The concept of artificial intelligence (AI), one of the most significant and intriguing gifts of technology, was conceived in the late 1940s, though it only began to materialize by the late 1990s. AI, which emerged almost 20 years ago, is a complex mathematical science that enables inanimate devices to learn, producing astonishing results. This foundational ability is why we still define AI fundamentally as machine learning.
However, such significant advancements and progress in AI would not have been possible without the acceleration of electronics, computer, software, and material technologies starting in the 1970s. Digital technologies provided infrastructure that evolved even faster than AI itself, revolutionizing many human concepts over the last 50 years, including cell phones, space shuttles, and automation.
Like natural intelligence, AI is fundamentally based on data and learning (teaching). While human intelligence processes data through models we have yet to fully understand, AI “learns” and “comes to life” through mathematical and digital models we develop.
In recent years, concerns about processing vast amounts of data more quickly and meaningfully have led to the development of neuromorphic computing circuits and quantum processors, the foundations of which were laid about 20 years ago. Quantum processors, initially theoretical and seemingly impractical, performed their first calculations about ten years ago. Despite being very basic, these computations happened almost instantaneously, with qubits operating in superposition states.
In 2019, IBM’s quantum computer solved a moderately complex predictive algorithm in under 2 seconds—an operation that took about 40 seconds on my relatively powerful personal computer. This near-zero computation time was staggering, though constraints remained daunting. The limitations were not just financial but technical, making quantum computing accessible for only a narrow range of applications.
Today, owning a personal quantum computer is not feasible. They are not designed for everyday tasks like listening to music, reading news, writing documents, or preparing presentations. These are efficiently handled by conventional computers.
Quantum computers are designed to run analysis algorithms on massive datasets—beyond the capabilities of traditional 1-0 mechanisms and sequential models. They leverage models from the quantum (subatomic) world, allowing for incredibly fast processing. For example, tasks that would take a supercomputer years to complete could be done in minutes on a quantum computer.
Why do we need greater capacity? For instance, the Intel Core i9-13900K processor, released in 2023, has a clock speed of 5.5 GHz and 24 cores, capable of performing approximately 250 billion operations per second. Despite this, some computations can take hours, days, months, or even years due to the need to process terabytes of data, distribute it across the entire dataset, and perform operations similar to a human brain. The human brain, estimated to have around 100 billion neurons, can perform roughly 20 quadrillion operations per second—still far beyond our current AI capabilities.
I had the opportunity to test quantum computers on three incredible projects: predictive analysis on healthcare data, optimizing a vast city’s smart data, and optimizing operations using four years of detailed data from a major airport. These experiences underscored why AI could be both fascinating and potentially dangerous for humanity in the coming years—a mix of anxiety and admiration.
I continue to develop digital solutions that could change the future and create a better world for future generations, particularly focusing on neuromorphic processing and computing. I am deeply grateful to my dear friends who support me in these endeavors.
Working with quantum computers is currently a challenging process. Nonetheless, simulators are available for new learners to experiment with quantum coding. For more significant projects, transitioning to a different segment and acquiring substantial permissions and authorizations is necessary. While IBM Q and Qiskit have provided successful results, you might have heard of China’s Zuchongzi and Jiuzhang 2.0 quantum computers, known for their super-fast problem-solving capabilities. In a small test on Jiuzhang, our predictive algorithm completed so quickly it was immeasurable, achieving an incredible error expectation of e-16 over 1,000 epochs of an advanced AI network model on entropy health data.
Moreover, we used neuromorphic models in quantum computing, performing data processing similarly to the human brain. One of the models we are still working on is our long-term Digital Mycelium Networks (DMN) project. We can now see that computations taking weeks on supercomputers can be completed in mere seconds. Exciting and terrifying!
We must be prepared for a future with machines that can manage us (humanity), equipped with robots, distributed intelligences, unlimited sensors, and the ability to learn, teach, and perform intuitive operations.
Next article, you’ll find how QC can be used in aviation industry.