A lot has changed over 30 years …
For many of us, this is our second (maybe third) time in the market focused on trying to move from hype to reality with Artificial Intelligence. The first textbook I ever read on AI was Artificial Intelligence by Patrick Henry Winston. This was the second edition, published in July of 1984. This was one of the texts assigned in a LISP programming course I was taking at NYU’s Stern school of business when I was pursuing my MBA in Information Systems. This was the era of expert systems and Winston’s book focuses extensively on problem solving paradigms, representing commonsense knowledge, and learning rules from experience. Many of us spent months developing expert systems designed to pick stocks or identify arbitrage opportunities, or improve industrial operations with very limited success. The focus was on emulating human thinking, and automating the process to a level where it could match a human expert. This was the age of problem representation and simplification many years before machine learning revolutionized how we think about AI.
The most recent book I have read on AI is Kai-fu Lee’s excellent AI SUPER-POWERS. This is really the story of how machine learning and large volumes of data are giving users super-powers, as well as how the global super-powers (the US and China) will leverage AI for economic advantage and competition. Kai-fu does an excellent job in explaining how we got it wrong in previous generations of thinking, It wasn’t about replicating the thinking of experts, but determining what the experts were missing in the data. Kai-fu walks us through the evolution of data driven machine learning from internet AI to business AI to perception AI and finally far in the future artificial general intelligence. For those of us focused on IoT and supply-chain the role and need for large volumes of data and the associated contextual attributes in this evolution should be obvious. If we want to quickly improve outcomes, performance, cost, and resiliency then instrumenting our supply chains is essential. Ultimately the data collected in our digital twins may lead to new insight that exceed the value of the operational efficiencies driving current implementations of IoT in the supply chain.