Even If This World Isn’t a Simulation, It’s Not Much Different
Is our universe, filled with endless possibilities, merely a part of some supercomputer’s program?
There’s been a lot of discussion lately about whether our world is a simulation. Take Elon Musk, for instance. Beyond his ventures into rockets, electric cars, and his controversial presence on social media, he’s also a believer of the “simulated universe” theory. Musk has candidly stated in public that he believes the likelihood of us not living in a computer simulation is one in billions. This bold statement is not just a disruptive view in the tech world; it’s a profound challenge to our philosophical and perceptual understanding of reality.
The notion of a simulated universe isn’t unique to modern technology. In fact, it has a long and rich tradition in the history of philosophy. Centuries ago, the French philosopher Descartes, with his famous assertion “I think, therefore I am,” introduced fundamental doubt about the nature of reality. British philosopher Putnam’s “brain in a vat” thought experiment further delved into the possibility that our understanding of perception and reality might just be illusions.
In the contemporary era, Nick Bostrom, a professor of philosophy at Oxford University, has proposed a startling hypothesis through mathematical derivation: the likelihood of us living in a world simulated by advanced computers is much higher than we might think. His paper “Are You Living in a Computer Simulation?” has sparked widespread discussion in the academic world. It makes one wonder if Musk’s views were influenced by Bostrom’s theories.
However, the core of this article isn’t just about pondering whether we truly live in a virtual world. A more urgent and intriguing question is: If there are indeed differences between our “real” world and the best theoretical virtual simulation, what are those differences? This question not only concerns technology and philosophy but also our understanding of life, the universe, and the self.
If We Were to Simulate a World
Simplified Simulations
We’re all familiar with computer-developed simulation programs. For instance, the world depicted in “The Matrix” is created by a supercomputer.
Computer simulation programs have become an important tool for exploring and understanding the world around us. The video games we play can be seen as simplified simulations, allowing us to experience and learn about various dynamics of the real world in a virtual environment. “Warcraft,” which accompanied my student days, was not just an entertainment tool; it also simulated aspects of war strategy and tactics. Recent games like “The Legend of Zelda” and “Genshin Impact” let players explore worlds filled with fantasy elements, experiencing the charm of role-playing games. And the “Civilization” series, which can keep you up until dawn, compresses the entire human historical process into a game, helping players grasp the evolution and conflicts of civilizations.
Speaking of the pinnacle of simulation, Microsoft’s “Flight Simulator” stands out. This game nearly recreates the entire Earth, along with its climate systems and geographical features, right before the players, offering an unprecedented flying experience. The game’s data volume is a staggering 70PB (70,000,000GB).
Simulated Worlds in Real-World Applications
In practical applications, we also construct simulated systems. While the focus may currently be on low-precision floating-point operations for AI, NVIDIA has not abandoned high-precision floating-point computation (FP64), crucial for simulation.
Simulation plays a key role in the development of autonomous driving technology. Testing new self-driving algorithms in real environments is not only costly but also risky. As a result, many companies have established dedicated simulation departments. In these virtual environments, autonomous vehicles can be safely tested and optimized in a variety of complex scenarios, from bustling city streets to adverse weather conditions.
These simulations not only save a significant amount of time and resources but also provide researchers with an unlimited experimental platform. Here, they can simulate and test various extreme and rare traffic situations to ensure the safety and reliability of self-driving systems.
When I worked at Uber ATG in 2019, the machine learning platform team I supported was part of the simulation department. We supported various machine learning tasks (perception, prediction, path planning, etc.) which required extensive offline simulations. This period allowed me to gain a comprehensive basic understanding of autonomous driving by interacting with various machine learning departments.
Times have changed, and Uber ATG is now a part of history. With Cruise’s decline, Waymo has almost become the last survivor in this field. However, the development of autonomous driving hasn’t stopped. In 2023, the field started using GenAI technology. NVIDIA’s GenAI-based simulation system, unveiled in 2023, is far more powerful than those from the past. This is attributed to both algorithm updates (the mainstream models in 2019 were based on CNN) and hardware upgrades (using NVIDIA 1080/2080 at that time).
At GTC 2023, NVIDIA promoted Omniverse, a universal simulation system. Omniverse supports many applications, including factory and production line planning. Previously, production line planning was a time-consuming and labor-intensive task, involving physical world testing. With Omniverse, many aspects can now be tested in a virtual world.
Meta Human
When my former employer Facebook first launched Avatar, everyone ridiculed the legless virtual Zuckerberg, questioning the outcome of a $100 billion investment. Meta’s stock price also plummeted from $380 to $90 during this period. Despite this, I spent a few hundred dollars on the then-immature Oculus Quest 2, just to experience a previously unknown sensation.
By 2023, tech influencer Lex Fridman used new technology in an interview with Mark Zuckerberg. Although the new generation avatar now only consists of a head, people’s ridicule has turned into admiration due to its realism.
Of course, this technology is not yet widespread due to cost issues, but such problems are always solvable. Perhaps when Oculus Quest 4 is released, everyone can experience a visual version of “The Matrix.”
The Limits of the Real World: Planck’s Constant
The Relationship Between Computer Clock Frequency and Simulation
When discussing computer simulations, a core concept that must be addressed is the computer’s clock frequency. Simply put, clock frequency is the number of cycles a computer processor can execute per second. This frequency determines the speed at which a computer can execute instructions, meaning a higher frequency allows for faster processing of tasks.
For instance, if a processor has a clock frequency of 3.5GHz, it means it can perform 3.5 billion cycles per second. During each cycle, the processor can carry out basic operations, such as arithmetic calculations or data transfers. For complex simulation programs, a high clock frequency means the ability to process large volumes of data more quickly and accurately, thereby providing more realistic and detailed simulations.
