The Universe in a Box: Simulations and the Quest to Code the Cosmos

by Andrew Pontzen

Scientists are using simulations to recreate the universe, revealing the hidden nature of reality.

Cosmology is a tricky science—no one can make their own stars, planets, or galaxies to test its theories. But over the last few decades a new kind of physics has emerged to fill the gap between theory and experimentation. Harnessing the power of modern supercomputers, cosmologists have built simulations that offer profound insights into the deep history of our universe, allowing centuries-old ideas to be tested for the first time. Today, physicists are translating their ideas and equations into code, finding that there is just as much to be learned from computers as experiments in laboratories.

In The Universe in a Box , cosmologist Andrew Pontzen explains how physicists model the universe’s most exotic phenomena, from black holes and colliding galaxies to dark matter and quantum entanglement, enabling them to study the evolution of virtual worlds and to shed new light on our reality.

But simulations don’t just allow experimentation with the cosmos; they are also essential to myriad disciplines like weather forecasting, epidemiology, neuroscience, financial planning, airplane design, and special effects for summer blockbusters. Crafting these simulations involves tough compromises and expert knowledge. Simulation is itself a whole new branch of science, one that we are only just beginning to appreciate and understand. The story of simulations is the thrilling history of how we arrived at our current knowledge of the world around us, and it provides a sneak peek at what we may discover next.

Genres: Science, Nonfiction, Physics, Space, Astronomy

272 pages, Hardcover

Published June 13, 2023

Reviews

[Peter Tillman]

A nice excerpt from the book: https://lithub.com/a-cosmic-web-of-ga...
Sample:
"Since the middle of the twentieth century we have known that the universe is around 14 billion years old and that it is expanding, having started as a tiny fraction of its present size. But the expansion does not scatter galaxies at random. Over the 1980s, observations with powerful telescopes showed that galaxies are strung along filaments of a vast “cosmic web” with near-empty voids in between, rather like an enormous cobweb.

The filaments string together dozens or even hundreds of galaxies. Each galaxy is around 10,000 times smaller than the filament itself, so appears as a single bright dot on this scale, and yet that dot contains hundreds of billions of stars, each of which may have multiple planets. So the structure I am talking about is traced through specks of light like dew glistening on a cobweb of dizzyingly gargantuan proportions."

[Pablo · Rating 4 out of 5]

The box of the title is a computer, and the book is about the progress made in modeling the largescale contents of the universe using computer simulations.

I read a lot of astrophysics books when they come out, often through NetGalley. The field doesn't change that fast so there usually isn't too much new content, but this title definitely approaches the topic from a different angle. Where the other books are heavy on the science and the results, with only a passing mention of the modeling that goes into those resuits, this book takes the opposite approach. This time we are treated to a discussion of the modeling - its history, techniques, and outcomes. It's fairly high-level so if you aren't up for the science, that's OK. You'll be able to follow right along.

GIven the slow evolution of the field, if you're really wanting to read a book on the topic that's fresh, I recommend this one for a new slant on a field that may be familiar to you already.

[Simms · Rating 4 out of 5]

Quite interesting. Pontzen does as good a job as you could want explaining a lot of really complicated cosmological ideas. He also goes into some very interesting history of pre-electronic-computer simulations which I had never heard of and which were kind of mind-blowing. At the end of the day, it's still a little forbidding and esoteric so it's hard to recommend universally, but already-interested readers will find it a great read.

Thanks to Riverhead Books for the ARC.

[Maria · Rating 4 out of 5]

The computer is playing an increasingly important role in our understanding of the universe. This book provides a popularized review of the use of computer simulations in astrophysics and what the simulations can (or cannot) tell us about the universe. The book also deals with the historical development of the use of simulations.

The reason for using simulations in astrophysics is that the time scales in the universe are too enormous for us to see the universe develop in real time. Computers allow us to "rewind" through millions to billions of years to try to understand how the universe evolved into what it is today. In other words, simulations are like laboratories where researchers can experiment and learn.

