In this converastion with Amara Graps, we discussed quantum technologies and her work with analyzing and communicating their progress to a divers audience of researchers, managers, and investors.
Amara is a former planetary scientist who has pivoted to the field of quantum technologies. Amara shares her journey and the challenges she faced during the pandemic that led her to explore this new area.
The discussion covers the various aspects of quantum technologies, including computing, encryption, and communication. Amara explains the different ways of representing qubits, such as energy levels, spins, and the nitrogen vacancy in diamond. She also discusses the competition between different modalities and the importance of focusing on the problems each modality is solving.
I highlight the exponential progress in classical computing and the potential for quantum computers to achieve a double or even super-exponential trend, leading to quantum supremacy. However, Amara believes that focusing on benchmarks and usability is more helpful than focusing on labels like quantum supremacy.
The conversation also touches on the role of education in growing the quantum technology community, with China being about ten years ahead in terms of funding and introducing quantum concepts to students at a young age. Amara shares her plans to develop metrics using keywords from journal articles to help investors evaluate the innovativeness of quantum technology companies.
Throughout the discussion, various examples are given, such as D-Wave’s impact on the field of probabilistic computing, the application of AI in quantum challenges like protein folding, and the potential for quantum drones in defense. The importance of engaging the deep tech community and creating ecosystems around quantum technologies is emphasized.
The episode concludes with an invitation to follow Amara’s account @QuantProgress on X for valuable information on quantum technologies and an acknowledgment of the positive-sum game that technology plays in improving people’s lives.
David: Quantum computing, quantum encryption, and quantum communication are all facets of the same ability to not only understand quantum mechanics from a theoretical point of view but also to take the engineering steps needed to take advantage of quantum phenomena in various ways. However, it is a frontier technology, so it is hard, and in a lot of ways, it is still in the future. What we are here to ask ourselves today is how far in the future and when is the right time for talented people to dedicate themselves to this area? For investors to look at opportunities to accelerate the exploration of applicable products and services, and whatever may come up. Also, thanks to the questions and comments that you’re welcome to send on the various channels – LinkedIn, Facebook, YouTube, X – where we are streaming live. Without further ado, I welcome my friend Amara Graps. Welcome, Amara, to Searching for the Question Live.
Amara: Thank you. It’s great to be here.
David: I was very happy when you accepted my invitation because I know you are very busy and you apply yourself very intensively in studying this field, which is exciting but also complicated. Originally, however, you are not from the field of quantum mechanics, quantum theories, or quantum engineering. You are originally a planetary scientist. How did you move from one field to the other? And how did you end up picking quantum as your next area of activity?
Amara: The pandemic pushed me, as it did many others, to explore new fields. My funding strategy as a planetary scientist grew more fragile due to the time needed to help my daughter with her schooling. I no longer had as much time to write grant proposals, which is how planetary scientists are funded. Health was the major issue during that time. The Foresight Institute was leading a series of salons during that first year, and I grew much closer to them. When it appeared necessary for me to pick up an alternate backup career to fund myself, I wasn’t sure what it would be. A board member of Foresight introduced me to a newsletter organization called Insight Quantum Technology. I thought, “I like to write about technology. I wonder if I can write about this field.” I looked up the statistics and saw that it was growing twice as fast as space, which piqued my interest. I wrote a few pieces that were easy for me to write for that organization, and that’s how it began. However, to convey this technology, you need to understand it yourself. So, I spent the last two and a half years reading research articles, and my head is now full of journal articles from that time.
David: Scientists have a lot of hidden qualities and superpowers, one of which is being very patient and able to wade through the jargon and difficulty in communicating or understanding a new field. Reading through scientific papers and their particular nature is often far from the language used in more approachable platforms. You leveraged that hidden superpower, taking advantage of the way you think and the way you learned to work in science. You then turned that into an income-generating opportunity. Tell me a little bit about Inside Quantum Technology News. Is it a website that everyone can go and read everything, or is it a paid platform? How does that work?
Amara: It’s a general news site, and they also offer reports for investing organizations. Those reports are not cheap, and you can purchase them to make investing decisions in this field. There’s a lot of news there, and it’s at a level that managers can understand. The articles I produced were geared towards the deep tech community. Half of my articles are still there and accessible. As I wrote longer pieces, around 10 to 15 pages, they put those in the pro section. I don’t know what the pro costs are now because I’m writing less for them. I have a large report that I’m working on, but I’m less of a news writer and more of a deep tech engager for groups that need that style of understanding and communication.
