Cartas al CEO: Machine Learning |
My
letter has been one of the chosen to be published in open as "aperitif"
of the book, and yesterday was published on the page of Kromann
Reumert, a Danish legal and business consultancy company that
collaborates with THINKERS50, under the title "Changing the Way we
understand technology" so I thought it might be interesting to share a
Spanish version here.
My letter is an attempt to explain to a
hypothetical CEO the importance of understanding the change in the role
of technology, how it has been going from being a simple automation
tool, to gradually take a completely different role, that of a tool with
a ability to analyze brutal data and is able to learn from these data
through various methods to reach processes of deep learning or
reinforcement learning that will change the World as we know it.
That
a machine is able to win the best human chess players, of jeopardy, of
go or of poker is not as such important in itself by the fact, but by
what it demonstrates: respectively, that a machine possesses a
computational brute force capable of surpassing any human brain, which
is also able to understand and process the natural language better than
many humans , which can even devise and invent new original strategies
competing against itself in a way that no human had ever been able to
do, and finally, who is able to study environments with uncertain or
unknown information and make better decisions than any human trained for
it. The letter, in fact, is an attempt to put into perspective
something that many still do not understand and are not able to
understand in its true dimension, but which is going to represent the
most important change we have experienced in the history of humankind.
I greatly appreciate my friends of BigML, company of which I am strategic advisor, the opportunity that for some time now provide me to learn and understand the technologies involved in the machine learning. Without that fundamental obligation to keep me up-to-date to be able to live up to it, it would have been infinitely more difficult to understand its true scope and dimension.
Then the Spanish text of my letter (original version in English here):
In
a few years, the idea we have of computers and computing in general has
changed dramatically, although the perception of a generation of
executives has not. Changing that perception and being able to
reinterpret what technology can do for our business is a more pressing
need every day.
For
many years, we understood computers as a form of automation. Any task
that had repetitive, intensive or tedious components could be able to be
automated using a computer with the appropriate program. The arrival of
computers in corporate environments was thus, choosing those areas
characterized by tedious routines, such as calculating and paying
payroll, accounting, etc., or those in which there was a legal
requirement for preservation of information.
The
idea that we had of a computer was that of a machine that could do the
same thing as a person, but faster, cheaper and with fewer errors. This
idea of computing as automation has been a constant in the approach of
the investments in technology since the beginning of the history of the
Corporative informatics.
For
some time, this approach has changed dramatically. When we see in the
news that a computer has been able to beat chess no less than the world
champion Garry Kasparov, who has defeated by wide margin the jeopardy
the best players in the history of the program, or has pulverized in the
game of go to the most recognized players in the world, and we do not
talk about the same computation that we knew so far : To carry out these
feats is not enough to have more power of calculation or to do the same
operations as a man but at more speed. It is not just a matter of brute
force: it is another way of raising things.
The
new frontier is called machine learning, and it will provoke such a
brutal change in the environment that it will reach the point of
minimizing what was at the time the impact of the Internet. A real
dimensional change that will define which companies prevail and which
simply disappear. It will change what we understand for work, and alter
society as a whole. And all of this, within about five years. At the
moment, the essence of the true competitive advantage is to be able to
provide sufficient data to feed machine learning algorithms that are
better, more efficient and more competitive than those of our
competitors. There's no more.