Quantum computation is overshadowed by fast progress in AI
Unlock free Digest editor
Roula Khalaf, editor of FT, chooses her favorite story in this weekly newsletter.
The rapid progress of the artificial intelligence that transferred the technological industry put another potentially transformative idea – quantum calculation – in the shade. It is difficult to focus on the distant and unproven benefits of quantum machines when the head of AI dominates the titles.
This, however, is more than just a question of perception. According to two leading figures in AI, quantum calculation could be much further, and of significantly smaller importance than many people working on the field like to claim.
Their comments forced the quantum computer industry on the defensive circuit and revived a question that is difficult to shake: where is the line between the hype and reality for the supposedly technology that changes the world that has yet to produce anything of practical value?
This year, the NVIDIA CEO, Jensen Huang, predicted that the useful quantum computers were still 20 years – much longer than the companies working on the field. Huanov’s own company is closely cooperating with many quantum companies, including the adaptation of miracle software to help researchers create quantum simulations. This did not prevent his comments that hit the stock of public trading companies.
Less dramatically in the reaction he caused, but potentially greater importance, was the proposal of Demis Hassabis, executive director and co -founder of Googleleepmind, that AI could take over many tasks that were thought to be just a quantum computer to fight.
One of the biggest hopes of quantum machines is that they will be able to model molecular activities much more detailed than traditional or “classic” computers ever. This could pass the way with new medicines or battery technologies. According to Hassabis, however, the AI that works on today’s computers is already showing skill in modeling complex systems and could handle this type of work.
It’s not surprising that such comments brought a quick response from the quantum crowd. Hartmut Neven, head of Google’s quantum effort, said this week that he is convinced that “the applications in the real world that are only possible on quantum computers” will arrive within five years. Exactly which form will take on is not clear.
Google’s big bet was at full measuring quantum machines that can far exceed classic computers. At the end of last year, he showed that he was able to overcome the “noise”, which is being built in quantum systems as a result of the innate instability of their basic components, known as Qubits – an important step while trying to increase and produce a practical system.
It is closer, however, the hope in the industry rests on what is known as Nisq-Buchic, Medium Quantum machines. They can handle short quantum budgets before they flood them noise, but they can still be used to produce something useful. Neven’s claim coincided with the publication of Google Research in nature in which she emphasized a new technique that could not make Nisq Quantum simulations more practical.
However, the supporters of the NISQ system have been claiming to be close to the breakthrough for years. As long as they are unable to show useful computer tasks that could never be resolved on a classic machine, doubts will remain.
But the progress of AI may have opened new ways for quantum calculation. Quantnum – originating from the merger of Honeywell’s quantum hands with Cambridge Quantum in the UK – this week discovered the way of using its quantum machines to generate additional data to train large language models that support most of today’s AI.
According to the director of Raj Hazra, the simulation of nature at the molecular level within the quantum computer produces data that cannot be generated in any other way. That could be useful, he adds, for companies that want to train AI models to detect drugs or explore new materials. But it is not shown to result in a meaningful progress over classic calculation.
Quantnuum’s work also indicates a broader point on the interaction of quantum calculation and AI: that the boundaries between the two fields change as both develop. It always seemed that two technologies would act side by side, where each would take over the computer work that are most appropriate.
Since AI has progressed with a fast pace, a quantum is promising even more than reality, how the marriage of the two will continue to be difficult to predict.