David: Let’s define what the series of quantum technologies are. I mentioned computing, communication, and encryption, and there may be others like quantum storage. What makes it hard to deliver the promise that they represent from a theoretical point of view and produce products and services that leverage them on a massive scale?
Amara: If you can represent a change of state that you can encode like a one or zero, that is the baseline of quantum technology. It’s simple in that sense, but there are different ways of presenting a zero or one state. What’s quite common is using energy levels. You go from the zeroth energy level to the first energy level, which is stronger than going from the first level to the second level. This helps reduce leakage from the other energy levels. You can identify that in a probabilistic manner because quantum computing is a probability kind of computation where it could be in any one of these states, and you represent that state only after you make the measurement. When you hear someone say, “That qubit is its vector on a Bloch sphere,” that’s describing how it looks mathematically. You have to think of it like a wave, and you are taking that function of a wave and adding up the amplitudes in a way that tells you if it’s a one or a zero. Several of the different ways of making a qubit, maybe three or four of them, are about energy levels. If you can represent the zeroth to first more than the first to second, that could make a qubit. Another way is by looking at the spins of that atom. If it’s just one up or down, it’s a binary function, and you can also represent a quantum computation with spins or both spins and energy levels. The nitrogen vacancy in diamond is one of those. You take a carbon lattice, like diamond, remove one of those carbons, put in a nitrogen atom, and next door to that is a vacancy. The spins interact at that location where the nitrogen is and the vacant slot next to it. Because the spins interact, it makes it a very sensitive detection of magnetic fields, as spins respond to magnetic properties. A lot of the most accessible and available quantum technologies are in sensing, particularly of magnetic fields, and many of them are with this nitrogen vacancy property. Superconducting quantum computers and those technologies are 40 years old because Josephson junctions are 40 years old. In that case, you have a superconducting material where there are these pairs of electrons, Cooper pairs. If you put two superconducting materials next to each other with an insulating material between, these pairs can move between, it’s called tunneling, from one superconducting material to the other. That creates a circuit that can be optimized in a way to give you a property where you have a zeroth ground state and a first ground state. It’s uneven in the 0th to 1st ground state, meaning it takes more energy to move from the ground to the 1st than to move from the 1st to the 2nd. Because of that, you’re more likely to stay in that ground state, or you have more control over that 1st state. The superconducting qubits are a zoo. There are many different ways of optimizing that circuit, and what I described to you is very basic.
David: People often ask me a similar question about what quantum computing or quantum technologies are all about. First, I tell them that everyday computers or our smartphones are already built taking into account quantum properties of matter because we are operating at a scale where we cannot ignore them anymore. For example, the 7-nanometer and soon 5, then 3, and already people are thinking about 1-nanometer feature size circuits are such that particles, being both corpuscles and waves at the same time, like the electron, at those feature sizes, you cannot ignore their quantum nature. However, the classical circuits that have dominated our computing infrastructure forever were designed to try and bridle and lock down this quantum feature so that the circuit would behave classically as much as possible. None of our iPhones take advantage of, for example, the electron’s ability to tunnel across power circuits, except that they are semiconductors, and that is itself based on tunneling. But the big leap into quantum technologies and quantum computing is represented, in my opinion, by abandoning this expectation of being able to lock down the circuit and completely embracing the quantum nature of the components, taking full advantage or trying to take full advantage of the phenomena that otherwise are disturbing and bothering the engineering of our parts. They become something that we really cherish and want to work with. That is, in my answer, the big leap, and it really requires a complete rethinking of hardware, software, engineering, and the whole ecosystem from the ground up. We built generation after generation of classical computers from relay-based to vacuum tube-based to the first electronic circuits that became integrated circuits. Now we are at a system-on-a-chip level and so on and so forth. That took 50, 60, 70 years, depending on when and where you start. How long do we still need? You said Josephson junctions are 40 years old, but I don’t have a superconducting quantum computer under my desk. So, whichever is the most promising approach, how long, in your opinion, before it can produce practical and scalable products?
