Can You Predict the Future?
…and why failing to do so is dangerous
Many years ago I heard a story. The story was of hockey great Wayne Gretzky being asked why he was great. His answer was that most people skate where the puck is but he skates where the puck is going to me.
As someone who has been ahead of his time in AI and other fields I can appreciate the statement and have lately been givingit some more thought.
Many years ago I studied AI for my doctoral work. I tried to combine several different fields into a unification theory of how to go from raw data to information. This process was discussed in another article so I won’t bore you here to rehash it.
Simply stated, 27+ years ago I was thinking how to go from data to information. Information was in the form of rules that model that data and provide insight on how the variables are related.
So where is the puck mowing towards now?
If I look around there are many successful instances of people and companies that can take data and convert it into information. One most people are aware of is how their clicks are converted into direct marketing to them.
For a moment lets keep it even more simple. Lets assume we have a bunch of data. That data is raw and we notice that the shape it follows is one of a line. We teach a machine learning algorithm to form a best fit line that minimizes the error. I know this is overkill and that there are simple procedures for finding a least square line.
BUT go with me… we have successfully modeled the data into information of the form y=mx+b where x is input, y is output, m is slope and b is where the line crosses the y-axis.
Where is the puck?
The next wave for AI is to go from Information to Knowledge. I am calling that IK 1.0
The line equation is a summary of how those two variables behave. An intelligent system that has taken data and converted it into an equation would then learn that the general equation form y=mx+b can be used to model other things.
For example y=mx+b can be
TotalCost= UnitCost*NumberOfUnitsProduced+ Fixed Cost
or
Force= Mass*Accelation + error
As you can tell the same information models different types of systems. Therefore, knowledge is when information has been processed and results are produced. That is the next wave. That is where the puck is going.
So why is failing to predict dangerous? We are living during times where our surrounding are changing drastically and so is our access to raw data. This will change which industries exist and how quickly new ones come along.
What does this mean?
Do you think in 5 years the US will need 5 million truck drivers? Will self driving trucks be the way? If it is, then how do you participate in the upcoming revolution in trucking? Without the future thinking you will not even know where the puck is going to be.
My four cents…
