He who moves most wins
Artificial Intelligence is all over the news and the internet today. It is going to make society an amazing utopia or a Malthusian horror. The reason for this is that AI has the potential to make a smart enough decision and act upon it in a shorter time span than anything else on the planet. It’s not a cold sentient being bent on destroying humanity, yet. So far it is a very good algorithm at finding otherwise hard to identify trends and clumps of data that represent good decisions assuming the utility function is correct. The important part there being the utility function which is a weird phrase if your not familiar with AI. It is essentially the representation of what is good in computer code. As an example, the utility function for AI software that controls an autonomous driving system would be to calculate the fewest deaths that would result from an unavoidable collision. However one of those lives might be your own. So the AI may decide to kill you in order to save the minivan full of kids behind you. So write your utility functions with care.
I find the topic of AI fascinating but I am only going to draw one concept out of it for the purposes of competition in business. That concept is the OODA loop. This concept was introduced to me by a naval historian friend of mine, and it immediately struck me as the same way that many video game AIs work. OODA stands for Observe, Orient, Decide, Act. It was originally used in military doctrine to measure the pace of a conflict. War is full of ever changing conditions, but it takes time for orders to be received and acted upon, and then more time to see what the impact of those orders were based on how well they were executed. This loop is for example about 30 minutes in 16th century naval warfare. A captain examines the sea, where his opponents are, his velocity, where his crew are, and how they are allocated among many other factors. He issues orders with the intent to let them execute and observe for about 30 minutes until new orders are issued based on the results of the previous set. 30 minutes is not a dogmatic number, just an average.
So you can imagine that this time has shrunk over the years as technology has speed everything up. So let’s take a minor switch and talk about video games. Many of them are military simulations, so the concept applies very well. AI in games uses the same logic. Look at the board, decide on a course of action, wait some time, reevaluate, and make a new set of actions. Games use the same logic to control your computer based opponents. The only difference between Easy, Medium, Hard, and Insane is the OODA loop time. On easy, the AI re-evaluates the board every n minutes, on Medium, it may be every minute, hard every 30 seconds, and insane every second. The utility function (destroy your virtual army) at each level is the same. What makes it harder to beat the computer is how often it can effectively reevaluate and execute. The faster it can decide, the faster it is outrunning your gray matter processor.
This is why small upstart companies beat larger established ones. They have a smaller OODA loop. I find this exercise to be a terrific tool to determine if you can really beat a competitor. It is somewhat predictive. Are you able to complete the cycle of evaluate your market, create new products, and serve customers faster than your competitors are completing that loop? Even if your market share is small, it will continue to grow if your OODA loop is smaller than the competition. If it’s not, find a niche where you can beat their loop time. Once you achieve this, then time is on your side.