Amara: The manipulation of these different energy states is now in an exploration phase where some are quite mature, like superconducting quantum computers. They now have an industry around them, for example, with dilution refrigerators that cool down those Josephson junctions so that the superconducting material can behave like a superconductor while removing the noise in the environment to maintain and keep coherent that state from its initial ground state. Several require those cryogenic temperatures and ways to cool it down, like dilution refrigerators. There are other modalities, which is the word you might hear that refers to the way of making a qubit, that are at room temperature, for example, photonic quantum computing, which completely embraces the wave nature of the photon. It defines the qubit in a kind of way, if you imagine wearing polarized glasses that polarize the light, so you have light removed, and it’s either in one direction or the other direction. You can imagine that the photonic computer has these ways of splitting the light so that it goes one way or another, but immediately after, it detects where it is. Because that type of quantum computer does not maintain the state very long, you have to detect it right away and immediately collapse to determine in which direction it is in order to get the one or the zero qubit. When you make that detection, what it’s doing is looking at the probability of all those different polarizations and collapsing that to get your one or the zero qubit. Those don’t need to have a cryogenic temperature. There are some parts of that circuit that are more noise-free if they’re cooled, but it’s moving away from needing to be required to run those kinds of quantum computers. Right now, you have a competition between the different modalities, but I don’t think it’s as strong of a competition as viewers might think. The competition is about how high of quality is the one or zero. It’s called fidelity. In classical computers, you can compute with them just because you know that it’s always a zero or always a one. With qubits, you don’t always know that that’s the case. In these different ways of making the qubit, the fidelity now is becoming more and more similar among all the different modalities. It’s 99 point something percent, like 99.93 or 99.98, which means it’s now possible to begin to make a computation, but there is error associated, and it’s not a completely error-free computation. In these last year or two, they’re all becoming more and more similar. You might say, “Okay, so the competition is which will be the highest fidelity modality.” Well, at the same time, you have to include the algorithm that is run on these different modalities. We all know with photonic computers, the way that they make their qubit, they work on certain kinds of statistic problems better. They’re more adapted to that. That’s their algorithm, called Gaussian bosonic sampling. Those kinds of computations are perfect for that kind of quantum computer. But the other computers that have a process more similar to classical, where you have gates like an AND, OR, or XOR in the way that you put your computation together, those gate computers are a bit in a race. It may still turn out that the ion trap quantum computer is better for some kinds of problems, the superconducting quantum computer could be better for some types of problems, and the nitrogen vacancy in diamond are better for others. I think that’s a better way to look at it because the field is still not funded well enough to have such a convergence of one modality. I think you need to let all of these flower and bloom, and the competition will take place more inside of one modality. There may be a bunch of nitrogen vacancy in diamond technology companies, and as time progresses, a few rise to the top. The same thing with superconducting quantum computers. When you include the problems that they’re solving, you’ll know better which quantum computer you want. It may be that you will combine an ion trap computer with a superconducting computer. In the cloud-based systems where you access quantum computers, sometimes you don’t know which one of these are in the cloud, or they’re abstracted out. That’s a direction that the quantum computing field is going when you work in the cloud – you can turn on a switch, and you don’t care which one of these different kinds of modalities are used for your computation. You can turn it on and fix that.
