Although Elon Musk has made no clear statements indicating he is pursuing quantum computing, here we investigate his motive and timing should he choose to do so.
Elon Musk is a serial entrepreneur, business magnate and investor who has founded or co-founded no less than seven companies. These are Tesla, Neuralink, SpaceX, The Boring Company, OpenAI, Zip2, and X.Com, respectively.
Always ready to surprise people, especially with his Twitter comments and rants, Musk’s dream of putting humans on Mars and building the best self-driving cars around deserves kudos. Yet, though the naturalized American has grand ambitions with hard engineering problems, what The Quantum Insider is interested in is whether Elon Musk has these aspirations in quantum computing.
Elon Musk’s Thoughts on Quantum Computing
For many years, Musk’s automotive and AI company Tesla has been working on what some consider to be the most powerful computer in the world, a reportedly “best-in-class” supercomputer.
Now, we could just explain all the ways that Tesla is as much — if not more — a software company, as it is a car company with all the cool gadgets inside its vehicles, but that would be defeating the aim of this article.
What we want to get at, though, is how important software is to Tesla and, by extension, to Musk himself. This decade’s Ironman said it best in a Twitter post in February of this year:
“Tesla is as much a software company as it is a hardware company, both in car and in factory. This is not widely understood.”
So, we know that Musk views Tesla as a Silicon Valley software company, rather than a pure-play automotive company.
What is driving (sorry for the pun) Tesla’s software IP forward, in fact, is its self-driving technology. However, this kind of innovation requires large amounts of computational power to manage the multi-trillion parameter networks, something current technology finds difficult to process efficiently.
Musk’s answer to the problem, though? Go in-house, building what you need for a specific task. In Tesla’s case, that solution is the aforementioned supercomputer. Unveiled last year, Tesla’s supercomputer was specifically designed to train self-driving AI technology and take it beyond its current capabilities.
Now, here’s where it gets interesting: the Tesla D1 Dojo chip. This state-of-the-art super chip delivers amazing bandwidth and computational power, much needed to train the neural network that powers Tesla’s AI technology, capable of managing the astronomical amounts of video data captured from its existing fleet of vehicles, used to train its neural nets.
The Dojo does have its critics, though, with its reported capabilities not that of a true high-performance computer (HPC), according to Gartner research vice president Chirag Dekate, whose stance is grounded on the fact it hasn’t been tested using the same standards as, say, Fugaku — the most powerful supercomputer in the world — and other supercomputers.
Built for just one specific task, Musk’s foresight in building the supercomputer to a general-purpose design means it can be used in other technological verticals and applications, these include natural learning processes (NLP) and deep learning (DL).
As Tesla’s data sources grow and its supercomputer turns into a rather less capable imposter of its former self due to the constraints of Moore’s law (the D1 Dojo chip is reported to have fifty billion transistors on it), Musk and co. will have to look at new ways to manage the vast amounts of data its network of cars produces.
This is where quantum computers could help, rather than the HPC capabilities of the current supercomputer Musk tasks with solving Tesla’s processing of its neural networks.
According to a paper published in Nature’s npj quantum information in March 2019, An artificial neuron implemented on an actual quantum processor, “artificial neural networks are the heart of machine learning algorithms and artificial intelligence, […] show[ing] that this quantum model of a perceptron can be trained in a hybrid quantum-classical scheme employing a modified version of the perceptron update rule and used as an elementary nonlinear classifier of simple patterns, as the first step towards practical quantum neural networks efficiently implemented on near-term quantum processing hardware.”
But what does this exactly mean, and what consequences does it have for Tesla and its business model?
Classical algorithms can be implemented on a quantum computer, the only thing is the performance won’t be any faster or better than if they were run on a classical machine.
Why Hasn’t Elon Musk Pursued Quantum Computing Yet?
Here’s the rub, though: when specially designed “quantum algorithms” are implemented on a quantum computer, they can lower time and memory complexities. Examples of these quantum algorithms come in the form of Shor’s algorithm, developed in 1994 by the mathematician Peter Shor for prime factorization; Grover’s algorithm, too — devised by computer scientist Lov Grover two years later — is a quantum search algorithm designed for searching an unstructured database.
The bad news is that such algorithms are too few and far between, as most existing quantum algorithms don’t match neural networks algorithms, not so good for Musk and his ambitions for Tesla, it seems, at least for helping him train the neural network that is so important to his company’s technology.
Another very real — if slightly sobering problem — is the hardware is just not ready yet.
What Would Elon Musk’s Engagement Bring to Quantum Computing?
Currently, it looks like the South African has not even thought about quantum computing as a solution to potential problems at Tesla, opting for his supercomputer for the time being as it may be many years before we pass the noisy intermediate scale quantum (NISQ) era and into the stage where quantum computers reach their true potential of “quantum advantage”.
While quantum computing may not assist Tesla right now, could quantum physics help Musk make inroads with some of his other companies?
Let’s take a quick look:
SpaceX: (spacecraft manufacturer, space launch provider, and satellite communications)
In a post published on the Microsoft Azure Quantum Blog in January of this year, entitled NASA’s JPL uses Microsoft’s Azure Quantum to manage communication with space missions, Anita Ramanan, Technical Program Manager Lead for Optimization at Azure Quantum, wrote about how NASA’s Jet Propulsion Laboratory (JPL) is leveraging Microsoft’s Azure Quantum to explore ways to communicate more efficiently with spacecraft exploring the solar system and beyond.
Communicating with space missions via its Deep Space Network (DSN) — NASA’s international array of giant radio antennas that supports interplanetary spacecraft missions — the JPL’s scheduling requests use the DSN antennae from the space missions and, unfortunately, this can be problematic for the Lab as it requires “intensive computing resources.”
