As quantum processors become larger and more sophisticated, controlling the qubits requires far more than simply sending them carefully calibrated pulses. Quantum computers must continuously measure their own state, process that information using classical computing, and respond with new instructions, often on timescales of microseconds or even hundreds of nanoseconds.

Cohen explains why Quantum Machines was founded around this problem, how FPGA-based pulse processing can replace conventional test equipment, and why tightly coupling quantum processors to classical computing may be essential for fault-tolerant quantum computing.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. The Quantum Computer Is Only Part of the Machine

The physical qubit gets most of the attention in discussions about quantum computing. But a functioning quantum computer requires an extensive classical control system to operate it.

Quantum Machines’ approach is based on what Cohen calls hybrid quantum control: tightly integrating quantum hardware with classical computing so that measurements from the quantum processor can be processed and fed back into the system with very low latency.

As Cohen explains:

“We developed a control system for quantum computers, which is the system that essentially talks to the quantum hardware and operates the quantum hardware, and it’s also the bridge between the quantum resources, eventually the quantum hardware, and classical compute resources.”

This architecture becomes increasingly important as quantum computers move toward error correction and more complicated algorithms. The quantum processor doesn’t simply execute a predetermined sequence of operations. It has to measure, interpret, and respond.

Cohen describes the goal as connecting two very different computational worlds: the quantum system performing the computation and classical processors handling the tasks for which classical computers are better suited.

II. Why Quantum Computers Need Classical Computing

One of the most interesting questions in the conversation is whether adding increasingly large amounts of classical computing to a quantum computer undermines the point of having a quantum computer in the first place.

Cohen argues that the opposite is true. Classical computing is necessary both to operate the quantum processor and to determine whether the quantum processor is actually providing an advantage.

“I actually think that it’s important to build these machines that couple a lot of classical compute to quantum compute, and then, yes, seeing that there is a real difference when you benchmark the system without the quantum accelerator and with the quantum accelerator.”

This points to a broader idea about quantum computing: the relevant machine is not just the QPU. A practical quantum computer will be a hybrid system, combining quantum processors with CPUs, GPUs, specialized control hardware, and increasingly sophisticated software.

The challenge is figuring out how to connect all of those components without introducing too much latency.

III. From Arbitrary Waveform Generators to Quantum Control Processors

When Cohen and his co-founders started Quantum Machines, quantum experiments were often controlled using general-purpose test equipment such as arbitrary waveform generators, or AWGs.

An AWG essentially stores the waveforms required for an experiment and then plays them back to the quantum processor. That works well for relatively small experiments, but Cohen explains why the approach becomes problematic at scale.

For a quantum computer with thousands or millions of control channels, the cost, physical size, power consumption, and amount of data that must be transferred to the hardware become enormous.

The alternative developed by Quantum Machines is a pulse processing unit, or PPU, implemented on FPGAs.

“We implemented it on FPGAs. Of course, the next steps would be to implement it also in silicon, on ASICs, but today we’re still an FPGA to keep flexibility and because of the volumes and everything.”

Instead of storing every individual waveform point, the PPU receives instructions describing how the waveform should be generated and produces it in real time.

That makes the system more compact and scalable—but, more importantly, it enables something conventional AWGs struggle to provide: real-time feedback.

IV. The Race to Close the Loop

Quantum error correction requires a quantum computer to continually extract information about errors occurring in the system and use that information to determine what to do next.

This means the control system needs to form a closed loop: measure the quantum processor, process the result, make a decision, and send a new operation back to the qubits.

Cohen explains why this was one of the motivations behind Quantum Machines:

“You couldn’t do any real-time feedback. I mean, you could, but you would have to go back to the software layer and generate new waveforms, and the latencies would be way more than the coherence times.”

The problem isn’t merely making the computer fast in an abstract sense. The feedback has to happen before the quantum state loses the information needed to act on it.

The episode explores how this works at several levels, from simple active reset of a qubit to more complicated feed-forward operations required for quantum error correction.

In one example, the control system measures a qubit and immediately decides whether to apply another pulse based on the result.

“You need based on the measurement to make a decision whether to play or not to play a pulse.”

That seemingly simple operation becomes extraordinarily complicated when it has to be performed across thousands or millions of qubits.

V. Nanoseconds, GPUs, and the Future of Fault-Tolerant Quantum Computing

As quantum computers scale, classical processing itself becomes a major systems-engineering problem.

Cohen describes an architecture in which specialized FPGA-based processing handles extremely latency-sensitive operations while larger computational workloads can be sent to CPUs or GPUs.

The system can currently close certain quantum-classical feedback loops in roughly four microseconds when communicating with a GPU, with even lower latencies possible for simpler processing closer to the hardware.

But the ultimate goal is to make these loops increasingly integrated into the control hardware itself.

Cohen describes the long-term direction:

“I think that in the long term, some of these things will go back actually into the control system and will be embedded in the silicon.”

This becomes particularly important for fault-tolerant quantum computing. Once error correction is running continuously, the quantum processor may need to operate for days or weeks while simultaneously correcting quantum errors and compensating for classical effects such as drift in qubit frequencies or pulse parameters.

Cohen describes a future in which calibration and error correction happen continuously while an algorithm is running:

“All of this has to happen, you know, in real time while you’re running very long programs. And I think that’s going to be extremely important.”

The implication is significant: the control system isn’t simply peripheral equipment attached to the quantum computer. It is becoming part of the architecture of the computer itself.

VI. Building the Picks and Shovels of Quantum Computing

Cohen and his co-founders started Quantum Machines in 2018, at a time when quantum computing was still far from the current level of commercial attention.

Their strategy was to build technology that quantum researchers needed immediately rather than betting on a particular qubit architecture winning the race to fault tolerance.

That strategy has allowed Quantum Machines to work across superconducting qubits, neutral atoms, trapped ions, spin qubits, and other approaches.

Cohen describes the common architecture they see across these very different physical systems:

“I would say, I don’t know, 60, 70% of the system is common across, for the control across the qubit types.”

The remaining portion has to be adapted to the specific physics of each platform—for example, cameras for neutral-atom and trapped-ion readout versus microwave measurements for superconducting qubits.

This makes quantum control an unusual position in the industry: the underlying qubits may be radically different, but many of the computational and control problems surrounding them are surprisingly similar.

VII. What Will Actually Make Quantum Computing Scale?

The conversation ultimately returns to the question that sits behind much of the quantum computing industry: what is actually preventing these machines from becoming useful?

Adding more qubits is clearly not enough. Those qubits need to be controlled, calibrated, read out, connected to classical computing resources, and eventually incorporated into fault-tolerant error-correction protocols.

Cohen argues that the quantum processor and its control system will have to evolve together.

“If we don’t build a much more scalable, in terms of, again, the cost and the power and the size, control system, and also build the right software to allow flexibility for the developers to develop also the better QPUs, then it just slows down the progress a lot.”

That may be one of the most important lessons from the episode. The path to useful quantum computing isn’t determined by the qubit alone. It depends on the entire stack of hardware and software required to make those qubits behave like a computer.

Full episode available now on YouTube and all major podcast platforms.

Listen to our full conversation with Yonatan Cohen on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. An Accidental Discovery That Changed Biology

One of the most remarkable stories in the episode is that DNA sequencing did not begin as an ambitious attempt to decode genomes. It emerged from a much smaller question.

Gilbert’s laboratory had isolated the lac repressor and wanted to know exactly which DNA sequence it recognized. Existing methods simply weren’t capable of reading DNA directly.

As Gilbert explains:

“The obvious question was, well, what is its structure? What is its sequence?”

That seemingly modest question ultimately led to one of the first practical DNA sequencing methods, opening the door to modern genomics. Gilbert reflects on the technical challenges, the surprising experiments that made sequencing possible, and how an attempt to solve one biological puzzle transformed an entire field.

II. Biology Becomes an Information Science

Long before the Human Genome Project officially began, Gilbert had already envisioned biology entering a fundamentally new era.

Rather than studying genes one at a time, he imagined every scientist having access to an organism’s complete genetic blueprint, a radical idea at the time.

The conversation traces the origins of the Human Genome Project, why Gilbert believed sequencing entire genomes would become inevitable, and how that prediction has evolved into today’s world of AI, computational biology, and massive genomic databases.

III. The RNA World and the Origins of Life

Gilbert’s 1986 RNA World paper became one of the foundational ideas in origin-of-life research.

The hypothesis emerged shortly after the discovery that RNA molecules could themselves act as enzymes, suggesting that life may once have relied on a single molecule capable of both storing information and carrying out chemistry.

As Gilbert explains:

“The RNA world is the idea that you begin with RNA, RNA molecules that copy themselves into other RNA, and therefore, you can begin an evolution truly with one molecular type.”

The discussion explores why DNA likely appeared later in evolution, what evidence continues to support the RNA World hypothesis, and which questions remain unanswered nearly forty years after the idea was first proposed.

