You’ve just booted up a game on a state-of-the-art quantum computer. You’re running 19 superconducting quantum bits on a processor held at near absolute zero. Anticipating its sheer power, you press start and…
Well, you don’t get much. Maybe a smiley face that’s winking and not winking at the same time.
What gives?
Quantum computers are nascent computer processors that promise to solve problems that are difficult or impossible for today’s computers. They use the mathematics of tiny particles, rather than computer logic, to guide their binary system calculations. And while programmers are making basic quantum games that sometimes amount to flipping coins, one day researchers hope they’ll be able to introduce strange new in-game weapons, improve procedurally generated levels, and create far more lifelike in-game artificial intelligence.
Companies have recently begun releasing real, rudimentary 20-or-so qubit quantum machines that programmers can actually access. These error-prone devices, created by the likes of IBM and the startup Rigetti, are just entering a new era in which they may finally be able to do something useful, beyond hard-to-grasp physics experiments.
Even in these early days, some physicists and programmers are hoping to generate excitement about these machines in the same way programmers first got people excited about classical computers: by building basic games.
“The first game for a normal computer was Spacewar!, which was made as a demonstration to show what the computer can do,” James Wootton, a quantum computation researcher at the University of Basel in Switzerland, told Gizmodo. “I think we need quantum Spacewar!”
Wootton himself is a self-proclaimed “Nintendo fanboy” (but not a “gamer,” he says) who has written his own recent history of quantum games.
Like most people working in the quantum computing field, he started with a degree in physics and studied quantum mechanics, the field that governs the confusing, counterintuitive behaviour of subatomic particles. His quantum computing work applies these behaviours as a new way to abstract problems into their most basic pieces, and perhaps tackle problems regular computers can’t.
So, what can’t a regular computer do? Computer scientists are still figuring that out. Computers are just systems that simplify problems as endless lists of communicating bits—physical systems that either equal zero or one, assuming their values based on the rules of logic. But current research demonstrates that some problems aren’t easily simplified down to logical operations—classical computers have trouble factoring very large numbers, for example. They also have trouble simulating the collective, probability-driven behaviour of subatomic particles. It would simply take too many bits or too much time.
A quantum computer would tackle these problems with qubits—quantum bits—instead of bits. Qubits return the same 0 or 1 that bits do, but take on states in between 0 and 1 during the calculation. So, a quantum algorithm begins by setting each qubit to 0, 1, or somewhere in between. The algorithm runs, then the machine measures and delivers the results. Just before measurement, one or several combinations of qubits are allowed and others aren’t, based on the quantum mechanical effect called “interference.” At measurement time, the machine will spit out one of the possible binary strings. So the outcome of the same quantum algorithm repeated many times on a five-qubit quantum computer would be many five-digit strings (like 00110). Only the strings of binary numbers permitted by the quantum algorithms will appear, and if more than one is permitted, each will appear with a frequency based on those probabilities.
Evangelists say quantum algorithms running on quantum computers will be able to factor large numbers in a fraction of the time of regular computers, or precisely and efficiently simulate the interactions between atoms, allowing computers to dream up new physical materials or molecules. Technology companies like Google and Microsoft hope quantum computers will offer more realistic artificial intelligence, while other businesses hope they will help find optimal solutions to complex problems, like how to best allocate planes to airport gates. Researchers are still figuring out which quantum algorithms demonstrate superiority over classical ones. Even after supposed quantum advantages appear, classical computers usually catch up quickly.
The fact is, quantum computers can’t do much right now. Energy from the outside environment causes qubits to lose their quantum behaviour, turning them into regular bits. The biggest commercial machines (not counting the D-Wave, which is a more limited kind of quantum device) have 20 or so qubits and are still very noisy—they don’t always return the answers they’re supposed to.
In this nascent environment, programmers like Wootton have already built gamified tutorials, citizen science experiments, text adventure games, and puzzle games. Wootton created his first (though not the first) game tackling quantum ideas, called Decodoku, using a classical computer just two years ago. Early quantum machines existed, but there was no way for the public to program or interact with them firsthand. He hoped to introduce players to quantum error correction: preventing the qubit from losing its quantum behaviour to the environment or from returning a 0 when it should have returned a 1. Decodoku is a simple and frustrating game in which coloured numbers appear on the screen and disappear if same-coloured numbers add up to 10. The numbers seem to spawn faster than you can clean them up, much like errors from the environment can overwhelm today’s quantum computers.
