Artificial General Intelligence – The Advent of Polymathic Machines

• Daniela Bortoletto - Professor and Head of Subdepartment of Particle Physics
• Chris Lintott - Professor of Astrophysics
• Andre Lukas - Professor of Theoretical Physics

In this talk I discuss the general state of development of Artificial General
Intelligence and speculate that even intermediate capabilities will be sufficient to alter the
path and modes of scientific exploration and engineering in the relatively near future.

While there is great progress, overall, in Machine Learning, now often called AI as well,
and many emerging uses of that capability, there are only a few efforts pushing the
frontiers toward machines that exhibit general intelligence on a level that meets or
exceeds that of humans. But progress in this pursuit of general intelligence has been
dramatic in the last three years, and there is reason to believe that we will see the arrival
of polymathic machine intelligences within five years, even if they are not fully
generalized. But these multi-domain-specific systems, which will significantly exceed
human capabilities to ingest and operate on the staggering amounts of data that will
come from new forms of instrumentation in most fields of science and engineering, will
contribute directly to completely new insights and capabilities.

I present some of the current work in the AGI field and show the path and expected rate
of progress from where we are today to these future, generally-intelligent machines. I will
provide highlights of capabilities and applications that are already emerging in language,
art, science, and computer programming. I offer an example of how intelligent systems
can fill gaps in our ability to deal with high-dimensional problems.

I end the talk with my thoughts about how the emergence of the polymathic machines
will force an inversion in the relationship of the researcher or engineer and these highlycapable
machines. This is very different than thinking of the machine intelligence as a
tool to help humans to figure it out! It requires thinking about how humans help the
machine to develop a level of precise understanding of the subject at hand, allowing the
machine to reason in an integrated way across all of it, in ways that humans will never
achieve themselves.

Craig J. Mundie is President of Mundie & Associates. He joined Microsoft in 1992 and retired in 2014 as Chief Research and Strategy Officer (since 2007) and the Principal Technology-Policy Executive (since 1998). Previously he was the CEO and co-Founder of Alliant Computer Systems.

Mundie was co-Executive Chairman of Bridgewater Associates (2015-2016) and continues as an advisor to the Board and CEO/CTO/CIO offices. He is a Director of the Institute for Systems Biology. He advises the Microsoft CEO and oversees Microsoft’s Quantum Computing Program. He is an advisor to the CEO, and the Board of Trustees, of the Cleveland Clinic. Mundie also is an early investor and/or technology strategy advisor to Exicure, Ironnet Cybersecurity, Somalogic, OpenAI, Helion Energy, Azul 3D, Stoicheia, and Seekout. Mundie currently serves on the Program Committee of the annual Bilderberg Meeting and on the Steering Committee of the Bilderberg USA Organization.   

Mundie represented Microsoft and served, consecutively, Presidents Clinton, Bush and Obama on the National Security Telecommunications Advisory Council (NSTAC) and was personally appointed to Obama’s President’s Council of Advisors on Science and Technology (PCAST) for both terms. Mundie served from 2009 to 2016 as the founding business chair, alongside General Keith Alexander, Director of the NSA, of the Enduring Security Framework Group which coordinated cooperative efforts combining the DoD, Justice and DHS departments with technology and telecommunications companies to address urgent cybersecurity challenges when they emerged. Mundie was the Business Chairman of the 2011 APEC Ministerial and business leaders meeting in Hawaii when the USA was the host country.

Mundie received BSEE (1971) and MSICS (1972) degrees from the Georgia Institute of Technology and had Honorary Degrees conferred by Moscow State University (Russia) and Rensselaer Polytechnic Institute.