However, as Ramanan points out, taking a quantum-inspired approach to scheduling optimization, “the Azure Quantum team developed a solution for a version of JPL’s scheduling problem with a limited feature set with the eventual goal to incorporate a broader set of requirements […], intend[ing] to reduce the need for lengthy negotiations and speed up the overall process.”
Quantum-inspired science, then, could be a possible panacea for, as Ramanan puts it, the “highly complex, multivariate problems […] quantum-inspired optimization algorithms” could solve.
So, quantum computing — or at least quantum-inspired technology — could be a lifesaver for Musk and SpaceX too. According to the opinion of former investment bank analyst at Morgan Stanley Adam Jones, published at SeekingAlpha in December 2020:
“From a SpaceX perspective, the commercial potential of quantum communications networks and its potential advantages of its rapidly deploying in-space comms architecture may provide significant optionality to the story and its valuation.
While, to our knowledge, SpaceX has not commented in detail on the enabling technology (ie. particle entanglement generators, quantum repeaters, random number generators, advanced cryogenics, etc) or the economic potential of quantum communications, we believe the company’s increasingly dominant position in space, satellite communications and DoD/government work makes this a natural extension of their capabilities. SpaceX continues to solidify its place as ‘mission control’ for the emerging space economy. Important milestones with Starlink, Starship and government contracts dovetail to support strong commercial momentum. We note that our recently revised EV valuation to over $100bn (bull case >$200bn) does not include any direct valuation attribution related to quantum communication networks, quantum-based metrology or cryptography.”
In the nearly two years since Jones uttered those words, SpaceX has yet to declare if quantum-inspired technology will benefit it in the 21st-century space race, though SpaceX does now use an AI-powered autopilot program to help its rockets navigate themselves from the launch to the ISS-docking station by measuring fuel usage and reserves, parabolic flight, weather, liquid engine sloshing, and other factors that affect SpaceX’s rocket flights.
Neuralink: (neurotechnology company that develops implantable brain-machine interfaces)
Reading An approach to interfacing the brain with quantum computers: practical steps and caveats, a paper published earlier this year by Eduardo Reck Miranda et al, you would suppose quantum computing and brain interfacing were a match made in heaven.
The abstract of the paper reads:
We report on the first proof-of-concept system demonstrating how one can control a qubit with mental activity. We developed a method to encode neural correlates of mental activity as instructions for a quantum computer. Brain signals are detected utilizing electrodes placed on the scalp of a person, who learns how to produce the required mental activity to issue instructions to rotate and measure a qubit. Currently, our proof-of-concept runs on a software simulation of a quantum computer. At the time of writing, available quantum computing hardware and brain activity sensing technology are not sufficiently developed for real-time control of quantum states with the brain. But we are one step closer to interfacing the brain with real quantum machines, as improvements in hardware technology at both fronts become available in time to come. The paper ends with a discussion on some of the challenging problems that need to be addressed before we can interface the brain with quantum hardware.
Now it is not for The Quantum Insider to say whether this is pure fantasy or as real as it gets, but the part saying available quantum computing hardware and brain activity sensing technology are not sufficiently developed for real-time control of quantum states with the brain gives us enough information to realize this is still a distant reality. As of the time of writing, there is scant information on Neuralink leveraging quantum technology to develop its implantable brain-machine interfaces (BMIs).
The Boring Company: (infrastructure and tunnel construction services company)
Finally, we have The Boring Company, Musk’s tunnel construction concern he founded in 2016 as a subsidiary of SpaceX to improve tunnelling speed so that it makes financial sense.
Or as Musk puts it:
“If you think of tunnels going 10, 20, 30 layers deep (or more), it is obvious that going 3D down will encompass the needs of any city’s transport of arbitrary size.”
Again, evidence of The Boring Company utilizing quantum technology for its drilling is lacking. Yet, definite use cases and authentic research efforts signal that the Pflugerville, Texas-based company’s adoption of quantum sensing technology is not too far away.
Work at The UK Quantum Technology Hub Sensors and Timing (led by the University of Birmingham), one of four Hubs within the UK National Quantum Technologies Programme, is proof quantum metrology is a legitimate technology that has the potential to change the game in drilling and construction.
Bringing together experts from Physics and Engineering from the Universities of Birmingham, Glasgow, Imperial, Nottingham, Southampton, Strathclyde and Sussex, NPL, the British Geological Survey and over seventy industry partners, the Hub has over 100 projects, valued at approximately £100 million, and has 17 patent applications.
As claimed by the Hub, quantum sensors can provide better sensitivity and reduced survey times, lowering survey costs and enabling a more prolific use of gravity surveys.
Isn’t this something The Boring Company could utilize to better serve potential projects like the tunnel from South Padre Island to Boca Chica Beach in South Texas or the proposed tunnel connecting the Ontario airport with the Rancho Cucamonga Metrolink train station?
Conclusion
Quantum computing could solve many of the headaches Musk has to contend with regarding his deep tech companies. Although Musk has made no explicit declarations, it’s only a matter of time before one or all of them avail themselves of quantum technology’s ability to solve what is currently unsolvable, or at least problematic.
What has been made clear from our review is that Musk is finding a rich suite of solutions through the use of High-Performance Computing, specifically in the areas of Artificial Intelligence, Machine Learning and Deep Learning. Whilst it is likely that society will need to continue innovating to find computational power to solve increasingly complex and challenging problems, Musk’s companies (amongst many others globally) are finding solutions today using increasingly powerful classical computing techniques.
So, will Elon Musk finally pursue quantum computing in 2024? Probably not.