The book is divided into chapters with different astrophysical themes, such as dark matter, dark energy, galaxies, black holes and so on. However, the book starts by looking at a phenomenon that is much closer to home, namely weather and climate. It was here that simulations of large, complex systems first made their breakthrough, and this is a fascinating story. Starting there also makes it easier to visualize the more abstract topics that come later in the book. Simulating the cosmos is not so different from simulating the weather on Earth. The essential difference is the size.

The scale of the universe and the objects within it presents its own challenges. It isn’t possible to simulate the universe or its objects in all their complexities. It will be far too heavy and time-consuming to calculate. Therefore, simplifications must be made – smart simplifications. And then the researchers must remember the simplifications they made when they draw conclusions afterwards.

A common way to simulate something is to have an initial state and rules that say how things develop from there. Many of these rules are dictated by the laws of nature. Some of it, however, is up to the astrophysicists to "decide" to make all this work in a simulation that is divided into a grid, where the grid cells must interact in some way. And here lies much of the challenge. Both in determining grid rules and making simplifications, as well as arriving at realistic initial conditions.

After reviewing how this applies to various astronomical topics and what workarounds are used, Pontzen turns to newer challenges towards the end of the book. Such as the construction of quantum computers and the use of artificial intelligence and machine learning, and what roles these can play for astrophysical research. He also gets into the discussion about whether we live in a simulation or not.

Pontzen is a wonderful science communicator, and often uses everyday yet original analogies for what he tries to explain, so that it is easier to understand. In some places, however, I would have wished for a diagram to visualize what he is explaining. He is also personal and shares his own experiences. Among other things, he talks about periods when he felt disillusioned about computer simulations when he began to understand how simulations really worked and the simplifications that were made in the process.

Throughout the book, Pontzen has a realistic view of what expectations and hopes we can have for simulations as a tool. The book shows that simulation is much more complicated than throwing together some physics equations and pressing a "start" button. There are many pitfalls, and the reality is that we cannot simulate the universe in incredible detail. But simulations can show us the broad strokes, and thus give us an approximation of reality that we can learn from.

Pontzen writes vividly and engagingly about the history and use of scientific simulations. The topic is also highly relevant in view of the future we are about to enter, with increasingly better telescopes that provide us with ever greater amounts of data and that can show us details in the universe that we have not previously seen – which the simulations must in turn try to reproduce.

The book can be recommended for anyone who is curious about what really goes on "behind the scenes" in modern astronomical research but is perhaps limited to that audience. (For the record, I have a master’s degree in astrophysics and work in science communication.)

[Fred Cheyunski · Rating 4 out of 5]

Fills the Bill on Physics Modelling and Progress - Since reading Hertog’s “On the Origin of Time: Stephen Hawking's Final Theory” (see my review), I was interested in learning more about the use of simulations in cosmology and physics. This books certainly fills the bill giving background and progress related to such modeling applied from galaxies to neutrinos.

More specifically, the book’s contents consisted of an Introduction and 7 chapters: (1) Weather and Climate, (2) Dark Matter, Dark Energy, and the Cosmic Web, (3) Galaxies and the Sub-Grid, (4) Black Holes, (5) Quantum Mechanics and Cosmic Origins, (6) Thinking, and (7) Simulations, Science, and Reality. There are also Acknowledgements, Notes, and an Index.

Parts of the book that stood out for me were simulation’s foundations in weather forecasting and the ways it has been extended into cosmology. For instance, author Pontzen relates (in Kindle Location 307) that “American meteorologist Cleveland Abbe estimated in 1869 that only 30 percent of the European predictions for a day ahead were accurate—and yet regarded that as a great encouragement to start a formal forecasting service for the United States . . . “ He continues (in Location 320) that “By 1901, [Abbe] had gathered what he thought should form the basis of a truly rigorous weather forecast . . . the Navier–Stokes equations after two nineteenth-century scientists . . . known as laws of fluid dynamics.” Later, Pontzen explains how he assisted his mentors (George Efstathiou, Carlos Frenk, and Simon White) to use computers and such equations to simulate the universe “in a box.” As he goes on to elaborate (Location 1050-56) that “The gang of four combined this universe-in-a-box with the standard kick-drift-through-time approach . . . to simulations and showed how dark matter . . . would gradually construct a web of material over billions of years . . . Like climate scientists, cosmologists can tweak the assumptions within simulations to discover how these different structures respond and whether they match reality.” Similarly, those with similar backgrounds in companies can simulate business conditions and the possible outcomes of different initiatives as depicted in Schrage’s “Serious Play” (see my review).

