Мысли вслух

Quantum Error Correction

2025-06-03T17:32:25Z

Nord Quantique has announced what it calls a breakthrough in quantum physics that will make quantum error correction a bit better. Errors are the top obstacle today preventing the industry from having usable quantum computers, since the individual qubits are prone to mistakes and degradation.

Typically, quantum computing companies address the error problem with redundancy. They add extra qubits to compensate for those that go off the rails. The problem is that the more qubits, the faster the errors multiply, and the more additional qubits the computer needs to compensate for.

Getting to a point where adding new qubits lowers the error rate is the break-even point, what some experts call the “escape velocity” for quantum computing. Several companies have claimed to reach this point, though, without seeing them build large-scale quantum computers there’s no way to really know for sure.

Nord Quantique claims to be one of the companies to have reached this point.

Their trick? To create a qubit that holds multiple photons inside, and then use the redundant photons for error correction. They claim that this reduces the need to have any extra qubits, making it possible to build large, usable, quantum computers.

According to quantum computing expert Bob Sutor, founder and CEO at Sutor Group Intelligence and Advisory, it takes, on average, 1,000 redundant qubits to error-correct one qubit.

The industry calls this “logical qubits” — so, on average, 1,000 physical qubits is equivalent to one usable, working logical qubit. In Nord Quantique’s approach, one physical qubit is the same as one logical qubit.

“It’s almost like turning the error correction problem inside out,” says Sutor. “Instead of having redundant qubits on the outside to create one good logical qubit, you’re focusing on the inside of the qubit.”

According to Sutor, quantum computers get interesting at around 100,000 qubits. At a ratio of 1,000-to-one, that will require 100 million physical qubits to accomplish. Today’s most advanced quantum computers have less than 1,200 qubits.

By reducing the ratio, quantum computers would need a thousand times fewer qubits, making them significantly easier to build and scale, while also requiring less power and computational error-correction overhead. Last year, Nord Quantique demonstrated that their multiple-photons-in-a-single-qubit approach was feasible, creating qubits with up to 30 photons inside.

The limitation then was that their error correcting photons were only able to compensate for two types of errors. In today’s announcement, they’ve figured out how to compensate for six types of errors, by adding an additional mode to the qubit. This makes the qubit more resilient and accurate, and opens the path to add even more error-correcting modes in the future.

According to Nord Quantique CEO Julien Camirand Lemyre, each qubit is about the size of a walnut. Since the company uses the superconducting approach to quantum computing, it still needs that giant chandelier to get the system cool enough, which takes about four-and-a-half square meters of floor space.

With the current technology, a single chandelier will be able to support more than 2,000 qubits, says Lemyre. So how close is the industry to seeing a working Nord Quantique quantum computer? Not that close.

Even though DARPA selected Nord Quantique for its quantum benchmarking initiative in April — one of fewer than 20 companies chosen — an actual computer is still years away. “We expect to have more than 100 logical qubits by 2029,” says Lemyre. “And then scaling from there to 2,000.”

That’s a long wait. But, according to Sutor, Nord Quantique technology could make an impact earlier. “It could be a technology that is ultimately adopted by other players,” he says.

Nord Quantique makes superconducting qubits, he says, just like IBM, Google, and Rigetti. That means its redundant-photons-in-a-qubit approach could be adopted by someone who is further along by swapping in the better qubit but keeping the rest of the architecture the same.

“You always have to ask with startups — are they going to be a great, big, huge company or is someone going to buy them?” he says. he new breakthrough does make their approach more attractive, he says — and shows that DARPA was right in picking the company for their initiative.

Nord Quantique’s news isn’t the only big recent announcement from a quantum computing company.

D-Wave, which uses an older and less flexible approach to quantum computing, announced record-high $15 million in revenues for the first quarter of this year, a 509% increase from this time last year, along with commercial deployments in the automotive, pharma, and defense verticals.

In general, over the first quarter of this year, private investment in quantum computing reached $1.2 billion, up 125% year-over-year, according to data from The Quantum Insider’s Intelligence Platform(More details: https://thequantuminsider.com/2025/04/25/quantum-startups-secure-1-billion-in-q1-as-commercial-race-accelerates).

