A few weeks ago I got a very exciting email, asking me what I thought about building a quantum computer out of Lego. Needless to say, it got my full attention, especially as I was very impressed with the Lego Atomic Force Microscope from back in 2012. As with all big ideas, I ended up talking about it with the sender over a coffee at the end of the week. This is the interim result. It’s going to be an adventure getting to the end of this project, and I’m looking for some help – maybe you!
A very very short background to quantum computers.
Today we have two mainstream types of quantum computer: Superconducting Qubits and Trapped-Ions. To make the first type you need (as an absolute minimum) a dilution refrigerator, superconducting qubits and some very fancy control systems. To make the second type you need high vacuum, high voltages, an ion source and some lasers. Trapped-ion quantum computers are generally cheaper to get started with, and indeed I found a really impressive guy on the internet has been building one in his garage. In summary, both mainstream types of quantum computers are very complicated and expensive. We cannot build one out of Lego, even with infinite money.
So, what can we build with $100?
We could get some lasers, some diffraction gratings, maybe an LCD or two and some single photon detectors to build a single quantum gate. There are some nice YouTube videos from @LookingGlassUniverse that are well worth a look. This would be very cool, but not exactly useful (and would need to be at least 2m long, unless we use lots of mirrors). Instead, we could build an optical computer. The team from Microsoft Research gave some very interesting presentations at CERN last year as part of the launch of CERN VentureConnect that I was fortunate to follow via Zoom. One of these caught my attention, discussing the use of off the shelf LCD elements for performing large matrix multiplications using light. This we could definitely do for $100 using a Raspberry Pi and hacking a few cheap LCD screens together.
Introducing the $100 optical computer!
The plan I have (so far, all plans change!) is to build a very simple optical computer that fits in a briefcase, technical details above. This will be one of the projects at the 2024 CERN Webfest hackathon, the launch of which is imminent. And as with almost everything I do for fun, it’s going to be as open source as possible (software and hardware). Once the computer is built, we will see how far down the road of computation we can go, with a quantum emulator/simulator as the ultimate goal. And, if we can do all this we should probably write a paper about it to publish somewhere.
To be clear, this is just a hackathon project. It isn’t a start-up, or a fundamental research project. We aren’t building a true quantum computer. It isn’t going to solve the worlds problems, like climate change, on its own – it’s just another powerful tool for us humans to use in achieving the challenge of the SDGs. It might be pretty neat for matrix multiplication, or it may not even work at all.
Do you need help?
Yes! I need 5 people to help me. It’s open to anyone, some will need to be available in the Geneva region, Switzerland, for the hardware building, but you could conceivably be remote if you are willing to have a few zoom calls around lunchtime in the Central European timezone. If you are interested in, drop me a line before the 26th July (best to email devine dot jd at gmail.com) with some info about yourself and why you want to join in this adventure. No pressure, a few short sentences is plenty! Across the team, I think we will need the following skills:
- Mechanical engineer, stability/structural calcs for our box (it will need to be sufficiently rigid) and mechanical drawings, potential 2D or 3D design/laser cutting/printing of a couple of components as necessary.
- Physicist, for optics (lenses, light transmission calcs) and quantum stuff, experience with Jupyter notebooks.
- Computer scientist/Software Engineer/Programmer (Python and/or C preferred), vector/matrix multiplication, image processing.
- Electronics engineer (low level/embedded development), wiring harnesses, SPI bus connections, hardware interfaces with Linux.
- Mathematician, but could also be a physicist or computer scientist with a strong analytical background (Build a simulator/emulator for the system in software, a digital twin), and figure out how to use this computational system to run ‘real world’ problems/port code from other systems and potential quantum emulation.
Recruits will get a confirmation by the end of the day on the 26th July. And if you don’t make the team, don’t worry. You will be able to follow along via our github repo, and if you have $100 you can always try making your own version.
Who will be our guides on this adventure?
I’m planning to mentor the team. I think I have most of the skills and knowledge to take the project all the way to completion (but we will test that!). I will be sharing my skills with the team members to encourage them in the right directions and give feedback, I’m also providing the hardware.
Beyond this, we will have access to some genuine specialists and experts in quantum and optical compute. The optical compute team at Microsoft Research in Cambridge are interested in this novel, low-cost, application of their work, I’ve already had a preliminary chat with them and plan to share progress as we build and test the box. We can ask them for advice as they have done this before with much more funding. I also have a contact with an expert in quantum computing at EPFL/ETHZ/PSI Centre for Photon Science who is going to give us some time with one of his PhD students. I have some friends and CERN Alumni who are passionate about useful things like matrix multiplications, who we can call on for support. In practice, this means we’ll have a few zoom sessions with the various experts over the two weeks of the programme, likely in the lunch hour slot.
How long is it going to take?
I don’t know how far we will get, but we will start in earnest on the 26th July 2024 and it will finish on the 10th August 2024. To be clear, this is strictly a ‘spare time’ project for me – mostly lunch times, perhaps a couple of hours in the evening and a few weekend hours on the 10th August. But as a participant, you are welcome to work on it as much as you like.
Is there anything else important to know?
This isn’t an official CERN project (which is why you are reading about it on my personal blog!), however it’s being done by the CERN MicroClub, which is a member based club at CERN, and will be part of Webfest 2024 that is supported by a bunch of different organisations. The main target audience for the hackathon are Summer Students at CERN, but really anyone (with at least some skills in one of the target areas) is welcome to apply.
Now I want to learn everything about Quantum Computers and Optical Computers, where do I start?
- The Analog Iterative Machine, a paper from the Microsoft Research Team in Cambridge.
- This great paper from the 1970’s that covers much of the basics for the system we will build.
- This patent from Microsoft, essentially covering the Analog Iterative Machine concept above.
- Qiskit, the IBM framework for quantum computing and quantum computer simulation/emulation. They also have some fantastic lectures on Youtube that are must-watches.
- Youtube videos from @LookingGlassUniverse about quantum computing.
- This is a nice looking lightweight quantum computer simulation library for Python, I think we might use it if we get that far and don’t find something better.
And of course if you are reading this after the 10th August, you can check out the project github repo and see how far we got. You can also check out the projects from last years’ CERN Webfest 2023.