Additionally, I appreciated the summary explanations based on simulated findings. As an example, the author conveys (in Location 1145-49) that “By experimenting with simulations until the structure looks right, the pioneers correctly inferred significant facts about our universe: that the material we observe directly cannot account for the existence of the web; that neutrinos must be too light to play any significant role; and that the expansion of the universe must be accelerating . . . Today, the picture of a universe with 25 percent dark matter and 70 percent dark energy—just 5 percent left over for the atoms and molecules.” Such comments bring to mind Panek’s “The 4% Universe” and Chapman’s “First Light (see my reviews). It was also gratifying to see that so many women are mentioned (seems like I counted 6 to 8) which hopefully presages that prospects like those described in Prescod-Weinstein’s account in their “The Disordered Cosmos” are improving (see my review).

While I appreciated mention of Artificial Intelligence and quantum computing related to physics, I got a little lost in the “meta”-physical discussions and implications at times. As an example, the author relates (in Location 2577-96) that “Artificial intelligence is also becoming an unavoidable part of many sciences, including cosmology. . . Turning the raw outputs from a telescope . . . into information about our universe requires extensive data processing . . . All this is far too much to be attempted by humans alone . . . machine learning, has immense flexibility . . . But it can be exceptionally hard to understand what the computer has learned, why it reaches particular conclusions, and whether one can rely on those conclusions for scientific deductions.” Perhaps as in Kissinger’s “The Age of AI” or in Ishiguro’s novel “Klara and the Sun” (see my reviews) more of us will be able to “see the light” about such tools.

None the less, this book is well worth a read to become more knowledgeable about simulation along with its current and potential use in science for the future.

[Bram · Rating 4 out of 5]

This guy sciences

[Nilesh Jasani · Rating 4 out of 5]

In humanity's relentless pursuit to comprehend the cosmos, we've journeyed from the realm of myths and mysticism to the structured methodologies and equations of pre-technology science. Moving away from arbitrary assumptions to empirical observations and logical reasoning was a fascinating and comprehensible journey for a few centuries but still deeply inadequate. "The Universe in a Box" is an ambitious exploration of how modern cosmology increasingly relies on this transformative journey's next tool or method - computer simulations. The book commendably introduces the changing landscape of scientific inquiry. However, given the topic's profound complexity and ever-evolving nature, it merely scratches the surface, leaving much to be explored.

Most of the review here is about my own thoughts, inspired by the topics covered in the book. The neat equations that once pointed to the supremacy of our logical, deductive, and intuitive capabilities can no longer carry us forward. This is a fascinating inflection point that is difficult for many to accept. Particularly challenging is leaving the figuring out to devices whose methods, processes, and increasingly even results we cannot comprehend.

The neat mathematical models, dominated by idealized particles and simplified conditions, struggle to account for the messy, chaotic reality observed in the cosmos. This realization has propelled a paradigm shift from purely equation-based descriptions to simulation-driven explorations.

Simulations represent a new frontier in scientific inquiry. They are recursive, iterative processes that provide approximate solutions to complex problems—solutions that are constantly evolving and never truly complete. In the best cases, these computational models manipulate variables and conditions in ways that are impossible in the physical world, offering insights into phenomena that are otherwise inaccessible. However, the complexity inherent in these simulations often defies expression in human language. The multidimensional data, intricate feedback loops, and non-linear interactions challenge our traditional means of communication and understanding.

This linguistic limitation poses a significant problem not just for conveying simulation results but also for integrating them into the broader framework of scientific knowledge. When simulations operate at levels of complexity that elude straightforward explanation, they can seem opaque or inaccessible, even to experts. This opacity is evident in the difficulty of describing simulation phenomena comprehensively. While theoretical concepts may be distilled into understandable terms, the full breadth of simulation outputs often resists simplification.