David: Very interesting. I appreciate you highlighting that the stage of development is such that the various approaches are not touching each other yet. They are not directly competing but looking at different use cases. Then multiple projects within a given approach could target those use cases, and that is where competition may lead to a certain level of consolidation before, across the entire industry, there is another level of possible consolidation. This is what really happened with personal computers as well, where there were a very large number of approaches, makers, and architectures. One used this microprocessor and operating system, another used another one, and then the winner or the winners were able to really grab very large proportions of the entire market. Another feature of quantum technologies, particularly quantum storage and quantum computing, that many people don’t fully grasp is that with classical electronics, we have become accustomed to an exponential rate of progress. Moore’s Law, which is not a natural law but an engineering challenge, has been kept on track by the combined effort of tens of thousands of people all over the world competing with each other. Then, when a breakthrough would arrive, cross-licensing the solution to make sure that the whole industry could progress for 50 years, more or less, just to make a simple number, doubling the power of our computers and smartphones every couple of years. This defined the past exponential progress that really profoundly changed the world. However, with quantum computers, we may, once we can build them at scale, find at least a double exponential, if not even a further super-exponential trend. I call those that have this feature jolting technologies, to draw an analogy from further derivatives, in this case, of acceleration, which is the jolt. This nature of quantum computing is that, on one hand, we may increase the number of qubits and our ability to increase the number of qubits, and that is the first exponential, the one that we are already doing with the number of transistors in our integrated circuits that are growing exponentially. But the additional exponential is that in a quantum circuit, adding a single qubit or whatever number is needed to take into account error correction and the fidelity of the qubits too, whatever that number is, whether it is one or a thousand, it doesn’t matter. Adding that exponentially increases the capacity of that particular circuit to compute. So there are multiple exponentials packed onto each other, and this is what promises to achieve what is called quantum supremacy – the completely hopeless aim of classical computers to ever catch up with the power of quantum computers and their ability to address problem sets that are out of reach forever of classical computers. Now, a few years ago, maybe a couple of years ago, I didn’t look it up, Google claimed, and IBM poo-pooed, the achievement of quantum supremacy by saying whatever calculation we made with our quantum circuits, classical computers would have needed 10 million years to complete. IBM countered, “Well, actually, if we carefully designed the classical computer, it is not 10 million years. It’s maybe a thousand years or a hundred years.” I don’t remember. Are these claims and counterclaims fruitful for the industry, or are they hindering what the people concretely working with their sleeves rolled up, trying to do to improve the performance? Are they attracting investors, or on the contrary, are investors frightened by the confusion that these claims and counterclaims provoke?
Amara: IBM and Google, they’re the big companies. IBM is the one that most often likes these terminologies like quantum advantage, then quantum supremacy, then quantum advantage. Last summer, it was quantum utility. I think it distracts from the work. The scientists know what to do, and they’re solving the problems and the bottlenecks, and they’re making progress bit by bit. When you focus on this quantum supremacy label or utility, I think that is not helpful. The way to gauge the progress is benchmarks. The benchmarks could be, if you focus on the usability of that device, that’s a very important framework where you’re looking at the power consumption aspects or how many cycles in your algorithm or how many users you have in your environment. This usability kind of framework, I think, is more helpful because if you build your ecosystem of people and you are learning which algorithms are useful, which are not, which modalities are useful, which are not for that problem that you’re solving, and you’re always thinking about the problem that you’re solving, then the progress is crystal clear. In my opinion, you look at the fidelity and the fidelity of entangling two qubits, a two-qubit gate, for example, and you track that two-qubit gate fidelity over time, it’s always increasing. You track how many qubits you need to run an optimization problem to solve the low energy state of the lithium molecule because then you’re getting into problems that the pharmaceutical industry is really interested in, or the financial industry, where you have a portfolio with different components and how to optimize that portfolio. These are optimization problems, and the optimization component of quantum computing is at least 25% of the use cases. This is helpful because that’s a very lucrative area. This drives how you route your taxis through a town to get your rider to their destination or your public bus system to properly bypass a traffic jam. There are so many kinds of optimization problems that have money associated with them that quantum computers can already help with. Today, the way those optimizations are frequently run is with a combination of classical and quantum computing. In the world of today, where we don’t have exactly ones and zeros, we have to correct those to get your ones and zeros. The combination with classical computers is helpful. We’ll be in this hybrid, what’s called the NISQ era, the near-term noisy intermediate-scale quantum environment, for probably three or four more years. Once you correct that qubit so it’s much more precise, then you have something closer to a universal computer that can run a wider range of problems. So that’s the state of progress in this field. I got a little bit distracted. Did I cover some of this exponential progress a bit?
David: Don’t worry. You are doing fine. There are fields that are intrinsically hard and require both theoretical and engineering advances. The theoretical advances depend on smart people getting the funding so they can dedicate their time to make sure they can crack the problem. The engineering needs not only money but also time to carry out the experiments. An example of a field like this is fusion energy, where for a long time we were ten years in the future in being able to do something. Now it looks like we may be actually changing that paradox where after ten years it would still be ten years in the future. This is as opposed, for example, to the personal computer industry, where the barriers to entry were low, a lot of people could start, and there was a lot of excitement exactly because of how numerous the initiatives were. Without knowing, I would guess that quantum technologies are a little bit in the middle. If you have billions of dollars, like the companies we mentioned, it’s better. But there could be breakthroughs and there could be very innovative approaches that open the floodgates of applications. An example of this is D-Wave Technologies, which 10 or 15 years ago, whenever they started, was derided by academia, even accused of being fraudulent because of how unorthodox their approach was. While they are still around, their annealing quantum computer is definitely different from many others, and it is only applicable to a narrow set of optimization problems. But they have customers, and the customers keep buying bigger and bigger computers. So based on the definition of utility that you gave, D-Wave is a big success, and they didn’t start with a billion dollars. I was there at the unveiling of the first D-Wave computer at the Computer History Museum with Steve Jurvetson, their investor, and Jordi Rose, the founder. It was wonderful and exciting, and it was the start of a lot of enthusiasm at the time. So where are we today in this hype cycle? If there are deep tech investors, Albert is asking the question, for example, how should they go about evaluating the investment opportunities? Are there investment opportunities that people can take advantage of if they don’t have a billion dollars around that they can put on the table?
