Source

The story begins in the spring of 2013 when Google Research announced the Quantum AI Lab. Back then, the Lab’s launch was powered by the most advanced quantum computer commercially available at the time, D-Wave Two from D-Wave Systems.

Nearly a decade has passed since — the Quantum AI Lab has achieved a lot, which we will describe in more detail below.

Google Quantum Computing Background

A pioneer in quantum computing research and development, Google is one of the world’s leading companies in this area. It was announced in 2019 that Google had achieved “quantum supremacy,” which means it had demonstrated the ability of quantum computers to perform calculations (however contrived) beyond the capabilities of classical computers. Sycamore, Google’s quantum computer, performed a calculation in 200 seconds that would have taken the world’s fastest supercomputer 10,000 years.

Google has continued to push the boundaries of quantum computing since then.

As well as developing quantum algorithms and applications for fields such as chemistry, materials science and machine learning (ML), the company is striving to develop even more powerful quantum processors. Quantum algorithms and applications can also be built on Google’s cloud-based quantum computing platform, the Google Quantum AI (QAI) platform.

Google’s Core Quantum Technology

As of 2023, there are certain milestones that Google wants to reach in order to be able to build a commercially viable quantum computer. One was quantum advantage, or doing a calculation faster than a supercomputer with a quantum machine, with 2019 marking the completion of this goal.

This year, the company’s researchers said they had shown that a system using error-correcting code could detect and fix errors without affecting the information. This is the first demonstration of a logical qubit prototype, which shows that increasing the number of qubits reduces errors.

In a company blog post from February 2023, Sundar Pichai, CEO of Google and Alphabet, reported that for the first time ever Quantum AI researchers demonstrated that increasing the number of qubits can reduce errors. Published in Nature, the result of this achievement — where the researchers were able to make a logical qubit that performed better than one we made from 17 physical qubits — means the company is now able to operate quantum computers in a much more efficient way.

Google Quantum Offerings

Since the beginning of Google Quantum AI, the company has been making the products and tools it develops for its own research free to the public.

Here are some examples:

QUANTUM VIRTUAL MACHINE (QVM)

In 2022, Google Quantum AI released the Quantum Virtual Machine (QVM), developed to simulate the experience of programming one of the quantum computers in its lab, including circuit validation and processor fidelity.

The machine works by combining the measurements (qubit decay, dephasing, gate and readout errors) from its Sycamore processors with the models from the physics research team, which can then simulate quantum processor-like output using the company’s models.

PYTHON-BASED CIRQ

Next up is Python-based Cirq, a software library for writing, manipulating and optimizing quantum circuits, which are then run on quantum computers and quantum simulators. It provides useful abstractions for dealing with today’s noisy intermediate-scale quantum computers, in which the hardware details are crucial to achieving state-of-the-art results.

POST-QUANTUM CRYPTOGRAPHY (PQC) ALGORITHMS

Another area of interest to the company is in testing post-quantum cryptography (PQC) algorithms. Having worked with the security community for over a decade, Google has been exploring options beyond theoretical implementations for PQC algorithms.

A post-quantum experiment with Cloudflare was announced in 2019. Using Cloudflare’s TLS stack, Google implemented two post-quantum key exchanges and deployed them on edge servers and Chrome Canary clients, which resulted in providing the company with more insight into two post-quantum key agreements’ performance and feasibility in TLS — the results also meant that Google has incorporated this information into its technology roadmap and meant that two years later Google researchers were able to test a range of network products that were incompatible with post-quantum TLS and found that they were incompatible with post-quantum confidentiality. Experimenting early resulted in the prevention of this issue from arising in the future.

Currently, the company is implementing PQC that addresses both immediate and long-term risks of protecting sensitive information and ensuring that Google is PQC-ready.

Who are Google’s Main Competitors in Quantum Computing

In the field of quantum computing, Google’s main competitors are the likes of IBM, Microsoft, Amazon and Intel insofar as they are tech giants who have an interest in quantum. In reality, however, we see a significant amount of collaboration across the ecosystem.

In quantum computing, IBM has been a pioneer for some years. Now, IBM Quantum offers access to its own quantum processors through its cloud-based platform.