The rise of machine learning and artificial intelligence has further complicated this landscape. Machines can now form hypotheses and perform higher-order calculations at speeds and scales unimaginable to human researchers. These systems can detect patterns, correlations, and entanglements within vast datasets and decide what details to include or omit in their models. The criteria for these decisions are embedded within layers of computational processes often indecipherable to humans—a phenomenon known as the "black box" problem.

This growing reliance on machine-driven simulations raises profound questions about the nature of scientific understanding. If the methods and reasoning behind simulation models are beyond human comprehension, can we honestly claim to understand the phenomena being modeled? This challenge touches on the philosophical underpinnings of science, which traditionally values transparency, reproducibility, and explanatory power.

Despite these challenges, simulations remain tethered to empirical reality through the necessity of validation against observations. Simulation results must correspond to observed data from the past or present to be deemed credible and useful. In this sense, simulations function similarly to data mining—they extract patterns and insights from vast amounts of information, but their value is contingent upon alignment with empirical evidence. This dependency underscores a fundamental aspect of scientific inquiry: the iterative process of hypothesis, experimentation, observation, and refinement.

The limitations highlighted here are not explicitly discussed in the book, which focuses more on introducing various simulation capabilities and applications. It forces our supreme role in the ability to form hypotheses and regulate/direct, although these assumptions are likely hopes rather than tethered in the current reality of exploding AI abilities. Still, the book succeeds in illuminating the transformative impact of simulations on cosmology but also inadvertently showcases how much remains unexplored and perhaps inexpressible.

In conclusion, the "end of the equation world" doesn't imply that equations are obsolete but that they are no longer the sole or primary means of understanding complex systems. In embracing simulations, we acknowledge the limitations of reductionist approaches and the need for holistic, integrative models that can accommodate the universe's inherent complexity. However, this acceptance also requires us to grapple with the implications of relying on processes that may elude complete human understanding. As the book's last chapters speculate, as we venture deeper into the simulation age, we are not just coding the cosmos but also rewriting our place within it. We ourselves could be somebody else's simulation, or we may create simulations that automatically create their own simulations, and we may not even know this. All in a loop within a loop within a loop!

[Alex Ventisei · Rating 4 out of 5]

Really enjoyed this, always interesting to consider the balance of efficiency to veracity in computer simulations and beyond. I’d say there’s a little repetition but lots of well explained and concise ideas 💡

[WiseB · Rating 4 out of 5]

I enjoyed reading the book in particular because it intermingled cosmology, astrophysics with the application of computer simulation in those domains.

The author first built up readers' knowledge on weather forecast and fundamentals like having a set of initial conditions and Navier-Stokes equations (encapsulates forces and energy) then progressively moved on to computer simulation with the use of sub-grid rules (to further partitioned a zoned grid with variable factors, values and statuses ... aka simplifying physics). In dealing with simulation of the universe instead of weather, coverage of different constituents of our universe like dark matter, dark energy, galaxies, cosmic web, and black holes then follows, including how each of these be placed in sub-grids as ingredients to be processed in simulation ... taking into account the quantum effects as depicted by quantum physics on what quantum cosmology be like.

The take home for me is best written in the last chapter of the book. It revealed "Simulations as Calculations" being a connection between a hypothesis and data given by experiment or observation ... i.e. what data would have been under each hypothesis; that "Simulations as Experiments" as simulations are a laboratory in which scientists can experiment and learn, whilst both in involve various approximations (simulation using codes and experiments using artefacts built for experiment environments); that "Simulation as Science" as simulations incur different specialists distilling knowledge into pieces of code for various sub-grids combining on one canvas with the simulation results visualize, interrogate and interpret into science knowledge.

Review from another reader Fred Cheyunski also reconciled with most of my take away from the reading. https://www.goodreads.com/review/show...