IV. Building One of the First Biotechnology Companies

Beyond academia, Gilbert also helped launch an entirely new industry.

Biogen was founded at a time when few molecular biologists believed academic scientists should participate in commercial ventures. The episode offers a fascinating firsthand account of biotechnology’s earliest days: raising venture capital, defining research priorities, and balancing scientific curiosity with building a company.

Gilbert summarizes the original mission simply:

“The initial idea is let’s make useful human proteins in bacteria.”

The conversation also explores how Biogen grew alongside recombinant DNA technology and why translating discoveries into medicines required scientists to rethink their relationship with industry.

V. Science, Art, and Reinvention

One of the most enjoyable parts of the interview comes near the end, where Gilbert reflects on creativity after leaving science to pursue digital abstract art.

Rather than viewing science and art as separate pursuits, he sees both driven by the desire to create something genuinely new.

As he puts it:

“In the science, the drive... is to find something new. But it has to be new and true... And there’s a similarity in the art. The driving impulse in the art is to create something that is new.”

That philosophy also shaped his scientific career. Rather than remaining in a single discipline, Gilbert repeatedly reinvented himself; from theoretical physics to molecular biology, from DNA sequencing to evolutionary theory, biotechnology, computational biology, and finally art.

His advice for young scientists is equally timeless:

“Don’t work on something you think is a good idea that you don’t like, but work in a field that you enjoy all the detail of.”

It is a fitting conclusion to a conversation with someone whose career has been defined by following curiosity into entirely new fields.

Full episode available now on YouTube and all major podcast platforms.

Listen to our full conversation with Walter Gilbert on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Long before molecular machines became a popular topic, Manson was helping uncover the experiments, genetics, and biophysics behind a question that seemed almost impossible at the time: how can a living cell build a true rotary motor?

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. A Fifty-Year Scientific Journey

Few scientists have the opportunity to watch an entire field mature from its earliest experiments to a complete molecular mechanism.

Reflecting on that journey, Manson says:

“My life’s quest is fulfilled. I now after 50 years, I finally know how what I knew 50 years ago actually works out mechanistically, and that’s about as cool as it gets.”

The conversation traces that history through the pioneering work of Howard Berg and many others who demonstrated that bacterial flagella truly rotate, identified the proteins responsible for chemotaxis, and ultimately used cryo-electron microscopy to reveal exactly how torque is generated.

It is a rare firsthand account from someone who participated in nearly every stage of that scientific story.

II. Nature Built a Real Rotary Motor

Bacteria contain genuine rotary nanomachines powered by proton flow across the cell membrane.

As Manson puts it:

“They’re all over the place... the ATPs you have in your mitochondria that every cell has, it’s also a rotary motor.”

The bacterial flagellar motor converts the proton motive force into rotation using a precisely organized collection of rotor and stator proteins. Rather than sliding like muscles, these molecular components rotate continuously, allowing bacteria to swim through environments where conventional propulsion simply would not work.

III. Why Bacteria Swim the Way They Do

The episode also explores one of the most elegant examples of behavior emerging from molecular biology.

Rather than steering directly toward food, E. coli performs a biased random walk by alternating between smooth swimming (”runs”) and random reorientation (”tumbles”).

As Manson explains:

“The random walk is biased and, and that’s how they actually move in a gradient.”

A single signaling pathway controls whether the flagellar motor rotates clockwise or counterclockwise. Counterclockwise rotation bundles multiple left-handed flagella together to propel the cell forward, while clockwise rotation causes the bundle to fly apart, randomly changing the bacterium’s direction before it resumes swimming.

IV. Why Rotation Works at the Molecular Scale

One of the most fascinating parts of the discussion is that rotary motors only make sense because bacteria are extraordinarily small.

At low Reynolds numbers, propulsion follows completely different physical rules than the macroscopic world we experience every day.

As Manson explains:

“You get a lot of benefits of being small.”

The episode explores why molecular motors do not require lubrication, why reciprocal motion fails at microscopic scales, and why evolution repeatedly arrived at rotating helical filaments as an effective solution for bacterial locomotion.

V. Evolution Finds Similar Solutions

The conversation concludes by examining one of evolution’s most striking examples of convergence.

Bacteria and archaea independently evolved rotating helical filaments to swim, even though the motors that power them are fundamentally different.

As Manson summarizes:

“Two different lineages, two totally separate evolutionary lineages... found a way to move by rotating the helical filament.”

It is a reminder that while evolution has no foresight, the laws of physics strongly constrain which engineering solutions are possible. Given similar physical challenges, life can arrive at remarkably similar designs through entirely different evolutionary paths.

Whether you’re interested in microbiology, molecular biology, evolution, biophysics, nanotechnology, or the history of scientific discovery, this episode offers a fascinating look at one of nature’s most sophisticated molecular machines, and the fifty-year scientific journey that finally revealed how it works.

Full episode available now on YouTube and all major podcast platforms.

Listen to our full conversation with Michael Manson on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Over the past twenty years, neutral atoms have evolved from a relatively small academic research area into one of the leading architectures for scalable quantum computing. Today, companies like QuEra, Infleqtion, Atom Computing, and Pasqal are racing to build increasingly capable neutral atom processors—but that wasn’t always the case.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. From Niche Research to a Global Race

When Saffman first began working on neutral atoms in the early 2000s, the field was largely overshadowed by trapped ions and superconducting qubits.

As he recalls:

“There was just a very small community of people really pushing to do qubits with neutral atoms... But we kept plugging away.”

Everything changed around 2015. Improvements in optical tweezers, atom rearrangement, and Rydberg gates demonstrated that neutral atoms could scale far beyond the small proof-of-concept experiments that had defined the field.

Today, what was once a niche direction has become one of the fastest-growing areas in quantum computing.

II. Why Rydberg Atoms?

At the heart of the neutral atom platform is a remarkable piece of atomic physics.

Ordinary atoms interact only weakly with one another, making them excellent quantum memories but poor candidates for computation. The solution is to briefly excite them into enormous Rydberg states.

As Saffman explains:

“These Rydberg atoms have very exaggerated properties.”

A Rydberg atom can be thousands of times larger than an atom in its ground state, allowing neighboring atoms to interact strongly through the Rydberg blockade mechanism. Those interactions form the basis of quantum logic gates and entanglement.

III. The Challenge Isn’t Physics… It’s Engineering

Much of today’s progress is no longer about discovering new physical phenomena.

Instead, researchers are working to eliminate tiny sources of noise that limit gate performance.

Discussing laser requirements, Saffman notes:

“We need lasers with very low phase noise and frequency noise to really get a high quality entanglement.”

As gate fidelities continue to improve, increasingly subtle engineering details—from laser stability to atomic motion—become the dominant obstacles to building practical quantum computers.

IV. Scaling Beyond Thousands of Qubits

Simply building larger arrays is not enough.

Future quantum computers must perform error correction efficiently while minimizing unnecessary operations.

One particularly promising idea discussed in the episode is using two different atomic species so that error-correction measurements can be performed without disturbing neighboring data qubits.

More broadly, Saffman emphasizes that the field is still searching for the best large-scale architecture:

“These architectural questions... are at the research forefront and have not been resolved as to what’s gonna be the best way to really scale up these systems.”

The conversation highlights how the next generation of advances may come as much from computer architecture as from atomic physics.

V. A Turning Point for Neutral Atoms

Perhaps the most striking moment comes when Saffman reflects on when the field began to feel commercially viable.

Looking back on the mid-2010s, he says:

“It was pretty clear to me that this wasn’t just a science experiment.”

That realization helped motivate his involvement with ColdQuanta (now Infleqtion) and reflected a broader shift across the quantum industry. Neutral atoms were no longer simply an interesting research platform. They had become serious contenders for building useful quantum computers.

Whether neutral atoms ultimately become the dominant architecture remains an open question. But after two decades of steady progress, they have firmly established themselves among the leading approaches in the race toward fault-tolerant quantum computing.

Full episode available now on YouTube and all major podcast platforms.

Listen to our full conversation with Mark Saffman on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

As AI systems continue to grow in scale, the industry is discovering that computation is only part of the challenge. Moving data between processors, racks, and data centers has become one of the most important bottlenecks in modern computing. Bowers explains why silicon photonics is increasingly viewed as a solution to this problem, how the field evolved from a niche research area into a commercial industry, and why the future of AI may depend on communication technologies built around light rather than electricity.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. The Bandwidth Problem

Much of the public discussion around AI focuses on GPUs, accelerators, and model architectures. But according to Bowers, the challenge of moving information is becoming just as important as processing it.

Modern AI systems require enormous amounts of communication between processors. As bandwidth requirements continue to increase, traditional electrical interconnects are struggling to keep pace.

As Bowers explains:

“They require tremendous amounts of data, right? Tens of terabits for a modern chip, and you just can’t deliver that with copper.”