Microsoft offers Azure Quantum, the company’s cloud-based platform that provides access to its quantum hardware.

Another is AWS (Amazon) and Amazon Braket, a fully-managed quantum computing service that lets customers access hardware from a variety of vendors, including D-Wave, IonQ and Rigetti.In a blog post from April 2021, AWS released a paper to the world, Building a fault-tolerant quantum computer using concatenated cat codes, which was written by a team at the AWS Center for Quantum Computing and describes an architecture that combines elements of active QEC and passive or autonomous QEC. This paper proves researchers there have demonstrated the feasibility of cat qubits based on superconducting circuits, and their highly biased error rates can be utilized to design additional quantum error correction circuits.

Finally, there is D-Wave, a Canadian company that specializes in quantum annealing, a type of quantum computing. The company has developed its own quantum annealing processors, which are available through its cloud-based platform Leap.

How is Google’s Quantum Funded?

Google is primarily funded from its own balance sheet. In addition to its own resources, Google receives grants from government agencies like the National Science Foundation and partners with academic institutions in order to conduct quantum research.

Who Are Google’s Most Significant Quantum Customers?

To advance quantum research, Google has also collaborated with government agencies and academic institutions. CERN, NASA and the University of California, Santa Barbara are some of Google’s prominent partners in quantum computing.

Google’s Quantum Partnerships

In recent years, and as part of its research and development efforts, Google has sought out partnerships in quantum computing with several institutions/companies, some of which we will highlight below:

Google is exploring quantum computing’s potential in financial services with J.P. Morgan. In this partnership, new algorithms will be developed for the optimization of portfolios, the analysis of risk and the detection of fraud.

The European Organization for Nuclear Research (CERN) and Google announced a partnership in 2021 to collaborate on quantum computing. In order to solve some of the most challenging problems in particle physics, the partnership aims to develop new algorithms and tools.

With the goal of exploring quantum computing’s potential for aerospace applications, Google has partnered with Airbus, too. As part of the collaboration, algorithms are being developed for flight optimization, scheduling and routing, as well as for improving maintenance and safety on aircraft.

Another to note is Google’s collaboration with Volkswagen, designed to explore ways to use quantum computing in the automotive industry to enhance battery performance, optimize traffic flow and develop new materials for electric vehicles.

Where is Google Heading in the Years To Come?

As one of the leaders in quantum computing, Google has been heavily investing in the technology. As Google continues to explore quantum computing applications in areas such as optimization, cryptography and ML, its quantum computing efforts are likely to focus on developing more powerful quantum processors and algorithms. The Google quantum computing team has already achieved several significant milestones, such as demonstrating quantum supremacy, or the ability for a quantum computer to perform computations that are practically impossible for classical computers. A 53-qubit Sycamore processor developed by Google enabled this achievement.

Google is likely to make further advances in quantum computing in the future by developing even more powerful quantum processors and solving even more complex problems. As well as exploring new applications for quantum computing, the company is likely to explore drug discovery, materials science and financial modelling as these seem to be areas where quantum computing can help the most. For more detail on this, check out the company’s “Our quantum computing journey”, which explains Google’s quantum roadmap in simple terms.

Before quantum computing is widely used in practical applications, there are still many challenges that must be overcome that companies like Google and others are making significant advancements in.

MAIN STRENGTHS

As we have already mentioned, Google has already achieved a key milestone by using more qubits to reduce the error rate of quantum calculations for the first time, according to its researchers, so this is obviously a step in the right direction.

CHALLENGES TO QUANTUM SUPREMACY?

In a story reported by The Quantum Insider last year, research conducted by the Institute of Theoretical Physics at the Chinese Academy of Sciences, led by statistical physicist Pan Zhang, showed that the team beat Google’s quantum computer Sycamore at the computational task examined in Google’s 2019 study.

Google claimed back then a supercomputer would take about 10,000 years to complete the task. Whereas on a supercomputer, it took the Chinese researchers about 15 hours to complete the task.

In order to solve that problem, Zhang and his team took the problem from a slightly different angle by using a 3D tensor network with twenty layers, each layer held 53 dots, one for each qubit. The team connected the dots to represent the gates, with a tensor encoding each gate. According to Nature, multiplying the tensors was all the team needed to run the simulation.

