High-performance supercomputing has traditionally been the backbone of scientific research. However, its role is evolving as it becomes a vital resource for training complex artificial intelligence models, bridging the gap between HPC and AI. This transformation is not just reshaping these technologies but also influencing the manner in which knowledge is created and positioned globally.
At the 74th Nobel Laureate Meeting in Lindau, Germany, WIRED spoke with Jack Dongarra, a prominent US computer scientist recognized for his significant contributions to HPC software over the last four decades and a recipient of the Turing Award in 2021.
Dongarra highlighted the critical role that AI currently plays in scientific discovery. He explained that AI aids in computing and simulating behaviors more efficiently. As traditional modeling techniques get complemented by AI, researchers are finding faster and better ways to derive solutions to complex problems.
He believes that AI’s impact will extend well beyond the scientific domain, predicting it will become more significant than the internet in our lives, with applications we have yet to fully explore.
Turning to quantum computing, Dongarra expressed excitement about its potential but suggested we have a long journey ahead. He described today’s quantum computers as primitive, noting that while they can provide a probability distribution for potential solutions, they do not yet yield definitive answers. Dongarra cautioned against the hype surrounding quantum technology, suggesting that it may face a downturn similar to past cycles seen with AI. He emphasized the need for robust quantum algorithms and infrastructure for quantum computing to mature.
Dongarra addressed the geopolitical climate, particularly the strategic competition between the US and China concerning technology development and sharing. The US has restricted specific computing technologies from reaching China, yet he noted that unofficial channels still distribute advanced computing resources into the country. China has responded by investing heavily in domestic technology development, potentially accelerating its advancements in sophisticated computing solutions.
As AI continues to grow, Dongarra predicts that the role of programmers will shift. He anticipates a future where software could be developed through natural language prompts, reducing the time and effort required in programming. However, he warned of the potential pitfalls associated with AI, such as errors in outputs, stressing the need for verification methods to ensure solution accuracy.
In summary, the convergence of AI and supercomputing is just beginning to reshape not only technological landscapes but also the geopolitical and professional environments surrounding these fields. As advancements continue, both challenges and opportunities will arise, signaling a pivotal era in computational technology.
For more details, you can learn about the Lindau Nobel Laureate Meetings.