Amara: Look at the problems that these different modalities are solving. You mentioned D-Wave, and I started thinking about another thing. In my planetary science career, I spent several years building an asteroid mining community. When Planetary Resources, the first asteroid mining company, began, the planetary scientists were poo-pooing, just like the other physicists were doing for quantum computers, let’s say, 15 years ago. But they didn’t disappear. Planetary Resources existed and they kept existing. After a few years of them not going out of business, planetary scientists started going, “Hmm, maybe we should think about and take them a little more seriously.” Quantum computing and asteroid mining are two opposite parts of the cutting-edge investment strategy. What happened in asteroid mining is there were engineers building space missions, but they were missing some of the science. My effort those years was bringing the scientists to them, and that meant answers to questions from the asteroid miners. We built science knowledge gaps and the first roadmap for asteroid mining during that time. What happens in quantum computing is that the science is there, wedded to the engineering. They’re there, but scientists are a bit expensive. So you need to invest in a way that can keep these scientists advancing continually. That’s where there’s some pain in the process because I understand there are layoffs looming. But when I see the list of open jobs at all these different lists, whoever loses their job in whatever quantum device company they’re working at, they can move sideways to another company. I have no doubt about that. So where can the money be made today? I think it’s a combination of education and use cases, and maybe that’s continuing the slow slog in improving fidelity of these two-qubit gates and more. That’s always in the background, along with error correction. All those things are definitely proceeding. But the education is where we could help grow the community, and maybe there’s money to be made there with small education devices brought into high school classrooms or even younger. For example, the Chinese quantum community, which is about 10 years ahead of Americans when I looked at the range of funding. First of all, the funding went all throughout the strata, from lower local municipalities funding quantum, as well as all the ministries, and of course the defense and dual-use purpose. That’s perhaps a big driver in China, that this potential military use was going sideways and entering, and you don’t know what was going on and what is going on in that for the Chinese quantum development. The money in education is such that primary schools were getting visits by the quantum companies, and they still are today in China. So you start at a very young age learning about these quantum technologies. STEM education overall is enormous in China. I looked at the numbers. There’s one of the ministries, a ministry of education, that even has parts in English, and it’s open. You can look up how many graduate students in engineering they have. In 2020, in the graduate programs, there were a quarter million engineering students in programs. I mean, that’s mind-boggling. Those are the people who are building quantum devices, who will potentially be building quantum devices. So I think education is a way now you can earn some money, but it’s not going to be millions or billions. It’s going to be painful and slow. The investment needs to talk to people like me, where I can show the process.
David: There you go. As a matter of fact, I am showing your handle @QuantProgress on X, where people can follow you. You are posting links to your articles and other important information about this, and it is a great way to make sure that everyone is aware of what you are doing. Then they can take advantage of your knowledge to guide their investment decisions or engage you for deep research like you have been doing for Inside Quantum Technology News. Talking about education, of course, I don’t know if anywhere outside of China people are going into primary schools to talk about quantum technologies, which is so exciting. I want to be a primary school student in China now! But everyone and anyone can take advantage of some free quantum computer time on platforms like the IBM Quantum Cloud, where you can register for free, start playing around with elementary algorithms, and then run them. Even if IBM ran those things in a simulator rather than a real quantum computer, it would still be absolutely fine. Let me ask you something else about how the industry evolves and how it interacts, actually forming an ecosystem. We mentioned how a lot of engineering work is needed to re-implement, in an approach embracing quantum phenomena, what happened in electronics for decades. Similarly, we need to completely redesign the vast and rich libraries of software algorithms for quantum computers as well, potentially multiple times, because those different architectures will need at least slightly different algorithms each time. Is it conceivable, or maybe someone is already working on accelerating this process using modern contemporary AI approaches? Can we accelerate quantum progress through AI? More specifically, are large language models a good quantum coding companion at all? Second, could it also be that quantum computers would become a good or maybe even better platform for running particular AI algorithms, where the gazillion calculations that these models need to carry out, especially in training time, could be enormously compressed by the super-exponential nature of the quantum computing platforms?