Although this takes nothing away from Google’s achievement four years ago, Zhang and co expected classical algorithms to improve in performance due to the original experiment’s purpose of exploiting quantum computer strengths and classical computer weaknesses, and eventually expect to develop algorithms specifically designed for this calculation.

The researchers of the report also said that as quantum computers improve, a slight tweak in performance might soon put them outperforming classical supercomputers — and even classical algorithms.

So there is hope. Yet, the research showed that no technology is foolproof.

Google Quantum Computing Key Takeaways

Humanity is on the verge of creating a useful, error-correcting quantum computer, whether today, tomorrow or fifty years in the future and there will be obstacles on the way to error correction, that’s a given.

From Richard Feynman’s idea of creating universal quantum computers for simulating other quantum systems more than forty years ago to the state of play today in our industry, what Google achieved in 2019 has proven quantum computing is a practical reality rather than a theoretical possibility and has led us to a new era of computing: Noisy, Intermediate Scale Quantum (NISQ).

Google — with its own hard work and investment and collaborations with universities and research institutions — is developing theories and algorithms that we can apply to error-corrected computers in the future.

When we get to that place, Google could be one of the first.

What does quantum computing have in common with the Oscar-winning movie “Everything Everywhere All at Once”? One is a mind-blowing work of fiction, while the other is an emerging frontier in computer science — but both of them deal with rearrangements of particles in superposition that don’t match our usual view of reality.

Fortunately, theoretical physicist Michio Kaku has provided a guidebook to the real-life frontier, titled “Quantum Supremacy: How the Quantum Computer Revolution Will Change Everything.”

“We’re talking about the next generation of computers that are going to replace digital computers,” Kaku says in the latest episode of the Fiction Science podcast. “Today, for example, we don’t use the abacus anymore in Asia. … In the future, we’ll view digital computers like we view the abacus: old-fashioned, obsolete. This is for the garbage can. That’s how the future is going to evolve.”https://embed.podcasts.apple.com/us/podcast/fiction-science/id1528078321

Computer scientists might take issue with Kaku’s digital doomsaying — but there’s little doubt that quantum computers will transform the field as much as artificial intelligence is transforming it today.

“Quantum computing could very well revolutionize what an Amazon Web Services or Microsoft Azure will want to do for the world in terms of computing,” says Louis Terminello, associate laboratory director for physical and computational sciences at the U.S. Department of Energy’s Pacific Northwest National Laboratory.

Kaku’s assessment of the potential impact goes a lot further: In his view, any problem that involves sifting through a multiverse worth of possibilities will become more solvable once the quantum revolution takes hold. Energy generation and storage, food production, climate modeling, disease treatment and genetic repair are all potential targets for quantum supremacy.

Why is that? In contrast to the rigid one-or-zero approach that serves as the foundation of classical computing, quantum computers would take advantage of the fact that quantum bits — better known as qubits — can represent multiple states when information is processed.

“Quantum computers, in principle, are infinitely more powerful than a digital computer that computes on zeros and ones, zeros and ones, because quantum computers are quantum mechanical,” he said. “The atom can spin in any direction. How many directions are there? An infinite number of directions.”

Tech titans haven’t yet settled on the best basis for quantum computing: Amazon, Google and IBM use superconducting circuits in their hardware. IonQ — which is creating a research and manufacturing facility in the Seattle area — favors a technology based on trapped ions. Other companies are taking advantage of the quantum properties of photons, or defects in silicon lattices. And Microsoft is placing its bets on topological superconducting nanowires.

Which technology will win out? Kaku says it’s too early to tell.

“How many quantum computer architectures are possible? An infinite number of them,” he says. “Now, of course, only a handful of them are practical and economical. But the point I’m raising is that Mother Nature has already devised millions of quantum mechanical systems, and we’re playing catch-up to Mother Nature. And so I think that one day, one or or a handful of these technologies will dominate the whole field, but we’re not sure yet.”

Even though full-fledged quantum computers aren’t yet ready to prime time, researchers are already trying to figure out how to simulate the quantum mechanisms behind important biological processes such as photosynthesis and nitrogen fixation. Coming up with new molecular methods to perform those tasks could be worth billions of dollars.

