Data
is not the new oil as some have claimed, and Machine Learning is not
the new electricity. The shift is so gigantic that itâ??s impossible
to come up with a fair analogy
We're
living through the most significant technological shift in human
history. Our smartphones are connected to thousands of computers and
access terabytes of data, and whether we realize it or not, our lives
are heavily impacted by algorithms- our news feeds, the products we
buy, our food, our transportation, etc.
The
shift is significant, not so much because we have access to unlimited
computational power and data, but rather, because a larger number of
things is now measurable.
The
confluence of unlimited computational power and unlimited data, in
conjunction with continuous advances in algorithms and hardware, mean
that Machine Learning (ML) is the driving force of this major
technological shift - and that is significant because computers will
play an increasingly important role in our decision-making and ML in
how the technology works.
On
one hand, measuring at scale allows Machine Learning algorithms to
make better predictions which help individuals and businesses make
decisions. On the other hand, it feeds the improvement of algorithms
that perform tasks that are better suited to computers.
Given
the quickly evolving landscape, how can developers and businesses,
old and new, best capitalize on this shift to have an impact and stay
ahead of the game?
First,
it's important to separate hype from reality. We won't have
human-like robots walking our streets and performing the very hard
jobs that require "soft skills" anytime soon, but AI and
Machine Learning are here to stay and they are the new reality of
computing.
Data
is not the new oil as some have claimed, and Machine Learning is not
the new electricity. The shift is so gigantic that it's impossible to
come up with a fair analogy. What is clear, is that every business in
the very near future will use ML and every developer will work on ML
as part of a standard set of computing tools. That's a reality today
for those that have embraced it.
Second,
understanding the ecosystem is critical. The Cloud plays a
significant role because it gives developers and businesses access
and flexibility in storing as much data as needed, and in instantly
scaling as needed, at low prices - size doesn't matter.
Individual
developers, small teams, mid-size, and large enterprises can all
leverage the Cloud and AI. And it also means that consumers, in many
verticals, will continue to expect competitive performance and free
or near free functionalities. So, we have the technology (cloud,
data), businesses (every business becomes a tech business), and
consumers (the "end point" for collection of data and
impact).
Third,
identifying internal opportunities and modifying processes is pivotal
in driving that shift. For developers, it means getting up to speed
on ML and understanding the subtleties of what different algorithms
offer. For medium-sized companies, that's having a strategy to
collect the right data, and execute on it, step by step. For large
companies, it's ensuring that data pipelines and processes align with
experimentation and leverage AI. In all cases, it's having clarity on
what needs to be measured so that the right metrics are in place.
It's
also important to avoid pitfalls, and there are many. On one extreme
sit the skeptics that think AI does not apply to them because they're
very far from being able to apply it. This is usually a misguided
notion, as the Cloud and many open source tools allow fairly quick
starts. Additionally, in most cases, data is already in existence in
one form or another. On the other extreme sit the dreamers who
believe that AI and the Cloud will magically solve everything.
In
practice, the biggest challenge most companies face is not having the
right expertise. For developers, learning ML has become increasingly
accessible, but a common issue is a disconnect between those with the
technical skills and those with the business experience. The reality
is that no matter what stage you are in - as a developer, or as a
company - the time to embrace AI is now.
AI
can be applied anywhere where you can collect data, measure, and make
predictions. Identify those opportunities and tie them to a specific
customer or business needs and start ensuring the quality of the data
is good and apply basic methods to start with as a proof of concept.
This includes asking the right questions. Then, look to the Cloud
because it gives you the flexibility to easily, quickly, and cheaply
try things out, and identify the right open source and learning
resources available.
In
this process, keep in mind what three factors above: separate hype
from reality and set reasonable expectations; have a clear
understanding of the technology (cloud, data), the business, and the
"customers" for the task, and have clarity on what is to be
measured and what the goals are.
If
you manage to put these things together and get started, even with a
small project, you'll already be participating in that gigantic
shift. Just be sure to have clear goals, iterate, and experiment.
Source:
Economic Times
