Sunday, October 14, 2018

The value of breadth

As the present regime shifts to one controlled by uncertainty and accelerating technology, the premium on breadth of knowledge compared to depth, continues to increase. Innovation appears to happen at the intersection of fields and not in secluded domains and that is an important issue that educational institutions need to consider as they design futuristic curriculums.

It is also highly problematic. As an example, business graduates tend to be broad and shallow and if the stated hypothesis holds true, they should do well in the future. However, it is more complex than that. As the value of the individual is intricately correlated with how she can improve the economics of societies, it is not just the breadth of knowledge that gets into the objective function but also applications of it. Here, shallow and broad knowledge suffers from interactions with institutional constraints. The politicians in Washington are certainly shallow and not necessarily broad. Without deep knowledge of technology, policy-makers are ill-equipped to do anything good for future generations, let alone for themselves. They could be well advised to go deeper, perhaps a few inches below the ground they walk.

A bifurcation is in the cards. There is an optimum shape - breadth over depth - that optimizes societal progress. The pride of the country, the graduate schools, are without a clue, chasing after the latest "trends," to optimize localized economics. They need and want to make money by selling the latest wares and they are beginning to resemble a used car sales tent. It is unfortunate. Those who want to advance knowledge and improve societies, need to get out of the artificial and rigid constraints placed on them - academic, political and economic - and attempt to shape thinking beyond their career horizons. That is a tough ask even for those who already secured tenure, but still counting the number of incremental papers they published in marginal journals.

The value of breadth is increasing but not for those who do not have any depth.

Tuesday, October 9, 2018

Right brained Artificial Intelligence

As Artificial Intelligence becomes more commonplace through hype and reality, it may be useful to characterize the plethora of methodologies and technologies that are part of the thoroughly confusing medley. Conventional AI appears to be largely driven by the left brain as engineers, data scientists, and technologists flock to the dream, ably assisted by capital, seeking returns somewhere. Generally speaking, that is a prescription for disaster as technology, data and mathematics do not typically solve any problem of importance to enterprises. Granted, game playing is interesting and faking human voices and interactions equally compelling but none of these are going to change anything in the lives of ordinary people. And, they add little value to the economy or even companies.

The search giant recently proclaimed that AI gets more aggressive as they get better. This observation is not substantially different from the twitter girl created by another giant, that turned nasty. What these companies seem to be missing is that building AI bottoms up from historical data will simply reflect existing information content. More generally, these AI agents should reflect society and such observations add no value to the emerging arena, except talking points. And as the hardware company found out down South recently, transforming an organization requires a bit more than a "pizza-sized box," albeit it has solved most of the world's problems already.

It is about economics, stupid!. And that requires the silent right brain. AI has enormous potential but only if they are developed with a right brain dominance. It is a tough task as the normally shy right brain prefers to work from the background and simply muffles out the noise created by the left hemisphere. The old-fashioned concept of, "seeing the big picture," is still very important before diving into the details. Education systems tend to churn out left only brains in great numbers and this is problematic for the emerging regime. Scientists, whether real or of the data kind, cannot solve the problems facing humanity, let alone companies.

As the singularity enthusiasts revise the date of arrival of the discontinuity, it is important to remember that civilizations did not advance by tactics in the past and it is unlikely in the future.