Notable investments include $360 million in IonQ, $230 million in QuEra Computing, $170 million in Quantum Machines, $150 million in D-Wave Systems, and 100 million Euros in Alice & Bob.

Another significant announcement was that of a rack-mountable, silicon-based quantum computer. Equal1’s computer has its own built-in cooling system, weights 440 pounds, and only has six qubits. Still, at about the size of a GPU server, it can fit into a regular data center.

Finally, earlier this month, Cisco announced a quantum entanglement chip(More details: https://www.networkworld.com/article/3978702/cisco-unveils-prototype-quantum-networking-chip.html). The research prototype, developed in cooperation with University of California, Santa Barbara, generates pairs of entangled photons that instantly transmit quantum state between each other, regardless of the distance between them.

comments

Ocelot

2025-02-27T16:41:31Z

Amazon Web Services launched its first computing chip designed specifically to support quantum computing today, on Thursday(February 27, 2025), joining IBM, Microsoft and Google in debuting such products.

Created between researchers at AWS and the California Institute of Technology, the new Ocelot chip’s starring feature is its error-correcting architecture that prevents external noise from disrupting calculations the chip and its corresponding hardware are working to execute. In a press release, AWS noted that Ocelot’s architecture has the potential to reduce the resources needed for sufficient error correction by five to 10 times.

The qubits — or quantum bits — Ocelot uses are called cat qubits, in honor of physicist Erwin Schrödinger’s famous thought experiment, Schrödinger’s Cat, which is used to explain quantum superposition. Cat qubits are a type of superconducting qubit, and are superimposed in various states between 0 and 1. This property intrinsically makes cat qubits resistant to noise and capable of supporting quick error correction, both vital for quantum computing.

“We believe that if we’re going to make practical quantum computers, quantum error correction needs to come first. That’s what we’ve done with Ocelot,” said Oksar Painter, the AWS head of Quantum Hardware. “We didn’t take existing architecture and then try to incorporate error correction afterwards. We selected our qubit architecture with quantum error correction as the top requirement.”

AWS said that it tested for error correction efficacy and also repeated tests on how well the logical qubits it composed held information to be processed. Painter said that the outcome of the experiments convinced him that Ocelot’s qubits can be scaled for more performative computing machines.

“Quantum error correction relies on continued improvements in the physical qubits," Fernando Brandao, the AWS director of Applied Science, said. “We can’t just rely on the conventional approaches to how we fabricate chips. We have to incorporate new materials, with fewer defects, and develop more robust fabrication processes.”

Ocelot remains a laboratory prototype, with AWS researchers continuing to refine its design and capabilities. Painter said that the company believes it still has several more stages of scaling to undertake with Ocelot, as well as more engineering problems to address, and will continue collaborating with academia to advance the fundamental research required to bring a cryptographically-relevant quantum computer to life.

“Right now, our task is to keep innovating across the quantum computing stack, to keep examining whether we’re using the right architecture, and to incorporate these learnings into our engineering efforts,” he said. “It’s a flywheel of continuous improvement and scaling.”

Other large tech companies are engineering their own quantum-designed computing chips and processors. Earlier this week, Microsoft announced its new quantum computing-designed chip, Majorana 1, which leverages subatomic Majorana particles to act as qubits.

Advances in the software and hardware crucial to scaling the burgeoning quantum computing industry are multifaceted, with disciplines like materials science, photonics and large-scale infrastructure needs as among the most pressing challenges researchers are working to solve.

comments

Majorana

2025-02-20T00:10:17Z

Microsoft announced a breakthrough in quantum computing today(February 19,2025), unveiling a new kind of quantum processing unit, using a new type of material, to create what it says is a “radically different type of qubit.”

The Majorana 1 – named after the Majorana quasiparticle – is designed to scale to a million qubits on a single chip that can fit in the palm of a hand. The goal is to bring the timeline for practical, reliable, large-scale quantum computers from decades down to years.

The new chip uses a new state of matter. Instead of solid, liquid, or gas, it is in a topological state. The breakthrough required developing a new material made of indium arsenide and aluminum, which Microsoft designed and fabricated atom by atom.

