1. Introduction: The Physics of the "Impossible"
In May 1981, at the MIT Endicott House, Richard Feynman delivered a lecture that would become the foundational text of a new epoch: Simulating Physics with Computers. As a historian of science, I find it fascinating that Feynman’s realization wasn't merely a creative hunch; it was a logical deduction. He observed that nature is "quantum mechanical in its essence," and therefore, probabilistic classical simulations are fundamentally doomed to fail at scale. To simulate the "impossible" complexity of the subatomic world, Feynman argued, we require a machine built of quantum elements—a universal quantum simulator.
What began as a chalkboard concept—a theoretical "lattice of spins"—has transformed over four decades into a sprawling cloud-based industry. Yet, the journey from that 1981 lecture to today’s multi-qubit processors has been anything but linear. It is a history defined by market-shaking leaks, scientific rivalries, and the realization that in the quantum realm, the line between hardware maintenance and software code is almost non-existent.
2. The Day Quantum Computing Crashed the Bitcoin Market
The intersection of abstract physics and global finance reached a fever pitch in September 2019. This was not due to a polished press release, but rather a leak. On September 22, a paper detailing Google’s "Sycamore" experiment was briefly posted to a NASA server before being pulled. In the forty-eight hours that followed, the value of Bitcoin and other digital currencies plummeted by more than 10%.
The market volatility of 2019 serves as a warning: the perceived threat to the cryptographic foundations of the blockchain is a real-world economic force. Investors reacted to the possibility that the "strong Church-Turing thesis"—the idea that any physical process can be efficiently simulated by a classical computer—had finally been upended. As the Kalai et al. analysis notes, the gravity of the moment was not lost on the public:
"Google’s announcement of quantum supremacy was compared by various writers... to landmark technological achievements such as the Wright brothers’ invention of a motor-operated airplane, the launching of Sputnik, and the landing of humans on the moon."
3. The "10,000-Year" Supercomputer Gap That Vanished in Seconds
Google’s 2019 claim of "quantum supremacy" was centered on a sampling task that Sycamore completed in 200 seconds—a feat they estimated would take a state-of-the-art classical supercomputer 10,000 years. However, in the annals of science, "supremacy" is often a fleeting title.
By 2021, the 10,000-year gap had been decimated. A team from the University of Science and Technology of China (USTC), led by Pan, Chen, and Zhang, demonstrated that by using improved classical algorithms—specifically tensor network methods—the same sampling task could be completed on a classical machine in a matter of seconds. While Google had moved to its "ABCDCDAB" architecture to make simulation harder, the USTC team’s Zuchongzhi and Jiuzhang experiments showed that classical computers are a moving target. Far from a failure, this "moving goalpost" represents a healthy scientific competition that forces both classical and quantum researchers to innovate at a breakneck pace.
4. Calibration Isn't Maintenance—It's Actually Part of the Code
One of the most surprising technical realities revealed by the analysis of the Sycamore chip is that "calibration" is not a background task; it is a fundamental layer of the computation. In the NISQ era, hardware is so finicky that a programmer cannot simply run "standard" gates—the idealized mathematical operations of a circuit—and expect a result.
Instead, the mathematical definition of the circuit must be "distorted" to match the specific systematic deviations of the physical qubits. This involves adding fixed rotations to the qubits before and after gates are executed to create "native gates." The stakes for this mathematical alignment are incredibly high:
"Using Sycamore’s standard gates instead of native gates halves the fidelity... moving to the original circuit with standard gates slashes the fidelity to zero."
In short, the hardware is so tied to its own quirks that ignoring one-gate adjustments doesn't just produce noisy data—it produces no data at all.
5. The First Universal Logic Gate Lived Inside a Single Atom
While today we chase hundreds of qubits, the origin of physical quantum logic is found in the archives of NIST, circa 1995. There, the team of Monroe, Meekhof, King, Itano, and Wineland demonstrated the first "controlled-NOT" (CNOT) gate.
The radical brilliance of this experiment lay in its economy: it used a single trapped ion. Rather than using two separate atoms, the researchers utilized two different "degrees of freedom" within the same atom. One qubit was stored in the atom's internal energy states, while the second qubit was the atom’s external motion—its vibration—within the trap. To achieve this, the atom had to be laser-cooled to its absolute limit:
"The two quantum bits are stored in the internal and external degrees of freedom of a single trapped atom, which is first laser cooled to the zero-point energy."
This 1995 milestone proved that the "vibration" of an atom could serve as a computational bit, a departure from classical logic that set the stage for every ion-trap processor in operation today.
6. Negative Probabilities: The "Imagined" Reality of Quantum Math
In his 1982 paper, Feynman revisited a concept he had wrestled with during his Nobel-winning work on electrodynamics: negative probabilities. While a final, observable event must always have a positive probability, Feynman argued that "imagined intermediary events" can be negative.
This is the key to why quantum computers can outperform classical ones. Classical "Monte Carlo" simulations, which rely on random sampling, are frequently crippled by the "sign problem." When configurations have negative weights, the statistical error grows exponentially, making the simulation "NP-hard" and practically impossible. Quantum computers don't just calculate these negatives; they bypass the sign problem by sampling directly from the quantum system. They effectively navigate a mathematical reality where negatives cancel each other out, a shortcut that classical math cannot take without an exponential explosion in time.
7. The Cloud Revolution: When Lab Experiments Became Systems
Prior to 2016, quantum computing was a "nights and weekends" endeavor. If you were a theorist, you had to beg an experimentalist for time on a bespoke, hand-calibrated device. The revolution occurred on May 4, 2016, when IBM launched the IBM Quantum Experience.
The catalyst was a moment of human consensus. In December 2015, at the ThinkQ conference, MIT professor Isaac Chuang asked a room of researchers if they would actually use a small-scale, cloud-based processor. Every hand in the room went up.
This shift required a fundamental change in mindset: treating the dilution refrigerator not as a "lab experiment" but as a "system." It meant automating twice-daily calibrations to ensure the hardware remained operational for a global audience. Today, that fleet of cloud-based systems runs two billion circuits daily, moving the field from the era of "bespoke physics" to the era of "computational utility."
8. Conclusion: Beyond the NISQ Era
We remain firmly in the "Noisy Intermediate-Scale Quantum" (NISQ) era. While we have reached the milestone of thousands of qubits, we still lack "fault-tolerance"—the ability to correct the decoherence that Feynman warned about in 1982.
There is a lingering tension in the field. While some suggest we are on a path of "doubly exponential" growth—a quantum equivalent of Moore’s Law—many historians and skeptics note there is no concrete evidence for this trajectory yet. As we look forward, the central question remains: Are we currently living in a bubble of "supremacy" claims, or are we witnessing the slow, difficult dawn of true quantum utility? Whether the 10,000-year gap stays closed or reopens will depend on our ability to finally master the noise that Feynman first identified forty years ago.
No hay comentarios:
Publicar un comentario