Around the same time, IBM put out its user-facing Quantum Experience, which allows even people without a physics degree to write their own programs on a small, five-qubit quantum computer. Wootton began writing games on the IBM platform, which he described as “simple applications that could help people understand the basics of quantum programming.” He called his first real game for a quantum computer “quantum rock-paper-scissors.”
Except it wasn’t rock-paper-scissors at all — it was more like a coin-flip guessing game you’d write for a graphing calculator. The player wants to switch a qubit’s value from 0 to 1, but the computer wants to keep it the same. Both the player and the computer are able to pick one of two mathematical operations that could bring the qubit halfway between 0 and 1, but in different ways — it helps to imagine 0 and 1 as the top and bottom of a sphere, the qubit values during calculations as points on the sphere, and the probability-changing operations as changing the qubit’s location along the sphere. If they pick the same operation, the qubit will flip its value, but if they pick different operations, the qubit will stay the same value. Another qubit determines the computer’s choice: It’s a superposition of both operations or a quantum state of both operations simultaneously. The computer can return only one of these, determined solely by the randomness of quantum mechanics.
A second game of Wootton’s is a multiplayer game called Quantum Battleships, in which a ship sits at the junction between two entangled physical qubits—each qubit represents half of the ship. The game measures and re-entangles these qubits a thousand times to determine the ship’s intactness: a higher percentage of times the two qubits have the same value is a healthier ship, and a higher percentage of times they have a different value is a more broken ship. Bombs change the probabilities for one of the two qubits. But the quantum idea of entanglement can do seemingly illogical things to the probabilities. Since the ship’s damage is based on correlations between qubits, and entanglement can present strange, stronger-than-expected correlations, the game might tell the player the ship took more damage than would be possible without a quantum computer. It’s as if you could put three red “hit” pegs into a two-holed battleship piece.
These are very rudimentary, text-based games. “I really want people to see the program and how it works, say ‘I can do better than this,’ and be inspired to do better than me,” said Wootton.
Other quantum programmers have joined Wootton in producing games on both IBM’s quantum processors as well as Rigetti’s. Physicist Mark Fingerhuth, CEO of a startup called ProteinQure that hopes to use quantum computing to design new molecules for use in medicine, maintains a registry of quantum projects on StackExchange, including quantum games. But they’re all simple. Some are gimmicks—Wootton used a 16-qubit IBM computer to make a superposition of two 8-bit ASCII characters, resulting in a superposition of emoticons. Others stick close to the rules of quantum mechanics or are citizen science projects meant to solve some quantum computing problem by abstracting it or to teach quantum computing topics.
Recently, for example, many have discussed the idea of “quantum supremacy,” wondering when there will be a problem for which a quantum computer can come up with a solution but a regular computer can’t. It hasn’t happened yet, but Wootton devised a game to benchmark quantum computers against each other and against classical computers, called Quantum Awesomeness. It’s a puzzle where a user must call out pairs of adjacent numbered, coloured circles that are most similar in a web of connected circles. The worse the player does, the more the puzzle increases in size and complexity, making the program itself become more resource intensive. It’s the player’s job to keep the game running as long as possible until the puzzles are so tangled that the game is unplayable, causing a game over. This game is a symbolic representation of Google’s quantum supremacy proposal. If a quantum computer will last longer against a terrible player than a classical computer does, this is a proposed demonstration of “quantum supremacy.”
You might wonder what a quantum computer will do for video games, given that the current state-of-the-art quantum games are basically just glorified random-number generators.
Some folks are already dreaming up ideas. This April, Microsoft and the University of Bristol asked PhD students to create a game “involving creative use of quantum principles.” The winning team built a game with quantum weapons, in which the goal is to blow up your opponent. One weapon, for example, was a bomb that could blow up in one of two places (represented by a qubit’s 0 and 1) which isn’t determined until the computer measures the qubit’s state. But before the computer measures the state, the player can entangle the bomb with a second bomb, correlating where both explode simultaneously.
