The Internet

As I discuss in Become an AI company in 90 days, AI is a broad, revolutionary technology like electricity, computers, and the internet. The internet provides a great recent case study on what to expect from an exponential technology.

Let’s use the number of worldwide websites as a proxy for the internet’s early adoption curve. Figure 4 plots the number of websites for the first 15 years of the internet’s existence. Nine years passed before the internet grew to 10 million websites. In its first decade, the internet grew at the rate of less than a million websites per year. The exponential growth curve was initially slower than linear growth . . . until it wasn’t.


Tim Berners-Lee is credited with creating the first website in 1991. I first heard the term the internet in 1993 when I started playing around with FTP, Gopher, and America Online. It was only a curiosity until I first tried Netscape Navigator as a graduate student at Stanford in 1994. I instantly became an internet evangelist and started creating basic static websites.

It was clear to me that the internet was about to change everything. But the technology was hardly useful for most people. I remember people asking me questions like, “Where is the internet?” and “How big is the market for the internet?”

In those early years, web “pages” were hard to build. The internet was difficult to access outside of a college campus. It was painfully slow, and even small images took forever to download. There were no standards for transmitting data fast or doing secure credit card transactions. Almost no one (including me) could have envisioned how much this exponential technology would change the world.


In early 1996, Mary Meeker of Morgan Stanley published the Internet trends report. Although it was designed for investors, the annual report quickly became the bible for anyone interested in tech and its impact on the world.

Meeker’s report was the first widely read publication that explained how the internet would reshape business and society for decades. It signaled the beginning of the transformational phase.

Here is an excerpt from the report’s summary:

Using the Internet as an information distribution vehicle offers companies the ability to reduce distribution costs, support costs, and cost of goods sold—and eventually to target focused customer bases. On the flip side, lower costs and easier distribution open markets for new competitors in publishing, marketing/advertising, commerce, and software development. Dislocations of traditional companies in these areas are likely in time [emphasis added] as new business models based on free trials, subscribers, advertising, and transactions emerge. . . . New companies will emerge and poorly positioned companies will die.

This was written in 1995! Google didn’t exist, and Amazon was only briefly mentioned under the heading “Coolest Commerce,” as shown in Figure 5.

Figure 5: Meeker’s screenshot and description of a fledgling Amazon.com.

The overarching lesson here is that in Internet trends, Meeker put every retailer and ad driven media company on notice. This publication was the bellwether that something big was about to happen, and those who didn’t act risked extinction. In hindsight we know that retail and technology companies had a small window of opportunity to act before Google, Amazon, and eventually Facebook would begin driving them out of business.

Unsurprisingly, the report highlights skepticism by some of the most informed industry leaders:

It’s likely that development of the Internet won’t be as easy as it sometimes appears. Bob Metcalfe, inventor of Ethernet and founder of 3Com, recently offered a list of provoking thoughts about what could crush the Internet, soon. . . .

But evidence that supported Meeker’s assertions continued to pile up over the subsequent five years of the transformational phase. Modems and computers got faster. Broadband began to roll out. HTML standards emerged, and the Java programming language was born. Innovative companies like Akamai and Cisco developed ways to transmit TCP/IP data more efficiently.

Still, the internet grew slowly even as dotcom valuations ballooned. Critics pointed to the slow adoption of the internet as evidence that it would never amount to anything. Of course, they were making the classic mistake of misreading the slow rise of the exponential curve as slow linear growth.

Critics were hard to ignore because they bolstered their arguments with indisputable facts. I bought some Amazon stock in 1998, and a colleague advised me to sell it because Walmart’s scale would allow it to crush Amazon if it so chose. (And I listened to him!)

As the internet moved through the transformational phase, the cost of entry continued to rise for legacy businesses. Talent was scarce, marketing costs increased, and legacy companies spent tremendous capital trying to catch up in the internet race. In 1999, Disney cut its way into the search space by purchasing Infoseek for $7 billion—a move they could have made for a pittance three years earlier had they taken Meeker’s advice.


The transformational phase ended in March 2000 with the collapse of the dotcom bubble. Speculative companies collapsed, and the critics felt vindicated. I shut down my own startup, Soapbox.com, when I couldn’t raise additional money for it.

But what felt like the end of the internet was actually the end of opportunity, as Figure 4 illustrates. Amazon, Google, and Facebook would continue riding the internet’s exponential growth exactly as Meeker predicted.

By 2000, it was too late for retailers and advertising-driven media platforms to prevail. The pricing power of Amazon Prime allowed it to drive smaller competitors out of business. Google’s AdWords platform pulled advertising dollars away from print media and led to the collapse of regional media properties. The traditional players couldn’t catch up or muster the capital to buy disruptive new entrants like Instagram, WhatsApp, and YouTube.


Let’s consider how a traditional retailer could have responded to the internet. When the internet’s foundational phase began, Toys “R” Us was so dominant in the toys sector it was considered a category killer. It drove every competitor out of business.

In 1999—the same year Jeff Bezos was named Time’s man of the year—Toys “R” Us attempted to catch up in the e-commerce race. Their rushed and botched execution resulted in a disastrous holiday season.

After failing to deliver gifts on time for Christmas morning, the company entered into a 10-year e-commerce partnership with Amazon. But it was already too late for Toys “R” Us. The company went through a slow, painful decline that ultimately resulted in bankruptcy in 2017.

Presumably, some executives at Toys “R” Us had read Mary Meeker’s internet research report in early 1996. But based on the company’s slow response, we can assume they didn’t take the potential threat seriously. We can assume that no company leaders asked the most important question: What if she’s right? Toys “R” Us had a window of opportunity to begin investing in e-commerce for practically nothing. They could have started developing toysrus.com when expectations, costs, and demand were all low. Or they could have invested in an e-commerce startup. Or they could have signed a favorable deal with Amazon.

During the transformational phase, the open window of opportunity allows companies to begin riding the exponential technology curve at relatively low costs. Toys “R” Us is a cautionary example of a company that failed to act until it was too late.

The Internet

Getting started? Get our book!

The complete guide for understanding AI, identifying opportunities, and launching your first product and become an AI Company in 90 days.