The problem is fundamentally one of physics. Copper traces become increasingly lossy at higher frequencies, requiring additional power and repeated signal regeneration. Optical communication, by contrast, can move information across vastly different distances with comparatively little penalty.

The result is that communication infrastructure is becoming a central consideration in AI system design.

II. Why Silicon Photonics Won

One of the most interesting parts of the conversation explores how silicon photonics emerged despite widespread skepticism from the photonics community.

For years, many researchers argued that silicon was fundamentally the wrong material for photonic systems. Silicon is an indirect-bandgap semiconductor, making it a poor laser material and seemingly an unlikely platform for optical communications.

Yet advances in heterogeneous integration, modulators, manufacturing, and packaging gradually changed that perception.

Looking back on the field’s evolution, Bowers recalls:

“There were a lot of different vectors that had to happen to convince people to switch to silicon photonics.”

Today, silicon photonics has become the dominant platform for many optical communication applications, benefiting from the manufacturing ecosystem developed for the semiconductor industry while incorporating high-performance III-V materials where optical gain is required.

III. AI Is Accelerating the Transition

The rise of AI has dramatically increased demand for bandwidth inside data centers.

Historically, optical communication was used primarily between racks and across long distances. Electrical connections remained dominant within servers and racks because they were simpler and cheaper.

That balance is now changing.

Bowers argues that the industry is approaching a major architectural transition:

“It’ll be photonically integrated, and it just saves you so much power.”

The shift is already visible in the emergence of co-packaged optics, where optical transceivers are integrated directly alongside high-performance switching chips.

As bandwidth requirements climb from tens of terabits per second toward hundreds of terabits per second, reducing electrical losses becomes increasingly important.

IV. From Pluggable Optics to Optical Computing

The conversation also explores the roadmap for future optical integration.

Data center networking has already progressed from pluggable optical modules to onboard optics and now to co-packaged optics integrated within the same package as switching silicon.

According to Bowers, the trend is continuing toward even tighter integration:

“The next step is on the chip itself communicating different portions optically.”

In this vision, optical communication will no longer be limited to connections between racks or servers. Future systems may use photonics to connect different regions of the same processor package and eventually different functional blocks on the same chip.

The distinction between networking and computing infrastructure could begin to blur.

V. Building Computers Connected by Light

Toward the end of the discussion, Bowers describes just how quickly bandwidth demands are increasing.

The industry’s largest switching chips have already reached extraordinary levels of throughput, and those numbers continue to grow.

As he notes:

“They’re selling 100 terabit chips, and in two years it’s gonna be 200 terabit chips.”

Meeting those requirements with traditional electrical interconnects would require enormous amounts of power and increasingly complex signal conditioning.

Silicon photonics offers a different path: one where information is carried by light, optical communication moves closer and closer to the processor, and future AI systems are built around photonic connectivity from the start.

The conversation provides a detailed look at one of the most important technologies behind modern computing, where semiconductor manufacturing, optical physics, networking, and AI infrastructure are converging to shape the future of data centers.

Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with John Bowers on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

The conversation explores the origins of Levin’s interest in bioelectricity, the experimental tools used to manipulate electrical signaling in living tissues, and how bioelectric circuits may function as a higher-level control system operating alongside genes and biochemical pathways.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. From Electronics to Developmental Biology

Levin’s fascination with bioelectricity began long before graduate school. As a child interested in both electronics and biology, he became captivated by a question that would eventually define his career: how do collections of individual components become organized systems with goals, behaviors, and structure?

A major turning point came when he encountered Robert Becker’s influential book The Body Electric.

“It occurred to me that this was, kind of a perfect merger of the things I was interested in.”

What attracted Levin was the possibility that electrical signaling might explain how developing tissues coordinate large-scale decisions. Neuroscience had already demonstrated the power of electrophysiology in nervous systems, but Levin wondered whether similar principles might operate throughout the body during development.

This perspective led him toward a field that, at the time, sat far outside the biological mainstream.

II. Building the Bioelectric Toolbox

One of the most interesting sections of the discussion focuses on the experimental techniques required to study electrical signaling in embryos.

Rather than relying on external electromagnetic fields, Levin wanted tools that could directly manipulate the endogenous electrical states generated by cells themselves.

“And so ion channels, ion pumps… as a postdoc, I began to assemble this toolkit.”

That toolkit combined voltage-sensitive dyes for imaging membrane potentials with molecular biology approaches that allowed specific ion channels and pumps to be expressed in targeted regions of developing embryos.

The goal was not simply to perturb development, but to read and write bioelectric information directly.

Levin describes how these methods enabled researchers to visualize electrical patterns across entire embryos and test whether specific voltage states influenced developmental outcomes. In many ways, the approach treats development as an information-processing problem, where electrical states serve as signals that cells can interpret and act upon.

III. Voltage Gradients and the Problem of Pattern Formation

A central theme throughout the episode is the challenge of understanding how cells determine where structures should form.

Genes can specify proteins, but they do not directly explain how embryos establish large-scale anatomical organization. Levin became particularly interested in questions such as left-right asymmetry and tissue patterning, where cells must coordinate decisions across significant distances.

As his work progressed, he discovered that cells were not responding simply to absolute voltages.

“Gradients are what the cells interpret.”

This insight shifted the focus toward spatial patterns of electrical activity. According to Levin, many of the important developmental decisions occur at the boundaries between voltage domains, where tissues can detect differences rather than fixed electrical values.

The result is a picture of development that resembles distributed computation. Individual cells possess only local information, yet large collections of cells can coordinate to generate coherent body structures through electrical communication networks connected by gap junctions.

IV. Two-Headed Worms and Anatomical Memory

Perhaps the most famous experiments discussed in the episode involve planarian flatworms.

Planaria can regenerate entire bodies from small tissue fragments, making them an ideal system for studying how tissues store information about anatomy.

By manipulating bioelectric signaling, Levin’s group demonstrated that worms could be induced to regenerate with two heads instead of a head and tail. Even more remarkably, the altered anatomy persisted through future rounds of regeneration.

“If you recut them, they make more two-headed worms.”

The significance of this result goes beyond regeneration itself. The worms retained the memory of a different body plan despite having unchanged DNA sequences.

For Levin, this suggests that bioelectric circuits may function as a form of pattern memory, storing information about target anatomy in ways that are distinct from genetic information alone.

The finding raises profound questions about where biological systems encode large-scale structure and how tissues remember what they are supposed to build.

V. Communicating with Cells

Toward the end of the conversation, Levin describes what he sees as the broader implication of bioelectricity research.

Rather than micromanaging every molecular pathway involved in development, bioelectric signals may provide a higher-level interface for communicating with living tissues.

“We are communicating to the cellular collective, and we talk about very abstract, high-level things.”

Whether discussing ectopic eye formation, brain repair in frog embryos, or regeneration after injury, Levin emphasizes that cells often possess latent competencies that can be activated by appropriate signals.

The challenge is not necessarily learning how to build tissues from scratch, but understanding the communication channels through which tissues coordinate their own construction.

From this perspective, bioelectricity becomes more than a developmental mechanism. It becomes a language through which multicellular systems organize themselves across scales.

The conversation offers a detailed look at one of the most unconventional and rapidly growing areas of modern biology, where electrophysiology, developmental biology, systems theory, and regeneration research converge.

Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Michael Levin on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

The conversation spans compiler design, heterogeneous quantum hardware, error correction, biomedical optimization, and the long-term relationship between quantum and classical computing systems.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. From Classical Architecture to Quantum Systems

Before becoming one of the leading figures in quantum computer architecture, Chong worked on classical supercomputers and emerging computing technologies during the 1990s.

Rather than approaching quantum computing primarily from theoretical physics, he entered the field through systems engineering and architecture.

“My emphasis as a classical architect was often in emerging technologies.”

That background shaped how he thinks about quantum systems today. Instead of focusing only on qubits in isolation, Chong became interested in the broader computational stack surrounding them: compilers, communication, scheduling, topology, and physical hardware constraints.

His transition into quantum computing began through conversations with Isaac Chuang, who encouraged him to think about scaling quantum systems beyond individual devices.

“You know, it’s time we thought about the architecture of these machines instead of, like, two devices at a time.”

That shift reflects a major transformation within the field itself. Early quantum computing was dominated by demonstrations of small systems and isolated quantum effects. Architecture asks a different question: how do you build large, programmable computational systems from fragile physical hardware?

For Chong, the challenge was always inherently interdisciplinary.

II. The Legacy of the NISQ Era

A major theme throughout the conversation is the importance of the NISQ era: noisy intermediate-scale quantum computing.

Although the field is now increasingly focused on fault-tolerant machines, Chong argues that the NISQ period served as a critical systems engineering phase that forced researchers to work across every layer of the stack simultaneously.

“We had this very sort of perforated stack of leaky abstractions.”