Majorana-based topological qubits is an approach that Microsoft has been pursuing for 20 years. Topological qubits are expected to be more stable than traditional qubits. It’s similar to how a knot in a rope remains in place even if someone jerks on the rope – the topological property of the knot itself keeps it in place.

“We took a step back and said ‘Ok, let’s invent the transistor for the quantum age. What properties does it need to have?’” said Chetan Nayak, Microsoft technical fellow, in a statement(More details: https://news.microsoft.com/source/features/ai/microsofts-majorana-1-chip-carves-new-path-for-quantum-computing).

Stability and reliability are key to making quantum computing work. “Whatever you’re doing in the quantum space needs to have a path to a million qubits. If it doesn’t, you’re going to hit a wall before you get to the scale at which you can solve the really important problems that motivate us,” Nayak said. “We have actually worked out a path to a million.”

Traditional qubits are extremely vulnerable to any change in their environment, which makes it difficult to scale up a quantum computer. But the new topological qubits need ten times less error-correction overhead, according to Microsoft. Traditional qubits also require analog controls, like turning a dial. Topological qubits, by comparison, can be controlled digitally.

“The results are real,” says Gabriel Aeppli, head of the photon science division of Switzerland’s Paul Scherrer Institute and a professor of physics at ETH Zurich. “In principle, topological approaches to quantum computing are ‘digital,’ and should scale better than more conventional approaches which can be seen as ‘analog.'”

Microsoft published a paper today in Nature magazine that describes how the topological qubits’ exotic quantum properties were created and how researchers were able to measure them.

The way it works is that four controllable Majoranas are joined together into the letter “H” with aluminum nanowires. Then, these individual H’s can be connected and laid on a chip, like floor tiles. “It’s complex in that we had to show a new state of matter to get there, but after that, it’s fairly simple. It tiles out. You have this much simpler architecture that promises a much faster path to scale,” said Krysta Svore, Microsoft technical fellow, in a statement.

The chips themselves are then combined with control logic and a refrigerator that keeps the whole system colder than outer space. Then, a software stack is used to program the chip and to connect with AI and classical computers – and all these individual pieces already exist, Svore said.

However, it will take years of engineering work to get everything to work together at scale, Microsoft said.

Meanwhile, enterprises can already experiment with quantum logic by accessing simulators and actual – small-scale – quantum computers through various quantum computing platforms. Microsoft, Amazon, Google and IBM, among others, all offer cloud-based quantum computing.

Majorana 1 can be “easily deployed inside Azure datacenters,” Microsoft said. However, the company did not say when the computer will become commercially available or what it will be able to do that other quantum computers cannot. Today’s announcement only proves that a topological qubit can be built.

“With the core building blocks now demonstrated—quantum information encoded in MZMs [Majorana zero nodes], protected by topology, and processed through measurements—we’re ready to move from physics breakthrough to practical implementation,” Nayak wrote in a blog post today(Details: https://azure.microsoft.com/en-us/blog/quantum/2025/02/19/microsoft-unveils-majorana-1-the-worlds-first-quantum-processor-powered-by-topological-qubits).

Now that the company has successfully demonstrated the world’s first topological qubit, the next step is to start building a scalable architecture around it, according to Nayak. A two-qubit system will demonstrate entanglement. Then an eight-qubit array will be used to implement error detection on two logical qubits.

The breakthrough is real and a huge engineering success, says Sankar Das Sarma, quantum physics professor at the University of Maryland. “But much more improvement is necessary before we can definitively say that this will lead to a commercial quantum computer.”

Microsoft’s single Majorana qubit is far behind what other quantum companies have in place. IBM, for example, has a 156-qubit quantum processor.

But the number of qubits is actually a minor measure, Das Sarma tells Network World. “What matters more is how error-free the qubits are. Microsoft’s qubits are intrinsically error-free giving them some unique advantages. Of course, they need to scale up and it is possible that they will succeed. We will see.”

In theory, at least, the new topological qubits will not only be able to scale faster, but do so more reliably, and take up much less space than today’s leading alternatives, Microsoft says.

“Scalability is absolutely real since it is made of tiny semiconductor wires,” says Das Sarma. “It scales up better than most other quantum computing platforms.”