Amara: In these hybrid classical-quantum systems, AI is already being used where it is preparing the initial conditions of your problem. Let’s say it’s an optimization problem, a variational type of problem, and AI selects the set of initial conditions that are fed and delegated to the quantum computer. The quantum computer does something, the hardest part, and then returns a result. The classical computer, which may have AI at the post-processing, can iterate and say, “Okay, this energy configuration is good, but let’s do another round and optimize it. Let’s feed in these new parameters back to the quantum computer and go around and around.” AI is even used midstream of the quantum calculation to make the quantum calculation go in a certain way and return results. So there’s definitely a blending of artificial intelligence and quantum computing in how these computations are made now. It’s a bit mind-boggling to define what hybrid computing is because the blending of quantum and classical is getting blurred and smoother. It’s a little bit hard. You have to go into the schematic of that computation to determine which is the classical, which is the quantum, and where is the random number part. Because you not only have the quantum computer, you have simulators of the quantum computer. As you’re optimizing your quantum circuit, you want to know which part should do what. So you can simulate what the quantum circuit does first to know where the load of the computation is and where to move it to run in a different way. This blending of the two types of computing is really interesting, and I think this is where education can help as well. The cloud services, you mentioned IBM, but there’s Amazon and there are others around. NVIDIA definitely has their hands in these cloud services because they’re accelerating with their GPU accelerators. Does that help? I got a little distracted.
David: Thank you. As I was speaking, I actually realized that we have a perfect example of AI being applied to a quantum challenge – the protein folding conundrum. With classical approaches, we have been very slowly cutting little pieces off and using x-ray crystallography. We were trying to understand how this structure is actually in 3D, very painfully and with very low speed of progress. Well, DeepMind, a division of Google, an AI-native organization, has been able to train an advanced AI system they call AlphaFold to really blow everyone away because they analyzed and then released the 3D shape of, first, a given number. If I am not mistaken, now they are at 100 million proteins. It was a triumph of AI technology applied to a quantum problem. Just to be explicit, the reason why this is a quantum problem is because the relationships, the repulsion and the attraction of these molecules, are based on quantum mechanical forces, and the classical analysis cannot understand and approximate the resulting 3D shape that these molecules take at the end. What is amazing is that not only did they achieve this result, but then they made the incredible decision to release the resulting database for free, including for commercial applications. So now it is something that hopefully thousands of new startups, universities, and researchers are leveraging and developing further.
Amara: I want to say a little bit about how D-Wave drove a whole field of computing that didn’t exist so strongly before they entered the scene. When D-Wave appeared in the early 2010s, there was a push to understand its speedup compared to classical computers. There was a very hyped press release in 2015 where the deep tech community was trying to understand how much speedup there was, and they were not finding it. In fact, if you looked at the Wikipedia entry for quantum annealing on that September 2015 press release, there was a spike of 9,000 times the entries of curious people going to Wikipedia to understand what quantum annealing means. Quantum annealing, traditionally from physics, is when you have a metal that’s heated up, then it cools down and relaxes into an optimum low energy state. That whole quantum annealing process is an analog for optimization that has been used since that time again and again, but not necessarily by quantum computers. I mean, D-Wave was the first. As the community of deep tech people were trying to find the speedup, they discovered that this analog of relaxing from a higher state into an annealed state could be simulated. So they put it in chips and classical computers. There’s a whole field in Japan – they’re the specialists with Toshiba and NTT. There are like four or five different probabilistic computing companies that have encoded that annealing process in a classical chip. But in learning how to optimize this annealing, they’ve figured out a bunch of tricks about sparse matrix computation, so you reduce the big computation space to a much smaller computation space. Well, that can inform the quantum computer people because you discover, “Oh, you don’t need so many qubits. You can do the same computation with fewer qubits.” The probabilistic field is enormous, and it’s circled back to inform the quantum computing field. I tend to think that quantum computing is a subset of probabilistic computing, but that depends if you’re a physicist or a mathematician, because sometimes people think one is a subset of the other. But I love this full circle informing where one field was trying to understand the speedup. They developed and created this whole new field. Now the algorithmic improvements and hardware improvements are circling back and informing quantum computing. So I love this interchange.