In classical computing, abstractions hide hardware complexity from higher-level software. NISQ systems did the opposite. Algorithm performance depended directly on coherence times, hardware connectivity, calibration quality, and pulse-level control.

That environment encouraged what Chong’s group called “physics-aware software”: compilers and optimizers explicitly designed around the properties of particular hardware platforms.

Rather than treating software and hardware as separate domains, the NISQ era pushed researchers toward full-stack co-design involving physicists, compiler engineers, experimentalists, and application scientists simultaneously.

Many of the techniques developed during this period, including pulse optimization, qubit placement, and communication-aware compilation, are now being adapted for fault-tolerant architectures.

III. Fault Tolerance and Specialized Quantum Accelerators

As quantum computing transitions toward error-corrected systems, the architecture problem changes fundamentally.

In NISQ systems, minimizing noisy two-qubit gates dominated optimization efforts. In fault-tolerant systems, logical operations and error-correction overhead become central constraints.

“In the NISQ era, you’re basically most concerned about two-qubit gates, and in the fault-tolerant era, you’re mostly concerned about T gates.”

This changes the cost model of computation itself.

Compiler design must now account for logical qubit placement, syndrome decoding, communication between encoded regions, and the enormous overhead associated with maintaining fault tolerance.

At the same time, Chong emphasizes that quantum computers are unlikely to replace classical machines outright.

“Quantum computers are very much very specialized accelerators.”

Instead, he describes a future where QPUs function more like GPUs or TPUs: highly specialized co-processors integrated into larger classical computing systems.

That perspective reframes quantum computing as part of a broader history of heterogeneous computing architecture rather than a standalone replacement for classical infrastructure.

IV. Quantum Optimization for Cancer Biomarkers

One of the most compelling sections of the episode focuses on Chong’s recent work applying quantum optimization techniques to biomedical discovery.

The project studies multimodal datasets involving DNA, mRNA, and pathology imaging to identify combinations of biomarkers capable of predicting cancer treatment outcomes.

“What we’re trying to do is find features in those three data sets that, taken together, can predict treatment outcomes for cancers.”

The challenge is combinatorial scaling.

Traditional biomarker methods often examine individual variables independently. But when interactions between multiple biomarkers are considered simultaneously, the search space grows exponentially.

To address this, Chong’s group explores hybrid quantum-classical optimization strategies inspired by QAOA-style algorithms that recursively simplify large optimization problems.

Importantly, the work is framed cautiously. The goal is not immediate universal quantum advantage, but identifying domains where quantum-assisted optimization may eventually outperform existing heuristics.

The episode repeatedly emphasizes that useful quantum applications may emerge gradually through hybrid workflows rather than sudden disruptive breakthroughs.

V. Building the Quantum Software Stack

Toward the end of the discussion, Chong reflects on the current state of quantum software infrastructure and where the field still lacks maturity.

“Right now every application that we look at is like a collaboration between an expert group of quantum people and an expert and a bunch of scientists.”

That observation captures how early-stage the ecosystem remains.

Unlike mature classical software development, most quantum applications still require close collaboration between domain scientists, physicists, compiler researchers, and hardware specialists.

Stable abstractions have not fully emerged because the underlying hardware and error-correction strategies are still evolving rapidly.

For Chong, this is not a sign of failure. It is characteristic of an emerging computing paradigm still searching for its foundational applications, abstractions, and architectures.

Throughout the episode, he presents quantum computing not as a single technological breakthrough, but as a deeply layered systems problem requiring simultaneous advances in hardware, software, architecture, and algorithms.

The future of the field, in this view, will depend as much on systems engineering and compiler design as on the qubits themselves.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Fred Chong on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. Measuring Below the Scale of a Particle

At the core of quantum sensing is a counterintuitive requirement: detecting signals smaller than the fundamental units themselves.

In Yacoby’s work, this is achieved through devices like single-electron transistors, which convert tiny changes in electrostatic potential into measurable current.

“The sensitivity… is roughly 10 to the minus four of one electron.”

This level of precision is necessary because measurements are rarely direct. When probing a nanoscale system, only a fraction of an electron’s electric field reaches the detector. The challenge is not just detecting charge, but detecting influence: small perturbations in a noisy environment.

These sensors effectively act as transducers, translating quantum-scale properties into classical signals that can be recorded and analyzed.

II. Momentum, Tunneling, and One-Dimensional Physics

One of the most powerful measurement techniques discussed in the episode emerges from tunneling experiments in one-dimensional electron systems.

In these systems, electrons are confined so tightly that their behavior is governed by collective excitations rather than individual particle motion. Understanding their structure requires resolving both energy and momentum.

A key insight comes from how magnetic fields influence tunneling between parallel quantum wires.

“What happens is that one dispersion shifts in momentum relative to the other.”

This shift allows researchers to align electronic states between two systems. When alignment occurs, tunneling is enhanced, producing a measurable signal. By systematically varying the magnetic field and bias, it becomes possible to reconstruct the full dispersion relation of the system.

This technique reveals a defining feature of one-dimensional physics: the separation of spin and charge into distinct excitations with different velocities.

III. Charge Quantization and Detection

The ability to measure charge at the single-electron level relies on a fundamental property: charge is discrete.

Single-electron transistors exploit this through Coulomb blockade, where electrons can only enter or leave a small island one at a time. This produces sharp conductance peaks corresponding to individual charge transitions.

“You could see each electron being added to this device one by one.”

Between these peaks, current is suppressed, creating a highly sensitive regime where small changes in the environment shift the system in and out of conduction.

By operating at these transition points, the device becomes an extremely precise charge detector, capable of resolving minute variations in nearby systems.

IV. Spin, Fields, and NV Centers

While charge sensors probe electrostatic properties, other quantum sensors target magnetic fields and spin.

Nitrogen-vacancy (NV) centers in diamond provide one such platform. These defects have well-understood energy levels that shift in response to magnetic fields, enabling precise measurements through optical readout.

“You see a dip in fluorescence… from which then you know what B is.”

By monitoring how the fluorescence changes as a function of microwave frequency, researchers can extract the local magnetic field with high spatial resolution.

This approach enables nanoscale magnetic imaging, including the possibility of reconstructing three-dimensional spin distributions in a material.

V. From Weak Interactions to Strong Correlations

A central theme in the episode is how measurement techniques must adapt across different physical regimes.

In weakly interacting systems, individual particles can be isolated and controlled, forming the basis of qubits in quantum dots and related platforms.

In strongly correlated systems, however, interactions dominate, and new emergent phenomena arise. One example is the fractional quantum Hall effect, where collective behavior leads to quasiparticles with fractional charge.

“Electrons are really not the elementary particles… charge waves are basically the elementary excitations.”

Probing these systems requires sensors that can access collective properties rather than individual particle behavior.

VI. Bridging Measurement and Quantum Technology

The same tools used to study fundamental physics are increasingly relevant for quantum technologies.

Charge sensors enable readout of qubit states. Spin-sensitive probes allow detection of local fields and coherence. Techniques like momentum-resolved tunneling provide insight into the structure of quantum materials.

Across all of these applications, the underlying principle is consistent: quantum sensors extend what is experimentally accessible.

They allow researchers to move from indirect inference to direct measurement, revealing properties that would otherwise remain hidden.

Modern condensed matter physics is not just defined by the systems it studies, but by the precision with which those systems can be measured. As this conversation shows, advances in quantum sensing are not just incremental improvements; they redefine what can be observed, and therefore what can be understood.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Amir Yacoby on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. Flattening Optics

Metasurfaces have been a theoretical idea for years: if you could control light with nanostructures on a flat surface, you could replace complex lens systems entirely. The challenge wasn’t just physics, it was manufacturing.

As Devlin puts it:

“It’s that shrinking it all, flattening it all down to a single surface, and then making it really easy to mass produce in the semiconductor fabs. That is the real innovation.”

The key shift is not just optical performance, but the ability to fabricate these devices using the same infrastructure as chips, turning optics into something that can scale.

II. From One Device to Ten Million

A recurring theme in the conversation is the gap between academic success and industrial reality. In research, building a single high-performing device is enough. In consumer electronics, consistency is everything.

Devlin describes that transition:

“You couldn’t just target the nominal anymore. You had to target the spread that was required to get good yield.”

This means designing not just for ideal conditions, but for manufacturing variability, ensuring millions of devices all perform within tight tolerances.

III. Seeing More Than Intensity

Traditional imaging systems capture intensity and color. Metasurfaces enable access to a fundamentally different signal: polarization.

That matters because polarization encodes information about material properties, not just appearance.

“Polarization is really good at telling you what objects are made of.”

This turns a camera into something closer to a sensor for physical properties capable of distinguishing between materials that might otherwise look identical.

IV. Why Spoofing Fails

Most existing facial recognition systems rely on geometry, that is, mapping the shape of a face. But shape alone can be replicated.