In fact, the Defense Advanced Research Projects Agency (DARPA) has selected Microsoft as one of only two companies to advance to the final phase of its quantum computing evaluation program, which is looking to achieve utility-scale quantum operations by 2033.

Microsoft had a tough road getting here. In 2018, Microsoft researchers published a paper on Majoranas that was subsequently retracted, and some researchers doubted the elusive Majorana quasiparticles could ever be harnessed into qubits at all. But Microsoft continued with the research, and, in 2022, was finally able to demonstrate the existence of Majorana zero nodes, first theorized in 1937.

Majorana zero nodes are quasiparticles that are their own antiparticles. Specifically, they’re a type of anyon – a quasiparticle that exists in two dimensions and, when braided, can approximate quantum operations.

This is all cutting-edge physics stuff and incredibly difficult to understand, much less calculate and turn into working devices.

“Ironically, it’s also why we need a quantum computer – because understanding these materials is incredibly hard,” said Microsoft’s Svore. “With a scaled quantum computer, we will be able to predict materials with even better properties for building the next generation of quantum computers beyond scale.”

comments

Quantum and AI

2025-02-14T17:04:55Z

I don’t think there are two more Deep Tech topics discussed than quantum computing and AI. Whether from research teams at major corporations, academics turned startup entrepreneurs, public and private investors, CEOs, educated commentators, well-versed industry analysts, stock boosters and day traders, or newly minted social media experts, these areas have captured the imagination of many people worldwide.

We can consider them separately, of course, but what is their potential when their capabilities are linked or at least used in the same neighborhood? Let’s look at three ways of using them together.

If AI is great, and quantum is great, why don’t we use quantum to make AI even better?

Quantum computing is a different programming paradigm from what we have in classical computing. The latter, with its roots in the 1940s, uses millions or billions or more 0 or 1 bits to do almost everything we have in our phones, laptops, desktop computers, embedded processors, and data center and cloud servers. Quantum computers use qubits, a portmanteau of quantum bits.

One qubit holds two pieces of numeric data, and while a relationship between them must hold, it has significantly more information than a bit. What’s even more exciting is that you double the number of data components every time you can add an additional qubit. So, two qubits have 4 pieces of information, three have 8, four have 16, and so on. This is true exponential growth!

This growth has led some to claim that quantum computers will be able to process far more data for AI applications than classical systems. Well, that sounds good, but how do you get all that data into the qubits? It turns out there are several schemes for doing so, but none of them are very fast. For this reason, be on the lookout for phrases like “small scale” or “prototype” when people tout their Quantum for AI innovations. For the most part, people are solving little problems and are waiting until quantum computers get big and powerful enough for commercial applications, probably with fault-tolerance and error correction.

Without the error correction, we have a very short time to do anything interesting with qubits. If loading data for AI takes a long time, we may not have any left to compute with the data. It also means that when we start the actual algorithm, we may be unable to perform many instructions. In summary, Quantum for AI sounds good if you have a small amount of data and don’t want to do much with it. I suspect classical AI will work fine here in many cases in the short run.

Vendors and researchers may dispute this for the sake of PR and funding, and I agree that some are making good progress. I think Quantum for AI will be one of the last significant use cases to become practical. Until then, be wary of announcements in this area that mostly combine the quantum and AI buzzwords. You get extra credit if you mention “Generational AI”! I’m not just being snarky: caveat emptor. Look for third-party expert verification of the work and its scale.

Why bother exploring this now? Much of what goes on with machine learning is finding patterns in data and then doing something insightful with them. Since the quantum computing model is so different from classical, we may be able to discover new patterns or classically findable patterns much faster. At this point in the discussion, speakers and authors usually throw in the impressive fancy terms “superposition,” “entanglement,” “interference,” and “measurement,” but we don’t need them at this high level.

Vendors will often demonstrate Quantum for AI on the hardware they build. That’s reasonable to show a milestone in each development area. However, I would rather you showed me a quantum chemistry example since I believe that’s the first general use case area that will be practical for quantum computing.

Can we use machine learning to make better quantum computers?