Over the past few decades, we’ve experienced chip iterations in line with Moore’s Law, doubling performance every 18–24 months. In the new era, Jensen Huang of NVIDIA proposed Huang’s Law for GPUs, suggesting a doubling of chip performance every six months. However, there’s always a limit to everything.
Planck Time and Length: The ‘Smallest Particles’ of the Real World
It’s highly probable that our time isn’t a continuously flowing river, but rather a string of tightly connected beads, with each bead representing the smallest unit of time. This is the role of Planck time in our universe — it’s the smallest grain of time, roughly just 5.39×10−445.39×10−44 seconds short. This concept challenges our traditional view of infinitely divisible time, prompting us to imagine a universe composed of these tiny grains of time. The distance light travels within a Planck time is known as the Planck length, approximately 1.6×10−35 meters.
This description might be too abstract, so let’s imagine it this way:
Planck Time: If we expand one second to match the age of the universe since the Big Bang, about 13.8 billion years. In this context, Planck time, in the expanded timescale, is an extremely small fraction of the original second, about 2.35×10−26 seconds. This expanded Planck time is also roughly one ten-millionth of the shortest time currently measurable by humans. Even this scale of magnification might be hard for us to comprehend…
Planck Length: Imagine expanding a standard 1-meter ruler to the scale of the universe. Suppose this ruler is enlarged to match the diameter of the observable universe, about 93 billion light-years. In this scenario, the Planck length on the expanded ruler would be about 1 centimeter.
Reality and Simulation: The Discrete Dance of Time
This “granular” treatment of time and space astonishingly resembles scenarios in video games or computer simulations. In these virtual worlds, time is also divided into segments, each representing an instant in the game or simulation. Just like Planck time in the real world, each “frame” in these virtual worlds is foundational to the entire narrative. This is not just a technological coincidence but a similarity in how both worlds — real and virtual — handle time.
Planck Time and the Computer’s Beat
If a computer’s clock frequency is likened to the heartbeat of its heart, then each beat represents the smallest attainable time segment in the virtual world. These time segments in computer simulations play a role similar to that of Planck time in the real world. They are the smallest units of time in which events occur in their respective worlds, determining the level of detail we can observe in changes. Whether in the vast universe or within the precise circuits of a computer chip, time seems to be composed of such tiny units, revealing the astonishing similarities between these two worlds.
Further Thoughts: The Nested Simulation Worlds
If our world itself is a simulation, how do we comprehend the relationship between this layer and other possible simulated layers?
The Upper Limit of Time Frequency in Simulation
As we discussed earlier, a computer’s clock frequency sets the upper limit for its basic operational speed. This means that no operation, no matter how fundamental, can exceed this frequency. In the context of a simulated world, this principle applies similarly. If our universe is the result of a simulation by a higher-dimensional entity, then the time frequency in our universe — our “Planck time” — would be determined by the “computer’s” clock frequency of that higher-dimensional world.
Surpassing the Universe’s Time Frequency
When pondering the hypothesis that our universe might be simulated, an intriguing question arises: can we surpass the set time frequency limit of our universe? This question takes us into a realm intertwined with philosophy and science.
If our universe is a product of a higher-dimensional simulation, everything within it — including the passage of time — is confined to the rules of the system simulating it. This implies that the rate at which we experience time passing is actually set by the “computer” simulating our universe. In this framework, surpassing this time frequency might be akin to trying to transcend the fundamental rules of our universe — seemingly impossible with our current understanding.
This also leads us to consider that our understanding of our own universe might be limited to the parameters set by the simulator. Whether we can break through these fundamental rules, despite our technological advancements, remains an unresolved question.
Building Deeper Levels of Simulation
If we attempt to create a next-level simulated world, its basic unit of time would be limited by our computer’s clock frequency. For example, if we use a 100GHz processor to simulate a world, then that virtual world’s Planck time would be 1/100G seconds. This means the world we simulate would be limited by our technological capabilities, with its flow of time not exceeding our set limit.
Technological Limitations in Simulation and the Meaning of the Universe
This mode of thinking not only reveals the limitations of technology in simulation accuracy but also offers a unique perspective to understand the temporal nature of our own universe. If our universe is indeed a simulation of a higher-dimensional entity, our understanding of time and space might just be part of a larger-scale simulation. This viewpoint challenges our fundamental perception of reality, opening a door to infinite possibilities.
In Conclusion
While delving into the profound theories of simulated universes, we can’t help but ponder that if true simulated worlds can nest many layers, it’s highly unlikely that we are at the “top layer.” Theoretically, the Planck time of a universe at the top layer could be infinitely small, whereas our universe has a Planck time of 10−44 seconds. Mathematically, it can continue to become infinitely smaller.
If we indeed live in a simulated world, how can we prove this? If not, given the existence of Planck’s constant, how is our world different from a simulated one? These questions provoke deep contemplation about the nature of our universe.
Regardless of the answers, our lives remain real, and our pursuit of understanding the universe and a good life doesn’t change. The meaning of our life is defined by ourselves. We should immerse ourselves in the performance on stage, even if we know it might be an illusion.
In the future, technological advancements may unravel these mysteries or bring more unknowns. We should remain curious, continue to explore the mysteries of our universe, and in this process, constantly find and define our own meaning.
Zhuangzi dreamed he was a butterfly, or was it the butterfly that returned to reality? Zhuangzi might have found it difficult to replicate his experience, but with over two thousand years of technological progress, we can now more easily replicate virtual experiences.