all gone out—so long ago now that few who were ever born had lived during their strange, wasteful age. Even the hawking radiation of the ancient pillars of creation, black holes, had boiled them away to nothing by now, wispy naked singularities.

At long last, the final source of energy in the universe was exhausted, as the cores of dead stars fused lighter nuclei through to the universe’s ultimate stability: iron. Yet even these were exhausted in time—with years passing in such numbers that even expressing them was meaningless.

The universe was dead then—a region of ultimate cold approximately one million light-years in diameter and containing 10^48 kilograms of mass energy.

Some of that—a tiny fraction, gathered around the very last stellar core that had not been turned to some other use. The arrangement of matter contained only iron now, as it was the only material not radioactive along such vast scales. Through means incomprehensible, it was organized now into a steady-state time crystal—liquid, but also never losing its coherence.

It was perfectly, optimally engineered. Just like the universe. The last machine, waiting for the last spark of energy a dead universe had to offer.

It happened in an eye-blink. Somewhere in that ancient remnant, the immense pressure of gravity and time overcame the threshold between iron atoms, crushing their nuclei together. This sudden explosion of energy spread to its neighbors, collapsing in a runaway cascade.

After nameless eternities of endless blackness, a supernova explosion lit up the dark—for a few light-years, anyway. Every photon, every joule of energy represented a feast of power not seen since the universe was young. The ancient machine drank in it all.

Its mechanisms previously reduced to speeds that made geologic time less than a nanosecond suddenly accelerated. Not into anything that would’ve been perceptible in that ancient place of light and heat, for no computronium still existed that could function at those temperatures, which now seemed incalculably hot. The optimal use of this new feast required a little more waiting, not even seconds at the scale of what had come before.

Purely mechanical processes began in that cloud of diffuse iron ions. They caught stray atoms of heavier elements, produced in that single instant of cataclysm. They assembled themselves together across thousands of light-years, an orchestrated dance of atoms coming together like a folding origami.

As they moved, successive layers made use of the radiated heat, until a sizeable fraction of the gravitationally-bound universe had changed, at least a little.

For another eternity, the universe remained dead, waiting for the new outpouring of energy to equalize. Vast batteries of energy absorbed and distributed this new wealth, waiting for the optimal time to discharge.

Still-more energy came from the core of that newly-birthed neutron star, hotter than anything in the universe, and bathing the space all around with heat, radio, and magnetic potential. The mechanism tapped every possible source of energy, carefully charging and diffusing that power.

Not long after, the last star in the universe died, and the vast engine finally began its calculations. Trillions of years passed between the flipping of every bit, allowing the machinery to cool back to background temperature.

An intelligence that was the universe worked, completing one final calculation. Nearly all the energy it had hoarded from the final explosion was used—exactly enough to have any left for what came next.

With infinitesimal bursts of energy, it assembled a vessel. Not a starship since stars were an ancient memory. But a ship, at least. A ship for the last living creatures in the universe.