As I mentioned, quantum computers use qubits as their basic information units. How many qubits do we need? The range of answers to this question is fascinating, with people claiming that dozens to the low thousands will be enough. My rule of thumb is that we will need 100,000 physical qubits, the manufactured or trapped natural entities that demonstrate the desired quantum behavior. Superconducting and silicon spin qubits are examples of manufactured qubits, and trapped ions, neutral atoms, and photons are natural. There are five other flavors (“modalities”) of qubits, but the first five and their variations are the qubit technologies used by the majority of vendors.

We explore quantum computing because, evidently, the biggest computer of all, Nature, uses the quantum mechanics model from physics to program all the small entities like electrons, photons, and atoms, and hence molecules and pretty much everything around us. We might as well emulate how the most significant computer works to solve our most challenging problems.

That sounds promising until we realize that Nature and the materials we create are quantum and don’t really care about interfering with our intentional calculations. Instead of getting qubits that maintain their values forever and operations on them that perform their jobs perfectly, we get added noise from the local natural environment. The noise causes errors in the qubits and their operations. For example, suppose you had a calculator, but electrical static caused the result of 2.0 – 1.0 to be 0.99. That’s an error in the electronics.

For another example, think about listening to a radio or a phone call with audio static. You may be able to make out the important information, but there might be so much noise that you just can’t understand what’s going on.

There are different kinds of noise in a quantum computer, and they come from several sources. Sometimes, we can detect patterns in the noise and use these to suppress or mitigate the errors.

Did I say “patterns”? Researchers and engineers have used several kinds of machine learning to detect noise patterns to improve the stability of quantum systems and the fidelity of what we do with them.

A 2024 survey paper published by NVIDIA researchers and colleagues nicely sums up the breadth of work in this area(More details: https://arxiv.org/abs/2411.09131).

It may very well be that we first do AI for Quantum to eventually get Quantum for AI.

Scientists and researchers in physics and computer science may not always be aware of industrial business processes and workflows to run their companies and provide value to their customers. It’s fine to work intensely in the hardware and algorithmic weeds, but someone has to raise themselves up and look at the context in which computation will occur.

Instead of thinking about quantum and AI somehow mashed up and working together, consider them working in separate processes. We also have traditional classical computation processes, including some modules that involve high-performance computing. Data enters each component, and then other data leaves and acts as input elsewhere. In mid-2024, Microsoft demonstrated a chemistry workflow starting with HPC, proceeding to AI, and ending with quantum computation(More details: https://quantum.microsoft.com/en-us/vision/quantum-for-chemistry).

Other vendors, including IonQ and Quantinuum, have also demonstrated such workflows. Note that the idea is not new: in 2020, IBM explained its view that the future of computation would be composed of bits (classical), qubits (quantum), and neurons (AI).

AI for Quantum has had value for several years and will continue to be a valuable tool as long as we build quantum computing systems.

Understanding the processing flow and data inputs and outputs among HPC, quantum, and AI systems is picking up pace now, and the differentiated value is likely to be middle-term: we should see practical results by the end of this decade.

We have much technology to develop before Quantum for AI gets significantly beyond the “we demonstrated that our work sort of isn’t too bad for small problems compared to classical AI” stage. This may seem harsh, but Quantum for AI is the most over-hyped quantum use case. I think we will need large error-corrected systems for this area, so I think this is long-term: seven to ten years and into the 2030s.

As our work on quantum evolves, so too will what we are doing with AI. The approaches we are taking now may not dominate either field ten years from now. Both are changing independently of the other. Our work to use them together effectively must take that into account.

comments

Quantum Teleportation

2025-01-05T16:10:07Z

Engineers at Northwestern University have achieved quantum teleportation using fiber optic cables already carrying internet traffic. This milestone could simplify the path to secure quantum networks by leveraging existing infrastructure.

“This is incredibly exciting because nobody thought it was possible,” said Prem Kumar, the study’s lead author and professor of electrical and computer engineering at Northwestern. Published in the journal Optica, the research shows how classical and quantum communication can coexist on the same network, paving the way for future applications like quantum computing and advanced sensing technologies.