David: Wonderful. Thank you. Albert, who asked the question before, is suggesting we take a look at TerraQuantum.Swiss. I don’t know if you know them. TerraQuantum claims to have raised over $100 million in cumulative funding, and they have a quantum-as-a-service platform and other things. So, Albert, thank you for that. We will certainly take a look, and Amara will take a look. You spoke about education, and certainly your successful pivot as an entrepreneur and someone who opens their own path into quantum computing demonstrates that not only in schools, not only in China, but elsewhere, efforts like yours are valuable and necessary. Being able to translate scientific publications that you diligently read into something that other people can understand, connecting the dots of an industry as complex and varied as that of quantum computing, is really useful and valuable. So what are your plans? What are your next steps as you strengthen the efforts and the investment that you have made?
Amara: In reading these journal articles, I pull out keywords from each article that helps me set a baseline of the progress in that subfield. If you’re a scientific author and you submit a journal article, the journal asks you to submit two or three keywords, right? Well, that’s not enough in quantum technology because it’s so multifaceted and interdisciplinary. This field is changing and growing consistently. If you’re an investor, you want to know if this new technology that you’re looking at, this company spun off from a university, say, which is common, is innovative or kind of legacy. I’ve been establishing a baseline of keywords for each of these subfields of quantum technology that I want to use to inform investors. Is that a good investment, or is it something that maybe isn’t helpful now but is consistent or legacy? You need a bit of an informational process to help them understand because some investors like to read the papers, but many don’t have time, and it’s intense. So they need a little bit of some metrics, like a 1, 2, 3, 4, 5 scale from innovative to legacy, where they can assess whether that technology is useful to invest in or not. I think that’s something I’m in the process of developing – some kind of metric using the baseline of keywords that I pulled out from these hundreds of papers I’ve read in the last few years. I’m also gaining more people who want to engage the deep tech community, because my style of writing is to inform more of the deep tech community who could be beta testers and allies. The managers have the purchasing power, and they’re less often the technology specialists, but sometimes they are. In China, they often are, by the way. You have the researchers who then spin off a company, and then they’re the CEOs. That happens a lot in China.
David: Lawyers are not running the country over there.
Amara: Yeah. So the deep tech community needs to be convinced. If you use a little bit more of the language that they’re comfortable with, then they’re convinced. Like when D-Wave was making their large claims, the deep tech community was not convinced. That avenue of progress kind of suffered a bit, but in fact, it drove probabilistic computing. So it’s not necessarily a bad result. It’s just what happened. I want to continue to engage the deep tech community in different ways. It might be helping businesses who are already leaders in their area create an ecosystem. For example, if you’re a company that specializes in testing superconducting quantum qubits, and your business model is about testing that, and you focus on foundries that make the chips of those kinds of superconducting quantum qubits, then you can make an ecosystem of quantum foundries. And help grow your deep tech community who are still learning what they need, because some processes need optimization. The superconducting quantum computing field needs more optimization because it’s really intensive to lower the temperature and isolate that qubit and do the tests to determine if that’s a good qubit or not. You have 50 different items you’re testing from the different ways that quantum computer is made. That testing device, this is Orange Quantum, is a company that is a leader, in my opinion, of that kind of testing. They’re based in the Netherlands. So I want to grow ecosystems like that. I want to grow an ecosystem to develop quantum drones, for example. I live in Latvia. There are hundreds of drone developers because it’s really intense what we’re preparing here. We know who our neighbor is and what our neighbor will do. So defense, defense, defense and drones is really high. Well, if you can bootstrap off of the technology from drones and start adding the capability of quantum, so that there’s quantum key distribution at the nodes a little bit. South Korea are the leaders in that. They have already developed quantum drones of a limited capability, but they did it because of their aggressive North Korean neighbor. So why don’t we model in our East flank countries what the South Koreans did? I think that’s a really fruitful avenue. I’m looking a lot at how I might bootstrap the drone development in this part of the world and add a little bit of quantum there, tracking as best I can the people who are working on quantum drones. I love ecosystems. I love growing this deep tech collaboration and education.
