i3 - July/August 2020 - 33

By Jim Harris

Business

T EC H TACT I C S

AI is Set to Eat Software

oftware is Eating the World," wrote Marc Andreessen in
a Wall Street Journal essay in 2011. To his point, Skype
software is facilitating more international long-distance
phone calls than the major telcos combined. Netflix has more
subscribers than the major U.S. cable companies. And, Uber is
worth more than the taxicab companies in North America combined. So, software is indeed eating the world.
Of the largest market cap companies -
Apple, Amazon, Alphabet (parent of
Google), Berkshire Hathaway, Facebook
and Microsoft - six of the seven are primarily software companies.
In a Forbes article from August 2019,
Martijn van Attekum, Jie Mei and Tarry
Singh argue that artificial intelligence
(AI) is set to eat software. Here are three
of the reasons they cite:

1. Exponential increases in computing power especially AI chips

A gigaflop is conducting a billion transactions in a single second. In 1961 a
gigaflop cost $153 billion on mainframes. In 2019 it cost a penny on a
smartphone. Computing is basically free
and it's happening at the edge.

2. Data is exploding

AI and machine learning (ML) require
huge data sets to accurately identify patterns and make predictions. With the
rise of the Internet of Things (IoT) and
the generation of more data in the next
two years than in all of human history
combined, the data is too vast for
humans to make sense of - but for AI
and ML this is their feedstock.

Geest in Exponential Organizations.
By 2020 that cost dropped 750-fold to
$20,000 using AWS, exponentially
decreasing the risk for entrepreneurs
and exponentially increasing the number of AI startups that will experiment
to disrupt industries.

Critical Question for Executives

Are your products, services and business
models getting smarter, better, safer and
cheaper every day because of AI?
Management consulting firm McKinsey
says AI can add $13 trillion of additional
economic value by 2030. We are about
to go through the same transition with
AI as we did with software. Early adopters will dominate those who delay.
Jim Harris is the author of Blindsided.
Follow him on Twitter @JimHarris.

Self-Driving Vehicles
as an Example
Tesla's vehicles have driven two
billion actual miles in self-driving
mode. The huge amounts of data
give Tesla powerful insight into
improving self-driving cars and
safety. Lessons that the AI learns in
an individual car are relayed to the
cloud, and improvements are
shared with every other car in the
next update. The AI for Tesla is
getting better, smarter and safer
with each mile driven. These data
insights enable Tesla to offer Tesla
owners in California car insurance
at 20% to 30% less than traditional
insurance companies. Tesla will
eventually expand this beyond
California. Tesla can innovate its
business model (in this case
entering the insurance market)
because it has access to the driving
data. It knows that its self-driving
cars are safer than traditional cars
with its insured losses a fraction of
a traditional insurance company. In
the long term, Tesla's risk will be
16X less than traditional insurance
companies - given that 94% of car
accidents are due to driver error.

The AI for Tesla is
getting better,
smarter and safer
with each mile
driven.

Courtesy of Tesla

3. The Plummeting Cost of Cloud
Computing and Storage
Amazon Web Services (AWS) has cut
its prices 70 times since it launched in
2006. In 1995 it cost about $15 million
to build the IT infrastructure for a
Silicon Valley startup, notes Salim
Ismail, Michael Malone and Yuri Van

C TA . t e c h / i 3

i3_0720_33_Business_TechTactics.indd 33

JULY/AUGUST 2020

7/13/20 9:06 AM

http://www.twitter.com/JimHarris http://cta.tech/i3

i3 - July/August 2020

Table of Contents for the Digital Edition of i3 - July/August 2020