At the heart of quantum teleportation lies a phenomenon called entanglement, where two particles remain linked regardless of distance. By manipulating these particles, researchers can transmit information without moving physical matter. This connection is not dependent on traditional methods of information transfer, such as light or sound, but instead seems to operate outside the constraints of space and time. Albert Einstein famously referred to this phenomenon as “spooky action at a distance”.

Imagine two entangled particles created in a laboratory and then separated, with one particle sent to Location A and the other to Location B. When a specific property, such as polarization or spin, is measured on the particle at Location A, the corresponding property of the particle at Location B is instantly known. This connection occurs regardless of the physical distance between the particles, whether they are a few meters or light-years apart.

In quantum communication, this property of entanglement allows for quantum teleportation. Unlike classical communication, where information is transmitted via signals such as electrical pulses or light waves, quantum teleportation transfers the quantum state of a particle from one location to another without physically moving the particle itself. This is achieved through a protocol involving entangled particles and a classical communication channel. For instance, when a quantum state is “measured” in a particular way at one end, the information about this state can be transmitted to the other end of the entangled pair, effectively recreating the original quantum state remotely.

“In optical communications, all signals are converted to light,” Kumar explained. “While conventional signals typically comprise millions of particles of light, quantum information uses single photons.”

The process involves transferring a quantum state from one photon to another through a technique known as destructive measurement. “The photon itself does not have to be sent over long distances,” said Jordan Thomas, the study’s first author. “Its state ends up encoded onto a distant photon.”

This exchange of quantum states over vast distances could pave the way for ultra-secure, ultra-fast data sharing, potentially transforming how we connect and communicate.

The security advantage of this system lies in its inherent properties. If an outsider attempts to intercept or measure the quantum state during transmission, the act of observation disturbs the state due to the principles of quantum mechanics. This disturbance alerts the communicating parties, making quantum entanglement an ideal foundation for secure communication systems, such as quantum key distribution (QKD).

ntegrating quantum and classical communication on the same cable posed a unique challenge. Fiber optic cables are already bustling with light signals carrying conventional internet traffic. In this chaotic environment, delicate quantum signals risk being drowned out.

“It would be like a flimsy bicycle trying to navigate through a crowded tunnel of speeding heavy-duty trucks,” Kumar said.

To solve this, Kumar’s team studied how light scatters within the cables. They identified a wavelength with minimal interference and placed their photons there, reducing noise from classical communication with special filters.

The experiment involved a 30-kilometer-long fiber optic cable, with quantum information and high-speed internet traffic sent simultaneously. By executing the teleportation protocol at the midpoint, researchers confirmed successful quantum information transfer despite the heavy traffic.

“Although many groups have investigated the coexistence of quantum and classical communications in fiber, this work is the first to show quantum teleportation in this new scenario,” said Thomas.

Looking ahead, Kumar’s team aims to extend the range of their experiments, explore real-world infrastructure, and advance techniques like entanglement swapping. These steps could enable even more sophisticated quantum applications.

“Quantum teleportation has the ability to provide quantum connectivity securely between geographically distant nodes,” Kumar noted. By integrating quantum capabilities into existing internet cables, researchers could sidestep the costly need for specialized infrastructure.

This achievement signals a future where classical and quantum communications coexist seamlessly, potentially unlocking the full promise of quantum technologies for society.

comments

Large Quantitative Model

2024-12-19T15:39:09Z

SandboxAQ announced it has raised $300 million in a new funding round at a $5.6 billion valuation. The Alphabet spinoff previously raised $500 million in its initial funding round in February 2023. The new funding will accelerate application development and adoption of SandboxAQ’s Large Quantitative Models (LQMs) and AI solutions...

SandboxAQ’s LQMs are trained on proprietary data generated using physics-based methods, a point of difference from general-purpose large language models trained on massive datasets. LQMs are being used in industries like biopharma to accelerate drug discovery and synthesis and in materials research for developing new manufacturing materials and compounds.

“Large Quantitative Models are the next wave of AI, as they provide a powerful ability to solve science and business problems for large industries including aerospace, biopharma, chemicals, defense, energy, finance, and more. The capital raise we are announcing today gives us additional resources to drive deep impact at scale,” said SandboxAQ CEO Jack D. Hidary, in a release. “LLMs and LQMs are complementary platforms that are both needed in the world of B2B applications. We are pleased to see the commitment of so many long-term investors in SandboxAQ.”

