The rapid growth of artificial intelligence and the deep innovations taking place at the quantum levels are at the heart of the massive disruptions and discontinuities driving the national and global economy. This discussion highlighted the urgent need to strengthen the nation’s capabilities in these areas, and it introduced the Council's Alliance for Transformational Computing initiative that seeks to advance a major strategic investment in research and prototyping at the leading edge of computing in concert with U.S. global allies.
President and CEO of the MITRE Corporation Dr. Mark Peters kicked off the session focused on the transformative potential of AI and quantum computing.
“We need to have a strategy, sustained investments, and partnership modality that enables the United States to retain superiority in AI and quantum computing."
Dr. Mark Peters
President and CEO
The MITRE Corporation
According to Deputy Director for Science and Technology at Oak Ridge National Laboratory Dr. Susan Hubbard, the national laboratories have emerged as a key testing ground for how AI can assist the research process. Combining scientific data and the world’s most powerful supercomputers, the national laboratories are on the cutting edge of implementing AI as a scientific tool, from data analysis to more advanced hypothesis generation and simulation. One climate model emulator at Oak Ridge National Laboratory, Orbit, uses 113 billion parameters. In the future, AI models might design, run, and analyze experiments with only one human in the loop, dramatically streamlining the scientific process and helping the United States keep pace as a global innovator through this incredible boost to researcher productivity. Despite its potential as a research tool, Dr. Hubbard points out the challenges that remain for the scaling of AI, including in maintaining energy efficiency and guaranteeing security.
Opengrowth.Ventures Founder and Chairman Mr. Gunjan Sinha echoed Hubbard’s point about AI being in its infancy. He made it clear that, while the explosion of large language models like OpenAI’s ChatGPT have given us a perception of a field that has already taken off, the growth going forward is going to be exponential. What we see today is just the tip of the iceberg, comparable to the first proliferation of internet browsers in the early 1990’s: a landmark step, but only the beginning. Mr. Sinha stressed that one of the Council’s and its initiative, the Alliance for Transformational Computing’s, most important tasks concerning the future of AI would be helping to design intentionally what the future of AI ought to look like, rather than just letting it unfold by chance.
“The sky is the limit for quantum computing applications."
Dr. Chris Langer
Fellow
Quantinuum
Quantum computers are rapidly catching up with their conventional counterparts and will soon radically change how we approach computation. According to Quantinuum Fellow Dr. Chris Langer, quantum computers, exploiting the unique characteristics of quantum particles, perform tasks that conventional computers cannot. However, quantum computations have error rates about a trillion times higher than conventional ones, meaning huge leaps in precision and error resistance will be needed before computers relying on quantum computations can have practical utility. But our current machines are getting larger and less prone to errors, with a few leading machines now on the cusp of commercial deployment. While it may yet be some years before quantum computing is widely available, there is a clear path ahead toward viable machines.
President and Glenn L. Martin Professor of Aerospace Engineering at the University of Maryland, College Park Darryll Pines, made clear quantum computing was still very much in the “punch card” stage of development. Resting on several enabling technologies and techniques, like ion traps, superconductors, and neutral atoms, quantum computing has no theoretical barriers, but is still grappling with experimental ones. Dr. Pines hoped to avoid a “quantum winter” like the AI winter that was finally ended by the proliferation of large language models. To do so, he suggested a focus on incremental steps and better fault tolerance to bring real-world viability closer.
When discussing the future application of the new computing methods, Mr. Sinha predicted an AI future dominated not by large language models, but by small language models. He again compared the current era of AI to that of the early internet where, over time, proliferation and more targeted development led to far greater innovation. Tomorrow’s AI would be more focused on specific tasks, and more integrated with a sensor infrastructure to better collect real-world data to enable it to do so. Thus, building out a better infrastructure for AI is critical; Mr. Sinha criticized the at-times single-minded focus on GPUs when, in his opinion, the host of other technologies that will complement them are just as important.
Dr. Pines notes the University of Maryland is working with one of its successful spin-out startups, IonQ, to advance the application of new quantum computing technologies. They have set up the “National Quantum Lab,” with user facilities designed to let startups test their quantum algorithms and to make it widely to those interested in collaborative efforts. Work in computational biology, solid-state physics, and battery designs is underway. And beyond quantum computing, other quantum technologies — including quantum imaging, sensing, navigation, and timing — are ready to revolutionize fields as diverse as navigation and medicine.
Dr. Langer argued the most consequential application of quantum computing was going to be the simulation of new materials and chemicals. For example, quantum computers could help identify room-temperature semiconductors, new fertilizers, medicines, and batteries. These high-value use cases could give quantum computing the funding needed for its initial commercial development.
“AI is in its infancy. We have to figure out exactly what the future state ought to be by design, not by accident.”
Mr. Gunjan Sinha
Founder & Chairman
Opengrowth.Ventures
When asked how to win the global race for AI and quantum, Mr. Sinha observed that many of the necessary systems are already in place, but they are not adequately advertised. In his view, greater communications could demonstrate the incredible work being done in universities and national laboratories related to advanced computing. Such a public awareness campaign would draw in the resources and talent, especially from young STEM workers, that could supercharge the industry.
“We need partnership in quantum and AI in the same way we have used the to build up classical computing over the years.”
Dr. Susan Hubbard
Deputy Director for Science and Technology
Oak Ridge National Laboratory
Dr. Hubbard highlighted the need to avoid seeing these advanced computing technologies as discrete units to be pursued separately and rather as parts of a potential groundbreaking whole. For example, hybrid conventional-quantum computers should not be seen as a mere step toward a fully quantum system, but rather as necessary to maximize efficiency in calculations. She imagined a system governed by an AI, assigning tasks to different computers based on whether a task was best done by quantum or conventional computers.
“We share facilities, we share people. That is why partnerships are strategic and geographical.”
Dr. Darryll Pines
President and Glenn L. Martin Professor of Aerospace Engineering
University of Maryland, College Park
Dr. Pines echoed the need for strong collaborations, nodding to the longstanding partnership between the University of Maryland and the National Institute of Standards and Technology (NIST). While originally a very basic agreement, it has over 20 years resulted in fruitful partnerships that have led to a robust quantum ecosystem in Maryland. This sort of co-location and place-building will be critical going forward.
To conclude, Dr. Peters asked each participant what needed to be the focus for the next five years. Dr. Pines argued for national quantum use facilities, democratizing access to the latest technology to spur new ideas, was a key way to stay ahead of the United States’s global competitors. Agreeing, Dr. Langer hoped, this would help avoid the “quantum winter” problem by speeding up compute times. Finally, Mr. Sinha looked back at how the Industrial Revolution and Information Revolution spurred the creation of the “blue” and “white” collar professions, respectively, and suggested that, with a new AI revolution on the horizon, a new “silver” collar would be needed to manage it.
