This last week saw MIT’s Future Compute conference with industry luminaries giving their take on what is happening in the world of Computing – not just Quantum Computing. Topics such as edge computing, high performance, quantum computing and of course cloud computing were the main stays of the talks from some of the world’s leading thinkers and technology businesses.
We’ll highlight some of the interesting pieces of information, snippets and experiences from some of the best thinkers in technology. Speakers from Microsoft, IBM, Honeywell and Psi Quantum spoke about the challenges of Quantum Computing.
Microsoft’s Mark Russinovich, CTO of Microsoft Azure, says get ready for Quantum Computing. The founder of a Winternals which was sold to Microsoft in 2006, says that we should all be getting ready for Quantum Computing. As we have reported Microsoft have now release Azure Quantum on general public release.
Psi Quantum is one of the most well funded Quantum start-ups on the planet. Founded originally in the UK, the team moved to find and secure better funding and have achieved VC “raises” in the hundreds of millions – we think one of the largest funding rounds in Quantum computing for a private company. Founded by Jeremy O’Brien and Peter Shadbolt, the company already has raised a reported half a billion dollars and is on series C investment round.
Jeremy O’Brien (from Psi Quantum) spoke about some of the problems that can be solved with quantum computing and of course discussing that Quantum Computing cannot simply speed up existing classical computing as many would hope. Sadly all too many people think that Quantum Computing can simply speed every existing processing demand, but that is simply not the case and remains one of the misnomers of Quantum Computing.
Pharmaceuticals, catalysts, optimisation problems are types of problems and sectors that can be assisted with quantum and there is a very very long list of applications that Quantum Computing can tackle. Rather than do things faster, it is all about doing these things at all – beyond the reach of classical computers. Some exponentially hard problems could never be solved with classical computing no matter the size and scale of those classical machines.
Jeremy O’Brien says that one million qubits is needed to do something useful and has been stating this for 25 years. Why one million qubits? Quantum Computers are hybrid devices – analogue systems. The analog features mean the error rate implies one needs around 10,000 qubits for an actual logical qubit. Therefore a really useful regime of Quantum Computing really requires something akin to million qubits.
Peter Barrett is a deep tech investor and holds a number of patents himself. His day job is the CTO & General Partner at Playground Global, a firm that invests in deep technology start-ups like Psi Quantum. He also sits on the board of Psi Quantum. Peter Barrett talks about some of the unknown chemistry that drives much of the world. We simply don’t understand natural processes such as photosynthesis, he states. We don’t know how medicine like metformin actually works. Beyond one million qubits it could allow us to engineer the world we live in.
Psi Quantum is currently mass manufacturing quantum chips in ‘tier one’ silicon fabrication locations (one of just three on the planet). The emphasis on using silicon technology and fabrication processes to drive the future makes sense since there have been over half a century of developments.
Jeremy thinks there really is no other choice for dealing with millions and billions of components, hence silicon photonics for the building blocks of the Quantum Computers they are building. Psi Quantum are building chips right now with millions of components currently in production.
The Architecture of the chips is new – called fusion based Quantum Computing – which is now public. Photons can remain alive for a short time, but computation can proceed for much longer – long enough to complete the calculation.
The one million qubit system will be the size of a small boardroom – the system resembles an HPC system that we have today when you account for the racking, cooling systems etc. There is also much work on the other components of the stack required. He thinks by 2025 we will have a million qubits that can tackle world changing applications.
Peter Barrett answered questions on what made Psi Quantum so investible? He was impressed by their ambition for a large machine rather then smaller iterations coupled with their ability to leverage photonics and silicon (which is a fairly mature technology). The obsession with sub 100 qubits is meaningless and really we should be thinking the question: “how close are we to 1 million qubits”. Peter thinks these machines are years away and not decades away and regards the event happening sooner rather than later in a matter of years rather than decades. Peter sees just a few years before one million physical qubits become a reality.
There is also a lot of excitement about chemistry – for we simply cannot do proper chemistry without quantum computers. If we cannot simulate the quantum subsystems then we cannot fully understand those systems. The world looks like a blacksmiths workshop trying to build an F35 fighter jet. Quantum Computing promises to be an amazing technology and when we look back on our world and will wonder how we managed with primitive cave man tools.
A big thanks to MIT Technology review for hosting Future Compute. We very much enjoyed attending. If you are interested in PsiQuantum mentioned here we have you can visit PsiQuantum’s website. Microsoft now have Azure Quantum on general release for those curious to open an account.