SandboxAQ began as Alphabet’s AI quantum computing unit led by Hidary. It was spun out of Alphabet into an independent startup in March 2022, with Hidary as CEO and former Google CEO Eric Schmidt as chairman.

“Leading global enterprises are realizing that they must look beyond the capabilities and limitations of LLMs and embrace LQMs in order to maximize the ROI from their AI investments,” said Schmidt in a release. “Jack Hidary and his team at SandboxAQ have shown the ability to create significant customer value across key industries such as biopharma, chemicals, and financial services. Jack is a world-class, high-integrity CEO leading a deeply technical team. With this round, SandboxAQ can move even faster to its goals and impact.”

One of the more famous SandboxAQ investors is Yann LeCun, a leading AI scientist and one of three Turing Award winners for deep learning.

“I am investing in SandboxAQ because of their industry-leading approach to quantitative AI,” said LeCun, who is also the chief AI scientist at Meta. “While LLMs are very helpful tools for consumers, it is quantitative AI that will define work in large sectors of the economy including biopharma, chemicals and financial services. I am impressed by the technical depth of Jack and his team and am excited to support their work. SandboxAQ has emerged as a leader in novel applications of AI that solve the most pressing challenges in the world and their technical success is impressive.”

SandboxAQ is investing significant resources in AI-driven scientific innovation, particularly in the healthcare field. In 2024, SandboxAQ’s AQBioSim division advanced its AI-driven drug discovery efforts, forging partnerships with two top academic research institutions to accelerate therapeutic development for neurodegenerative diseases. The company also deepened collaborations with major biopharma firms, using its LQMs to uncover novel biomarkers and streamline clinical development for investigational drugs. It also launched IDOLPro, a generative AI tool designed to create drug molecules with tailored properties, speeding up research and development.

he company is even working on a medical device under its AQMed division called CardiAQ, a magnetocardiography (MCG) investigational device under development designed to capture and analyze magnetic signals from the heart for faster and more accurate cardio assessments. The company also announced a new clinical research study with The Mayo Clinic, supplementing its ongoing collaboration with Mount Sinai Medical Center and a successful feasibility study with the UCSF Medical Center.

The latest round was funded by Fred Alger Management, LLC, T. Rowe Price Associates, Inc., Mumtalakat, Parkway Venture Capital, Breyer Capital, Rizvi Traverse, S32, US Innovative Technology Fund, Ava Investments, Eric Schmidt, Marc Benioff, David Siegel, Yann LeCun, IQT, and other prominent investors.

“We see significant growth potential and opportunity for LQMs across a broad range of industries, which is why investing in SandboxAQ is an investment in AI’s future,” said Jim Breyer, Breyer Capital. “Through SandboxAQ, Breyer Capital has a front-row seat to a generation-defining company with some of the most inspiring uses of AI technology ... uniting science and AI in bold and meaningful ways to address the world’s most pressing challenges.”

comments

Quantum

2024-12-14T20:16:01Z

Google’s announcement on Monday of its new quantum chip, called Willow, was full of eye-catching statements. For starters, Willow took less than a minute to perform a benchmark computation that would take one of today’s fastest supercomputers 10 septillion years to do. That number “exceeds known timescales and physics and vastly exceeds the age of the universe,” says Hartmut Neven, founder and lead at Google Quantum AI, in the announcement(More details: https://blog.google/technology/research/google-willow-quantum-chip).

And then there’s error correction. “We achieved an exponential reduction in the error rate,” Neven says. In fact, the more physical qubits the team added, the more it reduced the error rate, hitting a historic accomplishment.

Since qubits – quantum bits – are notoriously unstable, quantum computer companies use redundancy to improve accuracy, with multiple physical qubits combining into a single “logical” qubit.

Google also claims that this is one of the first times the error correction was performed in real time. And that the logical qubits have a longer lifetime than the individual physical qubits do.

It’s the first system that’s “below threshold,” Neven says. “It’s a strong sign that useful, very large quantum computers can indeed be built. Willow brings us closer to running practical, commercially-relevant algorithms that can’t be replicated on conventional computers.”

