Unlock Editor’s Digest Lock for Free
FT editor Roula Khalaf will select your favorite stories in this weekly newsletter.
The rapid advances in artificial intelligence that have revised the technology industry have put another potentially transformative idea: quantum computing in the shade. When the straight lash of AI dominates the headlines, it is difficult to focus on the farther and unproven benefits of quantum machines.
However, this is not just a matter of perception. According to two of the key figures in AI, quantum computing may be far less important and quite less important than many people working in this field claim.
Their comments forced the quantum computing industry into defensive crouches, reviving questions that are difficult to shake up. Where is the boundary between hype and reality?
This year, Nvidia CEO Jensen Huang predicted that useful quantum computers are still 20 years away. Huang’s own company works closely with many quantum companies, including adapting CUDA software to help researchers create quantum simulations. That didn’t prevent his comments from hitting stocks in the company they’ve been publicly recognised.
The reaction that triggered is less dramatic, but it has potentially even greater significance, and the proposal from Demis Hassabis, CEO and co-founder of GoogleepMind, believes that AI can only be tackled by quantum computers. It was a suggestion that many tasks were available to take on. .
One of the greatest hopes of quantum machines is that they can model molecular activity in much more detail than traditional or “classical” computers. It could pave the way for new medicines and battery technology. However, according to Hassabis, AI running on computers today has already proven to be proficient in modeling complex systems and can handle this type of work.
Naturally, such comments led to a quick response from the quantum crowd. Hartmut Neven, head of the quantum effort at Google, said this week that he is confident that “real world applications that can only be achieved with quantum computers” will arrive within five years. It is not clear what shape these will take.
Google’s big bet lies in full-scale fault-resistant quantum machines that could far outweigh classic computers. Late last year, it demonstrated that it was able to overcome the “noise” that accumulates in quantum systems as a result of the inherent instability of basic components known as qubits. A practical system.
But in a near term, industry hopes depend on what is known as NISQ (voracious, mid-scale quantum). These can only handle short quantum computations before being overwhelmed by noise, but they can be used to generate useful ones. Neven’s argument coincided with the publication of Google Research in Nature, which outlines new techniques that could make NISQ Quantum simulations more practical.
However, supporters of the NISQ system have long argued that it is close to a breakthrough. Questions remain until we can demonstrate useful computing tasks that are not handled by classic machines.
However, the straight forward movement of AI may have paved a new pathway for quantum computing. Quantinuum – formed from the merger of Honeywell’s Quantum Arm and the UK’s Cambridge Quantum – this week, quantum machines were used to generate additional data to create a large-scale language model that supports the majority of AI today. We have announced how to train.
According to CEO Raj Hazra, simulating nature at the molecular level within a quantum computer generates data that cannot be generated otherwise. That may be worth it, he adds. Adds companies seeking to train AI models for drug discovery and research into new materials. However, this has not been shown to bring about meaningful advancements over classical computing.
Quantinuum’s work refers to a broader point regarding the interaction between quantum computing and AI. It shows that both the boundaries between the two fields are shifting as they evolve. The two technologies are likely to work with each other, each working on the optimal computing task.
With AI moving forward at a fierce pace and quantum computing becoming even more promising than reality, it remains difficult to predict how two marriages will work.
richard.waters@ft.com