Polarization introduces a new constraint: material.

“Even if the 3D shape is perfect… the material difference is picked up… and it will immediately reject it just based off of that material property.”

In other words, a mask can match your face’s geometry, but it can’t replicate how your skin interacts with light. That difference becomes a powerful signal for security.

V. A New Information Channel

Beyond biometrics, Devlin emphasizes that polarization is just one example of a broader shift in sensing.

“It’s a new information channel.”

Once these signals are available at scale, in phones, laptops, and other devices, they can feed into machine learning systems and enable entirely new applications.

The implication is that we’re not just improving cameras. We’re expanding what it means to “see.”

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Rob Devlin on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

As demand for compute accelerates, the limiting factor is no longer just chips or algorithms, but the ability to generate and deploy energy at scale.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

Subscribe now

I. The Constraint Is Energy

The rapid growth of AI systems has exposed a fundamental bottleneck: energy infrastructure.

While it’s easy to focus on GPUs and models, Johnston emphasizes that scaling compute ultimately means scaling power generation. On Earth, that process is slow and constrained. Large energy projects face years-long permitting timelines, land-use conflicts, and regulatory friction.

The result is a mismatch between how quickly demand is growing and how slowly new capacity can be brought online. Even when the technology exists, deployment becomes the limiting step.

II. Why Space Changes the Economics

Space offers a radically different operating environment, one where several major cost drivers disappear.

There is no need for land acquisition or permitting. Solar energy is far more consistent and efficient, with near-continuous exposure and no atmospheric losses. And because energy is generated continuously, there is little need for large-scale storage or backup systems.

As Johnston puts it:

“One square meter of solar panel in space produces eight times the energy of one square meter… on Earth.”

That difference compounds quickly. Higher efficiency, combined with fewer infrastructure requirements, shifts the economics of energy production in a way that is difficult to replicate terrestrially.

III. Competing on Infrastructure and Cost

Data center costs break into three major components: chips, infrastructure, and energy. The chips remain the same whether they’re on Earth or in orbit, but the other two categories are where space offers leverage.

On Earth, infrastructure (cooling systems, backup power, and physical facilities) can rival the cost of the compute itself. Energy then adds a recurring operational expense.

In space, both are reduced. Cooling becomes a matter of radiating heat into the vacuum rather than running complex systems. Energy, once deployed, is effectively free. The optimization target shifts from operating cost to upfront deployment.

The key insight is that even if chips dominate total cost, improving the rest of the system still yields large overall gains.

IV. Launch Cost Is the Gatekeeper

All of this depends on one variable: the cost of getting mass into orbit.

Today’s launch costs are still too high for space-based data centers to be broadly competitive. But that is changing quickly with the development of fully reusable rockets.

“The break-even point is around $500/kg.”

If launch costs fall below that threshold, and potentially much lower, the economics flip. What is currently impractical becomes not just viable, but advantageous.

This is the central bet: that launch costs will continue to drop far enough, fast enough, to unlock entirely new classes of infrastructure.

V. From Prototype to Paradigm

Star Cloud has already taken early steps toward testing this idea, launching a spacecraft carrying off-the-shelf GPUs to validate that compute can run in orbit.

The current systems are experimental, using improvised thermal management and limited power. But they demonstrate a key point: the challenge is not whether computation can happen in space, but how to scale it efficiently.

From there, the problem becomes one of engineering and manufacturing. Building lighter radiators, optimizing power density, and designing systems that can operate reliably over time.

The broader vision follows naturally. If energy is cheaper and more scalable in space, and if launch becomes sufficiently inexpensive, then compute will move to where it is most efficient to run.

The Bigger Picture

This isn’t just a story about space, it’s a shift in how we think about computing infrastructure.

For decades, progress has been driven by improvements in chips and algorithms. But as systems scale, the surrounding constraints, energy, cooling, land, become just as important.

Space changes those constraints.

“If it’s cheaper to do it in space, we’ll be doing it in space.”

That framing removes some of the speculation. The question is no longer whether space-based data centers are possible, but whether the economics will favor them.

And if they do, the location of computation itself may fundamentally change.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Philip Johnston on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Gidney’s work spans quantum algorithms, error correction, and implementation. Across that stack, his core insight is consistent: once you try to build these systems, the bottlenecks shift from elegant theory to very concrete questions about circuits, gates, and arithmetic.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. The Problem with “Factoring 21”

Small-scale demonstrations of Shor’s algorithm often claim to factor numbers like 21. But these experiments can be misleading about what’s actually difficult.

Gidney points out that the key test isn’t whether a period is found, but whether the algorithm is implemented faithfully:

“If there are papers that claim they factor 21 with the quantum computer… did they do the multiplication under superposition?”

In many cases, they don’t. Instead, they replace modular multiplication—the core subroutine—with something easier that reproduces the same periodic structure.

“They’ve done things that are sort of inspired by Shor’s algorithm… replaced the expensive part… with something that happens to have the same period.”

This shortcut preserves the appearance of the algorithm while avoiding the part that actually scales poorly. The result is that these demonstrations say very little about how hard factoring really is on a quantum computer.

II. Why Factoring 15 Is Unusually Easy

The prevalence of factoring 15 isn’t just historical—it’s structural. That specific case is dramatically simpler than almost any other.

“Multiplication mod 15 is really cheap… but multiplication mod 21 and most other numbers… the number of gates you need goes up by like a factor of a hundred.”

The reason is subtle but important. Modulo 15, many operations reduce to simple permutations of bits, rather than conditional logic between them.

“They don’t actually interact with each other. They just move.”

This means the circuit avoids the entangling operations that dominate cost in general cases. As soon as you move to numbers like 21, that simplification disappears, and the true scaling behavior emerges.

This gap—between toy demonstrations and real instances—is one of the central challenges in communicating progress in quantum computing.

III. Arithmetic Is the Real Work

At a conceptual level, quantum algorithms are framed in terms of interference and superposition. But when implemented, they are dominated by something far more classical: arithmetic.

“They come up in… a lot of algorithms… additions and like table lookups and little subroutines like this.”

These operations appear everywhere, from factoring to chemistry simulation, and they must be implemented reversibly and fault-tolerantly.

“Most of the stuff I’ve worked on has been arithmetic dominated.”

This creates a disconnect between how algorithms are taught and how they behave in practice. The “quantum” part provides the speedup, but the bulk of the circuit is spent managing data: moving, adding, and transforming it in ways that preserve coherence.

IV. The True Cost: Fault-Tolerant Gates

Once error correction is introduced, the cost model becomes highly uneven. Some operations are trivial, while others are extremely expensive.

“It’s very common for error correcting codes to implement the Clifford set… but… the T gate… not really clear how to do… other than via… magic state injection and distillation.”

Because of this, entire optimization strategies have historically focused on minimizing T gates.

“If a T gate is a hundred times more expensive… you can basically just count the T gates… to estimate the cost.”

Even as implementations improve and that ratio shrinks, the asymmetry remains. Certain gates fundamentally require more overhead, and that shapes everything from circuit design to architecture.

Attempts to bypass this cost have largely struggled in practice:

“I’ve generally been kind of skeptical… they tend not to end with numbers.”

What survives are methods that work under realistic noise and scaling assumptions—most notably magic state distillation, which leverages the ability to discard faulty states:

“Being allowed to throw things away… is a major advantage.”

V. Building Intuition with Tools

As circuits become more complex, reasoning about them by hand quickly becomes infeasible. Gidney’s response has been to build tools that make these systems visible and testable.

STIM, for example, emerged from frustration with overly idealized simulations:

“No one was doing circuit noise simulations… they were mostly doing… code capacity simulations.”

But these simplified models can lead to large errors in understanding real performance:

“The threshold of the surface code is 10 times lower under circuit noise.”

Crumble tackles a different problem—understanding how information and errors move through a circuit:

“What Crumble does is… show you how Pauli errors would propagate through the circuit.”

This kind of visibility fundamentally changes how circuits are designed. Instead of reasoning abstractly, you can experiment, test, and iterate:

“It would’ve been impossible without that tool, at least for me.”

The same philosophy extends to learning. Gidney’s own entry into quantum computing came from building simple simulations and gradually refining them:

“You have to be able to know if the output is correct.”

That ability—to construct, test, and verify—is what ultimately bridges the gap between theory and working systems.

The Bigger Picture

Quantum computing progress is often described in terms of breakthroughs. But in practice, it’s the accumulation of many small, precise improvements that drives the field forward.

“They all kind of stack up… and then… they can make a big difference.”

From arithmetic optimizations to better error correction schemes, each layer contributes to reducing the cost of real algorithms.