“About 1% or so of the world’s energy goes to the process to refine nitrogen in the air to create fertilizer,” Kaku says. “But it’s very wasteful. … We need a quantum mechanical Green Revolution.”

On the energy frontier, quantum computers could help engineers design better reactors for generating fusion power — and help chemists design new types of materials for solar cells and batteries.

Kaku says chemistry is a prime target for the quantum revolution.

“Chemists who do not use quantum computers to model chemical reactions will go bankrupt,” he says. “They’ll be out of a job. They’ll be replaced by chemists who do use quantum computers. This means all medicine. All medicine can eventually be reduced to a quantum computer.”

Once quantum computers take hold, researchers could design synthetic molecules for medicines that address specific maladies.

“How do we find new drugs today? Trial and error,” Kaku says. “We have thousands of Petri dishes with chemicals in them. We tediously see whether or not they have any antibiotic properties. Why not do that in the memory of a quantum computer?”

Quantum calculations could also direct the course of gene-editing therapies with the potential of heading off diseases before they arise — an application that raises hopes as well as ethical concerns.

“Any discipline that requires the use of molecules and atoms can be helped by the quantum revolution, including cancer research, aging. Why do we die? Think about it for a moment: There are zero laws of physics that say that we have to die,” Kaku says.

Doesn’t immortality run counter to the Second Law of Thermodynamics? “If I have an open system and I use quantum computers to add extra energy from outside, I can begin the process of stopping the aging process,” Kaku says. “Think about that: the possibility of extending the human lifespan by reducing the buildup of errors in our DNA. … The applications are endless.”

He’s even hoping that next-generation computing will help him solve the mysteries of string theory and reveal the so-called Theory of Everything, which Kaku calls the God Equation. That hope is what led him to write “Quantum Supremacy” in the first place.

Kaku has been working on string theory for decades, and he’s the author of one of the leading textbooks about it. But he says the theory is “so complicated, with so many resonances, that the human mind has not been able to solve string theory.”

“What a frustrating thing,” he says. “So I said to myself, wait a minute. String theory is a quantum theory, like the atom. Why not use quantum computers to solve a quantum problem?”

By now, you’ve probably gotten the message that Kaku is bullish on the quantum revolution. Is he willing to admit there’s something that quantum computers can’t do? Yes, as a matter of fact.

If a movie like “Everything Everywhere All at Once” makes it look as if you can slip back and forth between quantum universes, Kaku says you should know that’s pure fiction. “It doesn’t work that way,” he says. “It turns out that it takes an enormous amount of energy and time to go between universes. So, believe it or not, it may be possible to go between universes, but it’s not for us.”

In other words, not even the quantum computer revolution can change everything everywhere all at once.

Source

Artificial intelligence has made incredible progress in the last decade, but in one crucial aspect, it still lags behind the theoretical computer science of the 1990s: namely, there is no essay describing five potential worlds that we could live in and giving each one of them whimsical names.  In other words, no one has done for AI what Russell Impagliazzo did for complexity theory in 1995, when he defined the five worlds Algorithmica, Heuristica, Pessiland, Minicrypt, and Cryptomania, corresponding to five possible resolutions of the P vs. NP problem along with the central unsolved problems of cryptography.

In this blog post, we—Scott and Boaz—aim to remedy this gap. Specifically, we consider 5 possible scenarios for how AI will evolve in the future.  (Incidentally, it was at a 2009 workshop devoted to Impagliazzo’s five worlds co-organized by Boaz that Scott met his now wife, complexity theorist Dana Moshkovitz.  We hope civilization will continue for long enough that someone in the future could meet their soulmate, or neuron-mate, at a future workshop about our five worlds.)

Like in Impagliazzo’s 1995 paper on the five potential worlds of the difficulty of NP problems, we will not try to be exhaustive but rather concentrate on extreme cases.  It’s possible that we’ll end up in a mixture of worlds or a situation not described by any of the worlds.  Indeed, one crucial difference between our setting and Impagliazzo’s, is that in the complexity case, the worlds corresponded to concrete (and mutually exclusive) mathematical conjectures.  So in some sense, the question wasn’t “which world will we live in?” but “which world have we Platonically always lived in, without knowing it?”  In contrast, the impact of AI will be a complex mix of mathematical bounds, computational capabilities, human discoveries, and social and legal issues. Hence, the worlds we describe depend on more than just the fundamental capabilities and limitations of artificial intelligence, and humanity could also shift from one of these worlds to another over time.