Even many decades from now, provided quantum computers scale up and deliver on their goals, classical computers would likely still perform a video game’s grunt work, like the sound, dialogue, graphics, controls, etc. They would connect with a quantum processor that could introduce new game mechanics or improve the physics engine. These games could incorporate the weirdness of subatomic particles on larger scales. One might have enemies that both exist and don’t exist in a spot before the player tries to shoot them. Or perhaps some future sequel to Portal could require a player to exploit the rules of quantum mechanics, instead of the rules of regular physics, in order to navigate through a level. Maybe there are obstacles you can only cross by applying the right quantum mechanical operations to buttons or entangling sets of switches.
But perhaps the most hyped quantum computing goal is better machine learning and artificial intelligence. A glimpse at the future of quantum gaming may come from a current quantum project related to music. James Weaver, a developer advocate for Pivotal Software, has created a program that composes original music using Rigetti’s quantum processor and its cloud-based programming environment, Forest.
How does it work? Two qubits make up four different notes — let’s say both qubits equalling zero, “|00>” as a quantum programmer might write it, is a C note. The first qubit equalling zero and the second equalling one, |01>, is a D, |10> is an E, and |11> is an F. The programmer can choose the first note, then define probabilities to determine the odds for getting each of the next notes. The quantum computer performs the calculation and picks the next note based on the given probabilities. It continues on, applying the probabilities to the qubits and measuring the outcome, to produce both a harmony and a melody line.
A quantum computer might eventually beat a classical computer at generating these guided probabilities. “Composing and appreciating music is based on probabilities at its core,” Weaver said. “Because quantum mechanics is inherently probabilistic, it’s an interesting way to compose music.”
You can extrapolate this to decades down the line when quantum computers are far more complex and quantum machine learning might lead to truly unpredictable gameplay. Jeff Henshaw, a founding member of Microsoft’s Xbox team and current group project manager of Microsoft’s Quantum Architecture and Computing Group (QuArC), described his ideas for future quantum video games to Gizmodo:
[Quantum machine learning] will give game developers an opportunity to create experiences that adapt to human input over time. In massively multiplayer scenarios, quantum-powered machine learning will be able to analyse the behaviours of legions of gamers and create experiences that challenge us better collectively, while adapting to each player’s unique style of play.
Large numbers of on-screen game-controlled enemies is a great example where another form of quantum advantage can help us: true randomness, powered by nature’s own subatomic behaviours. We’ve all played games where hordes are unleashed and scores of baddies spawn in patterns that feel random at first, but over time give way to programmatic patterns. This, in turn, hurts replay-ability. In a world where truly random behaviours can be informed by quantum processes, we can create environments, and scores of enemies, that feel natural in their behaviours even over infinite periods of play.
Maybe games like Grand Theft Auto could have infinite randomly generated environments and computer-controlled characters that learn of your character’s notoriety and act accordingly. It could make procedural generation even quicker and easier, and apply to elements beyond the appearance and content of the game—perhaps it could generate custom storylines and strange new multiplayer experiences, with a near-real accuracy of detail.
This is all speculation at this point, and getting to these futuristic gaming experiences will take work. “One has to advance the quantum algorithms, and determine in what capacity quantum computers can be used” in ways not accessible to classical computers, Krysta Svore, leader of the Microsoft Quantum – Redmond (QuArC) group at Microsoft Research, told Gizmodo.
Then there’s the hardware challenge. Many physicists think it would take more than 50 qubits for a quantum computer to do something that a classical computer can’t. This task will be something very specific, akin to Wootton’s Quantum Awesomeness game. A quantum computer that could factor numbers would require millions of noisy physical qubits containing built-in redundancies in order to fully account for errors, amounting to perhaps 1000 usable qubits. This would be an incredible engineering feat. Today’s state-of-the-art universal quantum computers have around 20 physical qubits, with promises of 50-, 72-, and 128-qubit devices in the coming year or so.
It’s up to early adopters to push the field forward through experimenting, Talia Gershon, senior manager in AI challenges and quantum experiences at IBM, told Gizmodo. Her team is trying to get people used to thinking with quantum mechanics and has worked with Wootton on a game called Hello Quantum to help users learn the ropes. (Think of a Maths Blaster for teaching quantum concepts.)
A typical first-person shooter won’t be relying on a quantum processor anytime soon. But quantum algorithm developers and companies working on quantum computing are hoping that eventually, this really will revolutionise how we interact with computers, and therefore, video games.
We just need more people learning quantum mechanics and how to program a quantum computer in order to get us there.
“Just go and have a hackathon,” said Wootton. “Hash out some ideas and see what’s possible.”
Leave a Reply