What emerges is a clearer picture: quantum computing isn’t just about exploiting physics. It’s about engineering—where the details of circuits, gates, and implementations determine what’s actually possible.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Craig Gidney on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Südhof’s work uncovered the molecular machinery that allows synapses to release neurotransmitters in response to an electrical impulse, often in less than a millisecond. His discoveries helped reveal how brain cells communicate with the speed and reliability required for thought, perception, and behavior.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. Synapses Are Where the Brain Computes

Neurons are often described as the fundamental units of the brain, but much of the brain’s processing actually happens at the connections between them.

Südhof describes the synapse as the interface where one neuron communicates with another:

“A synapse is the point at which two nerve cells in the brain form a connection that enables one nerve cell to send information to the other nerve cell.”

But synapses don’t simply pass signals along like wires in a circuit. Each connection modifies the signal before passing it on, effectively performing a small computation.

As Südhof puts it, the information “is never just transferred. It’s actually changed during the transfer.”

This perspective reframes the brain as a network of microscopic processing units distributed across trillions of synapses.

II. Converting Electricity into Chemistry

The signal traveling through a neuron is electrical: an action potential that propagates along the cell membrane.

But when the signal reaches a synapse, it must be converted into a chemical message.

Südhof explains the basic principle simply:

“At a synapse, the fundamental principle by which it transmits a signal is that it secretes a chemical messenger called a neurotransmitter.”

When the electrical impulse arrives at the end of the neuron, it triggers the rapid release of neurotransmitters stored inside tiny membrane-bound vesicles. These molecules diffuse across the synaptic cleft and bind to receptors on the neighboring cell.

The result is a new electrical signal in the postsynaptic neuron, completing the cycle of communication.

III. The Millisecond Problem

One of the major puzzles in neuroscience was how quickly this process occurs.

The arrival of an electrical impulse triggers neurotransmitter release in less than a millisecond—a remarkably short time for such a complex molecular process.

As Südhof recalls, this question became central to understanding synaptic transmission:

“How is it possible for an action potential that depolarizes the membrane to cause the secretion of the neurotransmitters within less than a millisecond?”

Answering this required identifying the proteins that organize vesicles, detect incoming calcium ions, and drive membrane fusion.

Over decades of work, researchers—including Südhof and his collaborators—mapped the molecular machinery that makes this speed possible.

IV. Calcium as the Trigger

The key signal linking electrical activity to chemical release is calcium.

When an action potential reaches the synapse, it opens voltage-gated calcium channels in the presynaptic membrane. Calcium ions rush into the cell and bind to specialized proteins that trigger vesicle fusion.

Südhof summarizes the critical step:

“Once the calcium channel opens… it binds to the synaptic tags, and that then causes immediate fusion releasing the vesicle and content.”

This mechanism ensures that neurotransmitter release occurs precisely when an electrical signal arrives, allowing neural circuits to operate with millisecond precision.

V. Why the Brain Uses Chemical Synapses

At first glance, converting electrical signals into chemical ones might seem unnecessarily complicated.

But chemical synapses provide a crucial advantage: flexibility. Unlike simple electrical connections, synapses can change their strength depending on activity and experience.

Südhof emphasizes that synaptic transmission is not fixed:

“No individual synapse has fixed relationship between input and output.”

Because synapses can strengthen or weaken over time, neural circuits can adapt, learn, and encode information.

This plasticity is one of the fundamental mechanisms underlying memory and learning.

The Bigger Picture

Synaptic transmission happens billions of times per second in the human brain. Each event depends on molecular interactions occurring at nanometer scales and microsecond timescales.

By identifying the proteins that control neurotransmitter release, Südhof helped reveal the machinery that makes neural communication possible.

What looks like a simple connection between neurons turns out to be one of the most sophisticated molecular systems in biology; one that allows the brain to translate electricity into chemistry and transform signals into thought.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Tom Südhof on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Gelbart’s career spans lasers, semiconductor manufacturing tools, ultra-precision optics, and high-resolution imaging systems. Across the conversation, one theme emerges clearly: most limits aren’t violations of physics, they’re misunderstandings of it.

The 632nm podcast features in-depth conversations with the world’s leading scientists and engineers.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. The First Laser Worked Because Someone Re-Measured

One of the central stories in the episode is the first working laser, built by Theodore Maiman. At the time, conventional wisdom said ruby didn’t have enough gain to lase.

Gelbart explains what changed:

“Everybody said ruby won’t lase because the gain is too low. Maiman re-measured the gain and found out they were wrong.”

Rather than accepting published values, Maiman checked the physics directly. The result was the first operational laser.

The lesson, Gelbart suggests, is not that physics was broken, but that the inputs to the engineering problem were wrong.

II. Impossible vs. Misunderstood

Throughout the conversation, Gelbart returns to a core distinction: a true physical prohibition versus a constraint we don’t fully understand.

“Very few things are impossible. Most things are just very, very difficult.”

Engineering, in his framing, is about identifying where the real boundary lies. Is a design blocked by thermodynamics? Or by vibration? Or by measurement error? Or by cost?

When engineers misdiagnose the constraint, they stop too early.

III. Exploiting the Fine Print of Physics

Gelbart is careful not to claim that engineers “break” physical law. Instead, they operate in regimes others ignore.

Discussing edge cases and loopholes, he notes:

“It’s like being a good lawyer in a trial, you can always find a loophole.”

Whether it’s pulsed operation instead of continuous, compliant mechanisms instead of rigid ones, or nonlinear material effects instead of linear approximations, progress often comes from asking: What assumptions are we making?

IV. Precision Is a System Property

Working at nanometer tolerances in optical and semiconductor systems, Gelbart emphasizes that precision isn’t about one perfect part: it’s about system-level design.

Thermal expansion, vibration isolation, metrology feedback loops, and mechanical compliance all interact. What looks impossible in a static diagram can become feasible once dynamics, averaging, or feedback are introduced.

This mindset, treating physics as a landscape to navigate rather than a wall to hit, recurs across his stories, from imaging systems to manufacturing tools.

V. Fundamentals Over Fashion

Gelbart is blunt about education and technical culture. Tools change. Software changes. Platforms change.

Physics does not.

The engineers who consistently solve hard problems, he argues, are those who deeply understand mechanics, optics, and electromagnetism, and are willing to re-measure, re-calculate, and question consensus when something doesn’t add up.

The Bigger Picture

Across lasers, manufacturing systems, and ultra-precision machines, Gelbart’s worldview is consistent: the limits of physics are real but they are often not where we think they are.

Engineering breakthroughs don’t come from ignoring natural law. They come from understanding it well enough to find the narrow path through it.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Dan Gelbart on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Rather than treating vision as an end in itself, Bear uses the visual system as a controlled experimental window into deeper questions about learning, memory, and brain development.

The 632nm podcast features in-depth conversations with the world’s leading scientists.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. Vision as a Tool, Not a Topic

Bear is often described as a vision scientist, but he draws an important distinction between the system he studies and the question that motivates him. Vision, in his view, is uniquely useful because it offers precise inputs, clear behavioral readouts, and well-mapped circuitry.

“Sometimes colleagues accuse me of studying vision and I always correct them. And I say, no, no, no. I use vision to study plasticity in the brain.” — Mark Bear, ~00:31:00

Because visual experience can be experimentally manipulated—by altering input to one or both eyes—it becomes possible to observe, in real time, how experience reshapes synaptic connections.

II. Why Experience Is Required

One of the earliest surprises to emerge from studies of the visual cortex was that genetic wiring alone could not account for the brain’s precision. Inputs from the two eyes must be aligned onto single cortical neurons, a process that cannot be fully specified in advance.

“That degree of precision seemed to be greater than what could be achieved with genetic instructions alone.” — Mark Bear, ~00:02:00

Instead, the brain relies on postnatal experience to fine-tune its circuitry—strengthening some connections while weakening others until the two eyes are functionally stitched together.

III. Critical Periods and Timing

Classic experiments in visual deprivation revealed a stark constraint on plasticity: when experience is absent early in life, normal wiring fails to develop—and later correction becomes impossible.

“Both losing the connection and restoring the connection is possible only during a finite period of early postnatal life, and that is what we call a critical period.” — Mark Bear, ~00:05:00

In humans, Bear notes, the most extreme visual critical periods close remarkably early, underscoring how tightly brain development is coupled to timing.

IV. What Closes the Window

If experience opens plasticity, what shuts it down? Bear points to the maturation of inhibitory circuitry as a key mechanism.

“The prevailing theory is that inhibition in the brain is late to mature… when it achieves a certain level of maturation that ends the critical period.” — Mark Bear, ~00:09:00

Early development is marked by an overproduction of synapses, followed by large-scale pruning as circuits stabilize. Plasticity fades not because the brain stops changing, but because its networks become increasingly constrained.

V. Plasticity in the Adult Brain

For decades, primary visual cortex was assumed to be fixed after childhood. Bear explains that this view has softened considerably.