Without further ado, we name our five worlds “AI-Fizzle,” “Futurama,” ”AI-Dystopia,” “Singularia,” and “Paperclipalypse.”  In this essay, we don’t try to assign probabilities to these scenarios; we merely sketch their assumptions and technical and social consequences. We hope that by making assumptions explicit, we can help ground the debate on the various risks around AI.

AI-Fizzle. In this scenario, AI “runs out of steam” fairly soon. AI still has a significant impact on the world (so it’s not the same as a “cryptocurrency fizzle”), but relative to current expectations, this would be considered a disappointment.  Rather than the industrial or computer revolutions, AI might be compared in this case to nuclear power: people were initially thrilled about the seemingly limitless potential, but decades later, that potential remains mostly unrealized.  With nuclear power, though, many would argue that the potential went unrealized mostly for sociopolitical rather than technical reasons.  Could AI also fizzle by political fiat?

Regardless of the answer, another possibility is that costs (in data and computation) scale up so rapidly as a function of performance and reliability that AI is not cost-effective to apply in many domains. That is, it could be that for most jobs, humans will still be more reliable and energy-efficient (we don’t normally think of low wattage as being key to human specialness, but it might turn out that way!).  So, like nuclear fusion, an AI which yields dramatically more value than the resources needed to build and deploy it might always remain a couple of decades in the future.  In this scenario, AI would replace and enhance some fraction of human jobs and improve productivity, but the 21st century would not be the “century of AI,” and AI’s impact on society would be limited for both good and bad.

Futurama. In this scenario, AI unleashes a revolution that’s entirely comparable to the scientific, industrial, or information revolutions (but “merely” those).  AI systems grow significantly in capabilities and perform many of the tasks currently performed by human experts at a small fraction of the cost, in some domains superhumanly.  However, AI systems are still used as tools by humans, and except for a few fringe thinkers, no one treats them as sentient.  AI easily passes the Turing test, can prove hard theorems, and can generate entertaining content (as well as deepfakes). But humanity gets used to that, just like we got used to computers creaming us in chess, translating text, and generating special effects in movies.  Most people no more feel inferior to their AI than they feel inferior to their car because it runs faster.  In this scenario, people will likely anthropomorphize AI less over time (as happened with digital computers themselves).  In “Futurama,” AI will, like any revolutionary technology, be used for both good and bad.  But as with prior major technological revolutions, on the whole, AI will have a large positive impact on humanity. AI will be used to reduce poverty and ensure that more of humanity has access to food, healthcare, education, and economic opportunities. In “Futurama,” AI systems will sometimes cause harm, but the vast majority of these failures will be due to human negligence or maliciousness.  Some AI systems might be so complex that it would be best to model them as potentially behaving  “adversarially,” and part of the practice of deploying AIs responsibly would be to ensure an “operating envelope” that limits their potential damage even under adversarial failures. 

AI-Dystopia. The technical assumptions of “AI-Dystopia” are similar to those of “Futurama,” but the upshot could hardly be more different.  Here, again, AI unleashes a revolution on the scale of the industrial or computer revolutions, but the change is markedly for the worse.  AI greatly increases the scale of surveillance by government and private corporations.  It causes massive job losses while enriching a tiny elite.  It entrenches society’s existing inequalities and biases.  And it takes away a central tool against oppression: namely, the ability of humans to refuse or subvert orders.

Interestingly, it’s even possible that the same future could be characterized as Futurama by some people and as AI-Dystopia by others–just like how some people emphasize how our current technological civilization has lifted billions out of poverty into a standard of living unprecedented in human history, while others focus on the still existing (and in some cases rising) inequalities and suffering, and consider it a neoliberal capitalist dystopia.

Singularia.  Here AI breaks out of the current paradigm, where increasing capabilities require ever-growing resources of data and computation, and no longer needs human data or human-provided hardware and energy to become stronger at an ever-increasing pace.  AIs improve their own intellectual capabilities, including by developing new science, and (whether by deliberate design or happenstance) they act as goal-oriented agents in the physical world.  They can effectively be thought of as an alien civilization–or perhaps as a new species, which is to us as we were to Homo erectus.