And, in an aside that should be of interest to any sci-fi fan, he says that the speed of the computation “lends credence to the notion that quantum computation occurs in many parallel universes, in line with the idea that we live in a multiverse.”

In a video presentation also released on Monday, Neven added that this week’s breakthrough means that practical commercial applications are as close as five years away. Most other experts estimate that this date is at least ten years in the future.

NIST, for example, recommends that federal agencies stop using classical encryption in 2035 and switch to quantum-safe encryption instead.

Google’s achievement might seem impressive, but industry experts warn that there’s still a long way to go before practical use of quantum computers.

“Could I solve a problem today that I couldn’t solve yesterday? The answer is no,” says Yuval Boger, chief commercial officer at QuEra Computing. “Nothing has changed in that sense.”

QuEra is a quantum computing company that uses the “neutral atoms” approach to building qubits, in contrast to Google’s “superconducting circuits” approach. In October, QuEra announced that it received a strategic investment from Google, though it did not disclose the sum. And, earlier this year, QuEra announced its own breakthrough in error correction.

“But they’ve passed one more waypoint,” Boger adds.

The first waypoint was to figure out whether quantum computing even works. That’s already been settled — it works. The second waypoint is whether the errors can be corrected. That waypoint has been achieved this year, he says, this week by Google — and, earlier, by his company.

“The third step – and we’re not there yet – is if it can scale,” he adds. “We’re now entering that stage.”

What about the breathtaking speed that Willow demonstrated? It turns out that the benchmark Google used was a random circuit sampling algorithm with no real applications.

“The new tech is not immediately actionable from a business point of view,” says Stefan Leichenauer, vice president of engineering at quantum computing software company SandboxAQ, which was spun off from Google in 2022. “The problems of scale are still present, and there’s a long way to go before they are solved.”

While practical quantum computing applications might still be many years away, security experts say the timeline for quantum-safe encryption preparation needs to be accelerated.

“Quantum computers will eventually break the asymmetric encryption we’ve relied on for over 50 years,” says Jordan Kenyon, chief scientist in Booz Allen Hamilton’s quantum practice. “It is a question of when, not if. If you are a CISO or CIO and are not actively planning, prototyping, or implementing the PQC algorithms NIST standardized this summer, there is no better time.”

It’s very likely that nation states will deploy quantum computers to crack the most sensitive messages even before the technology is widely available for commercial use. Plus, there are the “harvest now, decrypt later” attacks where adversaries vacuum up valuable communications or data, then decrypt them when the technology becomes available, says Forrester analyst Brian Hopkins.

“Even if quantum decryption is a decade away, attackers could steal encrypted data today,” he says. “Companies should begin transitioning to post-quantum cryptographic algorithms.”

And, to maintain flexibility as these algorithms evolve, companies should adopt a “cryptographic agility” approach, he adds, allowing them to swap out encryption methods as needed.

Beyond cybersecurity, there are other steps that companies can take to prepare for a quantum future.

Google’s Neven, for example, suggests that researchers, engineers and developers start educating themselves. There are even free resources out there, including a new quantum error correction course that Google launched this week on Coursera.

“Awareness and training is always a good thing,” says Jon France, CISO at ISC2 (International Information Systems Security Certification Consortium). “If not only for security reasons, but also for business opportunities.”

There’s always a chance that quantum computing will have a breakthrough moment that advances the timeline, he says. In fact, Google’s Willow announcement might just be one of those breakthrough moments.

“This may not have a practical impact for day-to-day use, but it continues to highlight the advances in the field and may move planning from long-term to mid-term,” he says.

Booz Allen Hamilton’s Kenyon suggests that companies should be working now to identify mission-critical use cases where quantum technologies may offer an advantage, prototype applications, and start investing in a workforce that can make it possible.

And, since there’s a shortage of quantum experts, this means a long-term commitment to upskilling.

“Quantum demands a breadth and depth of talent that is difficult to recruit, retrain, and retain,” she says. “But a diverse workforce is critical to identifying and delivering on the mission-critical use cases that will unlock the technology’s real-world potential.”

comments