“It’s become more and more apparent that there’s still a lot of experience-dependent plasticity in the primary visual cortex.” — Mark Bear, ~00:54:00

Perceptual learning and training can still refine sensory representations, especially when reinforced by sleep and repetition. Adult plasticity is slower and more limited, but it remains biologically real.

The Bigger Picture

Across the episode, Bear frames vision as a proving ground for a broader principle: experience is not merely layered onto the brain after development—it is an essential force in building it. From early critical periods to adult learning, neural circuits remain shaped by interaction with the world, constrained by biology but never entirely frozen.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Mark Bear on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Capasso reflects on his transition from Bell Labs to academia, the origins of flat optics, and why clarity and simplicity are not aesthetic choices, but scientific ones.

The 632nm podcast features in-depth conversations with the world’s leading scientists.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. A Question That Removed the Lens

Capasso traces the birth of metasurfaces to a blunt engineering challenge: how to deploy optical sensing in places where traditional components simply don’t fit.

“I need a quantum cascade laser in a drone without a lens.” — Federico Capasso, ~00:03:00

Rather than refining existing optics, the question reframed the problem entirely. What if the function of a lens could be encoded directly onto a surface?

II. When Subwavelength Changes the Rules

Metasurfaces operate in a regime where structural features are smaller than the wavelength of light, eliminating familiar diffraction behavior.

“As soon as the separation becomes less than the wavelength, the only order that survives is the direct light going through.” — Federico Capasso, ~00:00:20

This shift allows phase, amplitude, and polarization to be controlled locally—turning optical design into a problem of geometry rather than bulk material.

III. Why Dielectrics Won Early

Metalenses often relied on plasmonics, but Capasso made a deliberate pivot toward dielectric designs.

“The foundries are not going to make plasmonic lenses. That’s a fact.” — Federico Capasso, ~01:12:20

The decision was driven by scale. Dielectric metasurfaces fit naturally into existing semiconductor fabrication pipelines, making flat optics not just feasible, but manufacturable.

IV. Flat Optics and Simpler Aberration Control

Traditional refractive lenses fight aberrations with added elements and complexity. Metasurfaces reverse that logic.

“The control of aberrations with metal optics is much easier than with regular optics.” — Federico Capasso, ~00:53:00

By prescribing phase directly, correction becomes part of the design rather than an afterthought.

V. Simplicity as a Discipline

Capasso repeatedly returns to a principle inherited from Fermi and the Roman school of physics: avoid complication unless it is unavoidable.

“If there is a simple way, this is more valid.” — Federico Capasso, ~00:51:20

For him, simplicity is not about elegance for its own sake—it is how physics stays connected to reality, fabrication, and impact.

The Bigger Picture

Across the episode, Capasso argues that metasurfaces matter because they align physics with how the modern world actually builds things. Flat optics collapse complexity, reduce cost, and make optical functionality scalable in the same way electronics became scalable decades ago.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Federico Capasso on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Kim reflects on the early days of graphene, the role of Berry phase and topology, the rise of moiré materials, and why experimental physics still depends on listening carefully when data refuse to behave as expected.

The 632nm podcast features in-depth conversations with the world’s leading scientists.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. Why Graphene Carries a π Berry Phase

At the heart of graphene’s unusual transport properties is its Dirac-like electronic structure. Kim explains that the Berry phase in graphene is not an abstract theoretical construct—it produces direct, measurable consequences in experiments such as quantum oscillations.

“In graphene, when electrons make a cyclotron orbit, the Berry phase is π.”
— Philip Kim, ~01:37:00

That π phase shifts the Landau level sequence relative to ordinary metals, providing one of the earliest experimental fingerprints that graphene’s charge carriers behave like relativistic particles rather than conventional electrons.

II. From Band Structure to Topology

Kim situates graphene within a broader intellectual arc that connects condensed matter physics to topology. Long before topological insulators became a major field, graphene provided an accessible platform for thinking about Berry curvature, phase accumulation, and protected electronic structure.

“The quantum spin Hall effect was first discussed using graphene as a starting point.”
— Philip Kim, ~01:39:00

Experiments on graphene did not just confirm existing theory; they helped normalize topological thinking in materials physics, making abstract concepts experimentally tangible.

III. When “Bad Devices” Turn Out to Be New Physics

One of the defining stories of the episode concerns the discovery of moiré physics in graphene heterostructures. Early devices showed unexpected features that looked, at first, like fabrication problems.

“We almost threw those devices away.”
— Philip Kim, ~02:07:00

Instead of discarding the data, Kim’s group kept probing. The anomalies turned out not to be disorder, but signatures of a long-predicted fractal energy spectrum emerging from the interplay of magnetic field and lattice periodicity.

IV. Letting the Data Lead

As the measurements accumulated, the interpretation became unavoidable. What initially looked like noise revealed itself as structure.

“The data started to teach us what was going on.”
— Philip Kim, ~02:08:00

That realization marked a turning point, connecting graphene experiments to the Hofstadter butterfly—a theoretical result that had existed for decades without a clean experimental realization.

V. Being Wrong About Twisted Graphene

Kim is candid about his initial skepticism toward magic-angle twisted bilayer graphene. The required angular precision seemed implausible, and the theoretical assumptions too idealized.

“I was completely wrong.”
— Philip Kim, ~02:12:00

The subsequent observation of correlated insulating states and superconductivity reshaped the field, turning twist angle into a new control parameter for engineering flat bands and strong correlations.

The Bigger Picture

Across the conversation, Kim returns to a consistent theme: progress in experimental physics often comes not from executing a perfect plan, but from recognizing when nature is telling you something unexpected. From Berry phase shifts to moiré superlattices, the story of graphene is one of theory, fabrication, and surprise converging in real devices.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Philip Kim on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Sachdev’s work spans more than three decades, from early models of strongly interacting electrons to modern frameworks that connect strange metals, quantum criticality, and holographic duality. Along the way, ideas developed to understand electrons in solids have become central to how physicists now think about horizons, chaos, and quantum gravity itself.

The 632nm podcast features in-depth conversations with the greatest scientists in the world.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. When Temperature Becomes a Horizon

One of the most striking ideas in modern theoretical physics is that temperature in a quantum system can map directly onto geometry. Sachdev explains how this emerges from AdS/CFT, where excited quantum states correspond to black holes in a higher-dimensional spacetime.

“If you look at an excited state at some temperature, then in the string theory you get a horizon, and the Hawking temperature of this horizon is precisely the temperature of the CFT.”
— 01:34:00

Thermal behavior in quantum matter is no longer just statistical—it is gravitational. Properties like entropy, transport, and chaos can be read off from the physics of horizons, linking condensed matter directly to black hole thermodynamics.

II. A Model From 1992 Comes Back

In the early 1990s, Sachdev introduced models of strongly interacting fermions that were mathematically simple but physically puzzling. At the time, they seemed like abstract exercises. Years later, they resurfaced at the center of holography.

“What I had done in 1992 suddenly started becoming relevant to what I was working on in 2007.”
— 01:35:00

The SYK model turned out to be a rare solvable example of a maximally chaotic quantum system—and, unexpectedly, a precise dual of gravity in two-dimensional anti–de Sitter space. What began as condensed matter theory became a testing ground for quantum gravity itself.

III. Why Quantum Computers Aren’t Natural Materials Simulators

Quantum computers are often described as ideal tools for simulating quantum matter. Sachdev urges caution. The basic building blocks of today’s quantum computers don’t naturally capture the physics of electrons.

“Quantum computers are dealing with qubits, which are bosons really. They don’t have fermions very naturally—you have to build that in.”
— 01:38:00

Encoding fermionic statistics introduces significant overhead, especially on near-term devices. As a result, analog systems like optical lattices—where fermions exist natively—remain more faithful platforms for studying strongly correlated materials.

IV. Strange Metals and the Fan of Quantum Criticality

At the heart of Sachdev’s work is the strange metal—a phase that defies conventional quasiparticle descriptions and appears in high-temperature superconductors.

“The strange metal is a quantum critical phase.”
— 01:47:00

Unlike ordinary metals, strange metals sit inside a “fan” of quantum criticality that extends to finite temperature. Even away from the zero-temperature critical point, quantum fluctuations dominate transport and thermodynamics.

“You don’t have the resolution to decide which side you’re on, so you behave as if you’re critical.”
— 01:51:00

Temperature itself blurs the distinction between phases, allowing quantum-critical behavior to persist far beyond absolute zero.

V. What SYK Ultimately Taught Us

Sachdev is clear-eyed about the limits of idealized models. The SYK model is not a literal description of real materials—but its value lies elsewhere.

“The SYK model has taught us a lot, much more than we expected. It taught us a lot about strongly entangled, chaotic dynamics.”
— 02:13:00

Those insights now guide the construction of more realistic theories that combine quantum criticality, fermions, and collective modes—offering a possible route toward understanding high-temperature superconductivity itself.