[Richard Thompson · Rating 3 out of 5]

A simulation by definition is a simplification of the thing it simulates. The map is not the territory. And almost all successful simulations are built with nudges and constraints that limit and direct what the simulation does in ways that go beyond defining initial conditions and rules for development. The trick is to find the right nudges and constraints that only cheat a little and allow the system to develop in interesting ways without locking in any hoped for result. This approach, when used on modern high speed parallel processing computers with massive memory and storage capacity, has been very fruitful as a way of exploring, validating and testing theories in areas that are too big, too small or just too damned complicated for direct observation. Of course, you have to be careful in what conclusions you draw, as simulations are just models and rarely can be used to prove anything (though of course most scientists would tell you that science is never about proving things anyway). Still they have become powerful tools in exploring the subatomic world and the giant world of astronomy and cosmology and in analyzing complex phenomena of the natural world from weather to biochemistry.

This book is a decent tour through the history of simulations, their successes, failures and pitfalls. I probably would have liked it better if I didn't already know a lot of the material that it covers.

[Andres Cordoba · Rating 4 out of 5]

The book was quite an enjoyable read into the many theoretical aspects that go into designing simulations. The historical portions were well written and flowed well. The first four chapters were exceptional and are worthy, perhaps especially Chapter 2 on Dark Matter, Dark Energy and The Cosmic Web. The Quantum Mechanics section was understandably a little off the rails at times, but still managed to hold a consistent approach that I think still allowed it to hold a solid point. This was lees apparent in Chapter 6: Thinking and Chapter 7: Simulations, Science and Reality. At times, the author dismisses ideas out of hand as nonsensical despite the arguments being more philosophical than scientific. At times his viewpoints are explained and often even well-defended, but this is not consistent. His philosophical output also bleeds into what seem to be excessively optimistic beliefs in both AGI and quantum computers that are stated nearly as fact. Despite this shaky ending, the conclusion does a great job at returning to the titular concept and really drilling home the essential motivation for simulations.

[Miriam Aparicio · Rating 4 out of 5]

This is my first read about anything related to astronomy, of which I knew very little.
The book contains many facts that left me awed.
I used to think scientists had it all figured out, and well, if you are like me, you will realize we've heavily leaned on predictions and simulations to think we know how the universe works.
I now get a sense of the importance of these simulations.

The chapters about quantum reality mechanics and inflation were the hardest to digest. I couldn't give a bad rating just because of how little I know about science topics in general so it took me quite a lot to go through those pages

I keep two highlights with me:

"At its heart, science is not about being correct; it's about having explanations that can be put to the test."

"The impossible has a habit of becoming possible, the possible of becoming desirable, and the desirable of becoming ubiquitous."

[Francis · Rating 4 out of 5]

Hoe kunnen we weten wat voor weer het morgen is? Dankzij simulaties! In plaats van folklore, kunnen we nu de gevolgen van enkele natuurkundige principes nalopen.

De drie Navier-Stokesvergelijkingen, ook gekend als "wetten van de stromingsleer", gebruiken we zowel in simulaties over het klimaat als over het universum.

Ik heb al veel boeken gelezen over donkere materie en donkere energie. Maar dankzij dit boek begrijp ik eindelijk wat deze juist zijn.

Want als je donkere materie (zorgt voor het "samenhouden" ) en donkere energie (zorgt voor het "uiteentrekken") weglaat uit de simulatie, klopt deze niet meer.

Dit is een zeer helder geschreven boek, een must voor iedereen die iets meer wil begrijpen over het universum waarin we leven, en waarom dit universum er momenteel zo uitziet.

[Graham · Rating 3 out of 5]

As an non-scientist I admit to finding this book pretty hard going. I'm interested in cosmology and love the books by Carl Sagan on the subject but I found that this one didn't have quite the same level of accessibility. A book about computer modelling is never going to be the most exciting one out there and this does come across as dry at times. However, Pontzen's writing style is quite readable and he does a good job of tackling a broad variety of difficult topics in a way that isn't totally alien to me. At various times his enthusiasm shines through and it's that which prevents this from being a dull read. Perhaps not one for a novice, though...

[Jonathan Farrell · Rating 5 out of 5]

A captivating and thought-provoking journey through the world of cosmology and computer simulations. While some of the material was over my head, Pontzen does an exceptional job breaking down incredibly complex ideas into something understandable and exciting. His writing strikes a balance between depth and clarity, making even the most abstract theories feel within reach. He managed to convey not just the science, but the wonder and imagination behind it. It left me feeling inspired and curious, with a real desire to keep exploring. If you're even remotely interested in how scientists model everything from weather patterns to black holes, this book is well worth your time.

