Scaling Artificial Intelligence
A zettabyte is 270 bytes or 1 billion terabytes, which is 1 billion gigabytes
History of Data
Six years ago we commented to a Webinar audience, “2,300 years ago, Macedonian information workers gathered all the knowledge in the world into the great Library of Alexandria. Its 200,000 books were made available to the fewer readers than that in the whole world.” Six years ago Forbes estimated that total world knowledge amounted to about 44 zettabytes, 44 trillion gigabytes of data, a previously unimaginable amount.
In addition to an unimaginable amount of information, it was also nearly useless, with usage rates declining precipitously as gross amounts of data exploded. Bernard Marr estimated that collectively we were using only about .5%, one-half of one percent, of the available data. He added that increasing average usage to merely .55% would dramatically increase individual company profits, productivity and the economy.
Because the growth and growth rate of new information are both increasing exponentially, it’s impossible to estimate current usage, but our somewhat informed guess is that it hasn’t moved appreciably even though total Internet users have increased from about 1 billion 20 years ago to more than 5 billion today. Data growth has been that extreme.
Comparisons Over Time
The numbers are truly staggering. From the Library of Alexandria with its 200,000 volumes holding the complete knowledge of the world to the Library of Congress, with its estimated 200 terabytes of digital and physical data, is both quite a leap and not even a drop in the bucket.
Eric Schmidt posited that from the very beginning of humanity to the year 2003, an estimated 5 exabytes of information was created, about 0.5% of a zettabyte. Today we produce that much in a couple of hours.
Today we have dozens of million-square foot data-centers that can and do move a significant percentage of Forbes’ estimate all by themselves. Artificial Intelligence centers, with far fewer people working than in manufacturing, transportation, retail and services, deliver high economic output through automation and optimization.
The old saying was “Knowledge is power.” The new one is “Data is riches and growth.”
In 1955, U.S. manufacturing employment was 16.6 million, or 32% of the total. There was tremendous popular anxiety that automation would create unemployment at an existential level. Instead total employment has grown from 65 million to 161 million, who enjoy a standard of living almost unimaginable 70 years ago.
We believe the hype on AI as a competitor to knowledge workers is overblown. Its real value is its ability to empower them, and all of us.
At TFT our role is to enable customers to put all that information to good use more efficiently and accurately than through “copy, cut and paste.”