The Bigger Picture

Sachdev’s central message is not that black holes explain superconductors, or that gravity replaces condensed matter theory. It is that deeply entangled quantum systems obey universal principles, whether they appear in a laboratory crystal or at the edge of a horizon.

Progress, he emphasizes, comes not from unlimited resources or brute-force computation, but from the slow convergence of ideas across fields—condensed matter, quantum information, and gravity—each illuminating the others.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Subir Sachdev on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

Through the noise of hype cycles and funding booms, Fowler has remained one of the field’s clearest thinkers. In this conversation, he walks us through why quantum computing is still worth pursuing—and why it will take much longer than most people expect.

The 632nm podcast features in-depth conversations with the greatest scientists in the world.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. The Five-Year Fallacy

A recurring theme in quantum computing is the hype cycle—especially the idea that the field is always five years away from success. Fowler addresses that directly.

“Every quantum platform goes through the ‘we’ll nail it in five years’ phase. And you won’t, because the number of problems that you have to solve is huge.”
— 00:03:30

He argues that unrealistic expectations have slowed the field, not because progress is impossible, but because underestimating the engineering challenge leads labs (and funders) to chase short-term milestones instead of long-term architectures.

II. Why Surface Codes Still Win

The surface code remains the dominant approach to fault tolerance, not because it’s elegant, but because it tolerates error better than any alternative.

“The surface code has a much higher threshold… your gates don’t have to work that well to get the error correction working well.”
— 00:09:10

The appeal is simple: surface codes turn noisy, imperfect hardware into a system that can suppress logical error exponentially. Fowler emphasizes that this robustness, not qubit count, is the real bottleneck to useful quantum computation.

III. When Radioactive Particles Break Your Qubits

Even when qubits behave theoretically, real materials don’t. Fowler describes a very physical—and unintuitive—source of catastrophic failure.

“Radioactive particles in the material… can hit your chip and cause errors. Sometimes… it can heat up the whole chip and cause everything to have an unacceptably high error rate.”
— 00:11:30

These hardware-level disturbances create correlated errors that software error correction can’t fix, which is why quantum computing still requires intense materials engineering—not just clever algorithms.

IV. From Circuits to 3D Structures

Fowler’s most forward-looking work centers around “lattice surgery,” which reframes computation as geometry. Instead of gates on a timeline, the quantum program becomes a 3D object.

“You turn a quantum computation into a complex three-dimensional branching structure… that does all of the work in one go.”
— 00:19:20

This approach allows compilers to optimize quantum computations globally, leading to dramatic reductions in overhead compared to gate-by-gate scheduling.

V. Why Google Walked Away

One of the most candid parts of the interview is Fowler’s explanation of why he left Google—and why the company abandoned fully open-source quantum software.

“Ultimately they felt it was a business risk to run a collaboration with everyone.”
— 01:34:30

Export-control concerns meant Google could not support the open, international collaboration Fowler believed was necessary for rapid progress. His commitment to open tooling eventually led him to continue the work independently.

The Path Forward

Fowler is neither an optimist nor a pessimist: he’s a realist focused on engineering. His central message is that quantum computing won’t succeed through secrecy or competition—it will succeed through shared tools, verified models, and collaborative compilers that map quantum algorithms onto real hardware.

Whether the field is five years away or fifty, what matters is building the infrastructure that makes progress inevitable.

🎧 Full episode available now on YouTube and all podcast platforms.

Listen to our full conversation with Austin Fowler on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

The 632nm podcast features in-depth conversations with the greatest scientists in the world.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. The Logic of Elite Overproduction

At the core of Turchin’s theory is a simple idea: when too many people compete for too few elite positions, societies destabilize.

“When you have 10 times as many elite aspirants for positions, 90% of them are frustrated... and then many of those frustrated, they turn into what you call counter-elites.”

He compares the dynamic to a game of musical chairs—one that becomes violent when the number of players far exceeds the number of seats. The result isn’t just competition, but the formation of factions that challenge the legitimacy of the system itself.

🎧 Listen at 00:09:00

II. The Return of the Wealth Pump

Turchin argues that this instability was set in motion decades ago by economic policy changes that redistributed gains upward.

“They stopped increasing minimum wage... shipped lots of industries overseas, and there was massive immigration also. So all those factors... resulted in what I call the wealth pump.”

The “wealth pump,” he explains, transfers income from workers to elites, creating both a growing pool of frustrated professionals and a narrowing band of opportunity at the top.

🎧 Listen at 00:15:00

III. Measuring Collapse with Data

Unlike traditional historians, Turchin builds quantitative models to test how these forces interact over time. His team encodes historical data—population, wages, unrest—into computational systems that simulate the pressure points of complex societies.

“The idea is... not to predict the future... but to actually predict the effects of different types of interventions.”

These models, he says, work less like crystal balls and more like climate forecasts: they show how certain policies might amplify or defuse underlying instability.

🎧 Listen at 00:46:00

IV. Technology and the Science of History

New technologies are transforming this work. AI systems can now extract and classify data from thousands of historical texts—what once took researchers weeks now takes hours.

“We now use AI to extract data from historical texts—what used to take a human weeks, a large language model can do in hours.”

Turchin sees this as part of a broader revolution in the humanities: a shift toward computational social science capable of testing hypotheses across centuries and civilizations.

🎧 Listen at 00:41:00

V. A Cautious Optimism

Despite his reputation as a prophet of collapse, Turchin insists that these patterns aren’t destiny—they’re feedback loops that can be understood and changed.

“We should... have a public-domain computational model that people can actually see how it runs... so they can get convinced that if you do this, then there is going to be an unintended consequence that will actually make things worse.”

By making these models transparent, he hopes policymakers and citizens alike can learn how to stabilize societies before crises escalate.

🎧 Listen at 00:56:00

Full episode out now on YouTube.

Listen to our full conversation with Peter Turchin on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share

The 632nm podcast features in-depth conversations with the greatest scientists in the world.

If you haven’t already, subscribe to 632nm wherever you get your podcasts.

Subscribe now

I. The Paradox of Quantum Error Correction

Lukin begins with the contradiction at the heart of quantum computing — you can’t copy or measure quantum information, and yet it can somehow be protected:

“You cannot copy quantum information. Number two, in quantum mechanics, if you measure the state, you destroy it. Nevertheless, it turns out that you can use this redundancy to actually encode and protect quantum information. But actually it requires you to prepare these entangled states, these big superpositions that sometimes are called Schrödinger cat states.”

What makes this possible, he explains, is redundancy built not from copies, but from entanglement — a subtle kind of correlation that spreads information across many qubits without ever making it directly observable.

🎧 Listen at 00:02:00

II. All Evolution in Quantum Mechanics Is Analog

Every quantum operation is continuous — a smooth rotation in Hilbert space. That’s what makes controlling qubits so demanding:

“All evolution in quantum mechanics is analog evolution. There is no such thing as a digital operation that just snaps something from zero to one. Every evolution is continuous, so any small imperfection will make your state drift away continuously.”

In classical computing, tiny voltage errors are erased by digital thresholds. In quantum computing, even a microscopic phase error accumulates — a challenge that makes error correction not just useful, but essential.

🎧 Listen at 00:09:00

III. Ancilla Qubits: Measuring Without Collapsing

To detect errors without destroying information, Lukin explains, physicists use special “helper” qubits that indirectly sense when something has gone wrong:

“You can measure some auxiliary qubits—ancilla qubits—which don’t contain information themselves but are entangled with the data. They tell you something about whether an error has occurred, but not what the encoded quantum state actually is.”

This approach allows researchers to extract classical error signals from a quantum system, preserving the delicate superpositions that store computational information.

🎧 Listen at 00:15:00

IV. Entangling Logical Qubits Through Rydberg Interactions

Lukin’s group builds these ideas into hardware using neutral atoms excited into Rydberg states, whose strong interactions allow for controlled entanglement across entire arrays:

“You take one group of atoms and bring it so that the atoms sit nearly on top of each other. Then you shine one pulse of light which entangles them all in parallel. That’s a transversal operation between two logical qubits.”

These transversal gates perform the same operation across many atom pairs at once — a key ingredient for fault-tolerant quantum computation.

🎧 Listen at 00:28:00

V. Testing the Limits of Quantum Mechanics

Ultimately, Lukin views each advance in quantum error correction not just as an engineering step, but as a test of physics itself:

“When we make these larger and larger superpositions and keep them stable, we are really testing whether the laws of quantum mechanics still hold for systems of this size. Every experiment is a test of the theory itself.”

The pursuit of stable superpositions isn’t only about computation — it’s about seeing how far quantum mechanics can stretch before giving way to new physics.

🎧 Listen at 00:41:00

Full episode out now on YouTube.

Listen to our full conversation with Mikhail Lukin on your favorite podcast platform:

• Apple Podcasts

• Spotify

• Or search for “632nm” on your preferred podcast player

If you know someone who would enjoy this, please share it with them!

Share