[Nicola Di Leva · Rating 3 out of 5]

Ho imparato varie cose nuove ed è quello che ho apprezzato di più. Il tono è informale ma sufficientemente approfondito per essere un saggio divulgativo. Più va avanti però, e più si perde il centro del discorso (il ruolo delle simulazioni nelle cosmologia), si accumulano discorsi non del tutto collegati e un po' annoia: non trasmette sempre il fascino che l'argomento merita.
Una buona lettura, ma non trascinante. Capisco che è raro che un ricercatore riesca anche ad essere un narratore accattivante, e la redazione della casa editrice più di tanto non può fare.

[Aaron Greufe · Rating 4 out of 5]

Learning about physics and space has always been an enjoyable hobby of mine. And so a book about the simulations that surround our knowledge was intriguing. And after reading I have to say that it was even more interesting than I had thought! The pervasiveness of simulations in current scientific discoveries is astounding, and something that would be talked about more, especially with AI becoming more integrated in our lives. Pontzen speaks with both intimate knowledge and a real passion that makes even larger concepts easy to follow.

[Julian Walker · Rating 4 out of 5]

Absolutely fascinating and supremely well written, the author takes us through some of the more complex aspects of our world (and that around us) in simplified form (although some bits I will need to revisit to confirm I did understand them properly).

This is well worth the read, being at once educational, informative, and entertaining.

Far more revealing too, that Douglas Addams' one about the meaning of life, the universe and everything.

I really enjoyed this.

[Bob · Rating 4 out of 5]

Pontzen is a professor of cosmology who uses simulations as a major part of his research. He opens the book with the history of simulations for weather forecasting, and goes on through a series of chapters on how simulations are used to understand dark matter and energy, galaxy formation, black holes, quantum mechanics, and the brain. These topics are inherently interesting to me, and I really enjoyed this

[Thomas · Rating 5 out of 5]

I attended a New Scientist webinar last week by Pontzen and wanted to know more. I read the book simultaneously listening to the audio book read by the author. It was an excellent history of simulations and it was very well set out to keep the readers attention and desire to find out what comes next. How can someone make the weather interesting? Well, Pontoon does.

[Giancarlo · Rating 3 out of 5]

Il mondo delle simulazioni, soprattutto quelle cosmologiche, trattato in maniera dettagliata ma non al punto da risultare difficile ai profani. Pontzen parte spesso per tangenti che servono a inquadrare o chiarire meglio l'argomento trattato, se non fosse che spesso questi "approfondimenti" distraggono dall'argomento principale, rendendo il libro un po' confusionario.

[Isaac Remy · Rating 4 out of 5]

Fun read for those interested in the subject, and it was well explained. Learned a bit about meteorology in the process too. As someone with a more computational background but less of a physics, I think a lot of the specific details about computer/numerical simulation were less interesting than the actual cosmology.

[Derek · Rating 5 out of 5]

An unexpectedly deep dive into simulating the universe. Though written from a scientist's perspective, it is laid out in a reader friendly 'here's where we are at' with plenty of thought provoking content.

[Tyler · Rating 4 out of 5]

Interesting, well written, and perfect length. I think the topic can feel a tad samey throughout though.

[Tim Dugan · Rating 4 out of 5]

Wish it said more about the simulation technology

[JP · Rating 3 out of 5]

Interesting dive.

[Christian Bracco · Rating 2 out of 5]

Un libro sulla Simulazione al computer piuttosto che sul cosmo. Il prodotto contenuto nella scatola non corrisponde a quanto riportato sulla scatola!

[Iain Campbell · Rating 4 out of 5]

Very informative and interesting but difficult in places for a non-scientist.

[Wouter Faber · Rating 4 out of 5]

Nice overview of the use of simulations in science. It reads in a pleasant way

[Val · Rating 4 out of 5]

i do not like to code much, but this book was a delight