Welcome, 

Project Eleven is an applied quantum computing and cryptography group. We protect systems from quantum attacks and will enable a transition to a quantum safe future. Our work sits at the intersection of quantum computing, cryptography and blockchain. 

Google’s Willow announcement in December brought a huge amount of attention to quantum computing. It received deserved acclaim for good science, being the first experimental demonstration of error correction scaling. Google showed that quantum error correction scales: adding more physical qubits leads to exponential improvements in the quality of logical qubits. This was proposed in the late 90s and is just seeing experimental realization today. 

In the past two weeks, Microsoft and Amazon have followed up with announcements of their own. Microsoft announced Majorana 1, a breakthrough in topological qubits and Amazon announced Ocelot, a prototype chip combining cat qubits and superconducting qubits. Both teams are pursuing unusual, immature qubit architectures. Their rationale is that these immature architectures have better error profiles and that this should drastically reduce the number of physical qubits required to create fault-tolerant qubits, which will lead to victory for the Seattleites. 

There’s been a bit of wrangling, but Microsoft’s general claim is to have got topological qubits working. These are a new type of qubit - add them to the existing set of superconducting, neutral atom, trapped ion and photonic qubits - and Microsoft has been trying to realise them since 2009. The chip they’ve announced has 8 topological qubits, based on Majorana particles. If true, this would be a scientific breakthrough and a great achievement by Microsoft. However, MSFT made similar claims in 2018, which they later retracted, and this latest claim has been met with skepticism. Topological qubits are tricky devices, and it’s difficult to prove that you have qubits rather than an intricate and expensive chunk of metal. Researchers await more evidence. 

A positive result from MSFT would mean that topological qubits are now a going concern. They may or may not scale, but the questions to be answered moves from “is a topological qubit possible?” to “what can these qubits do?” First, they’ll need to prove that they really do exist. See Scott Aaronson, the Wall Street Journal and NewScientist. 

With less fanfare, Amazon announced Ocelot. Amazon is focused on achieving fault-tolerant qubits via cat qubits. Like topological qubits, by virtue of their design cat qubits are resistant to certain errors. This should reduce the scale and sophistication of error correction needed to achieve fault-tolerant qubits. The chip contains 5 cat qubits and 4 superconducting (transmon) qubits linked together to form a single error corrected qubit, with a total logical error rate of 1.65% per cycle. Achieving a similar error rate on Google’s Willow chip required 49 superconducting qubits. See Amazon Science, Alice & Bob

Alice & Bob, a French QC company, is also working on cat qubits. They’ve shipped chips that include a single cat qubit, but have yet to publish work showing multiple cat qubits linked together. In general, the progress on cat qubits is promising, but scaling qubits and improving device performance has proven to be extremely difficult, with many well funded efforts struggling to make material progress. 

As an aside, it appears that the gold standard for research announcements from big tech companies is now a Nature paper. Much of the science making up the Willow and Majorana announcements was already available on the Arxiv before getting published in Nature and bundled into a PR announcement. 

For the cryptography-curious amongst us, check out our primer on RSA. We walk through the basics of the protocol, which is the single most popular asymmetric encryption protocol in use today. 

A 2019 paper estimated that breaking RSA-2048 requires 20 million physical qubits. Though that figure hasn’t been revised or recalculated yet, a number of breakthroughs have occurred since then which should lower the number of qubits required. Some of the advances include lattice surgery and magic state cultivation. We are optimistic that advances will continue and the resource requirements for running Shor’s algorithm, and other algorithms, will continue to decrease. This is good news for quantum computing, but does makes the problem of preparing for Q-Day more urgent. 

Until next time, 

Project Eleven 

  

Links

Microsoft Majorana

Amazon Ocelot

RSA