Fortunately, though (and again, whether by careful design or just as a byproduct of their human origins), the AIs act to us like benevolent gods and lead us to an “AI utopia.”  They solve our material problems for us, giving us unlimited abundance and presumably virtual-reality adventures of our choosing.  (Though maybe, as in The Matrix, the AIs will discover that humans need some conflict, and we will all live in a simulation of 2020’s Twitter, constantly dunking on one another…) 

Paperclipalypse.  In “Paperclipalypse” or “AI Doom,” we again think of future AIs as a superintelligent “alien race” that doesn’t need humanity for its own development.  Here, though, the AIs are either actively opposed to human existence or else indifferent to it in a way that causes our extinction as a byproduct.  In this scenario, AIs do not develop a notion of morality comparable to ours or even a notion that keeping a diversity of species and ensuring humans don’t go extinct might be useful to them in the long run.  Rather, the interaction between AI and Homo sapiens ends about the same way that the interaction between Homo sapiens and Neanderthals ended. 

In fact, the canonical depictions of such a scenario imagine an interaction that is much more abrupt than our brush with the Neanderthals. The idea is that, perhaps because they originated through some optimization procedure, AI systems will have some strong but weirdly-specific goal (a la “maximizing paperclips”), for which the continued existence of humans is, at best, a hindrance.  So the AIs quickly play out the scenarios and, in a matter of milliseconds, decide that the optimal solution is to kill all humans, taking a few extra milliseconds to make a plan for that and execute it.  If conditions are not yet ripe for executing their plan, the AIs pretend to be docile tools, as in the “Futurama” scenario, waiting for the right time to strike.  In this scenario, self-improvement happens so quickly that humans might not even notice it.  There need be no intermediate stage in which an AI “merely” kills a few thousand humans, raising 9/11-type alarm bells.

Regulations. The practical impact of AI regulations depends, in large part, on which scenarios we consider most likely.  Regulation is not terribly important in the “AI Fizzle” scenario where AI, well, fizzles.  In “Futurama,” regulations would be aimed at ensuring that on balance, AI is used more for good than for bad, and that the world doesn’t devolve into “AI Dystopia.”  The latter goal requires anti-trust and open-science regulations to ensure that power is not concentrated in a few corporations or governments.  Thus, regulations are needed to democratize AI development more than to restrict it.  This doesn’t mean that AI would be completely unregulated.  It might be treated somewhat similarly to drugs—something that can have complex effects and needs to undergo trials before mass deployment.  There would also be regulations aimed at reducing the chance of “bad actors” (whether other nations or individuals) getting access to cutting-edge AIs, but probably the bulk of the effort would be at increasing the chance of thwarting them (e.g., using AI to detect AI-generated misinformation, or using AI to harden systems against AI-aided hackers).  This is similar to how most academic experts believe cryptography should be regulated (and how it is largely regulated these days in most democratic countries): it’s a technology that can be used for both good and bad, but the cost of restricting its access to regular citizens outweighs the benefits.  However, as we do with security exploits today, we might restrict or delay public releases of AI systems to some extent.

To whatever extent we foresee “Singularia” or “Paperclipalypse,” however, regulations play a completely different role.  If we knew we were headed for “Singularia,” then presumably regulations would be superfluous, except perhaps to try to accelerate the development of AIs!  Meanwhile, if one accepts the assumptions of “Paperclipalypse,” any regulations other than the most draconian might be futile.  If, in the near future, almost anyone will be able to spend a few billion dollars to build a recursively self-improving AI that might turn into a superintelligent world-destroying agent, and moreover (unlike with nuclear weapons) they won’t need exotic materials to do so, then it’s hard to see how to forestall the apocalypse, except perhaps via a worldwide, militarily enforced agreement to “shut it all down,” as Eliezer Yudkowsky indeed now explicitly advocates.  “Ordinary” regulations could, at best, delay the end by a short amount–given the current pace of AI advances, perhaps not more than a few years.  Thus, regardless of how likely one considers this scenario, one might want to focus more on the other scenarios for methodological reasons alone!

From