It has been 25 years since I formed my first impressions of the Internet. I thought that it would shift the balance of power away from large organizations. I thought that individuals and smaller entities would gain more autonomy. What we see today is not what I hoped for back then.
In 1993, I did not picture people having their online experience being “fed” to them by large corporations using mysterious algorithms. Instead, I envisioned individuals in control, creating and exploring on their own.
In hindsight, I think that four developments took place that changed the direction of the Internet.
To me, blogs symbolize the “old vision” of the Internet, and Facebook epitomizes the new trend.
When you read blogs, you make your own deliberate choices about which writers to follow. With Facebook, you rely on the “feed” provided by the artificial intelligence algorithm.
Blog writers put effort into their work. They develop a distinctive style. In general, there are two types of blog posts. One type is a collection of links that the blogger believes will be interesting. The other type is a single reference, for which the blogger will provide a quote and additional commentary. On Facebook, many posts are just mindless “shares” where the person doing the sharing adds nothing to what he or she is sharing.
Bloggers create “metadata.” They put their posts into categories, and they add keyword tags. This allows readers to filter what they read. It has the potential to allow for sophisticated searching of blog posts by topic. On Facebook, the artificial intelligence tries to infer our interests from our behavior. We do not select topics ourselves.
The most popular environment for reading and writing blogs is the personal computer, which allows a reader time to think and gives a writer a tool for composing and editing several paragraphs. The most popular environment for reading and posting to Facebook is the smart phone, which favors rapid scrolling and photos with just a few words included.
Before August of 1995, ordinary households were kept off the World Wide Web by significant technical barriers. Until Microsoft released Windows 95, people with Windows computers could not access the Internet without installing additional software. And until America Online provided Web access, the users of the most popular networking service were limited to email and other more primitive Internet protocols.
The fall of 1995 began the period of mass-market adoption of the Internet. Another important leap occurred early in 2007, when Apple’s iPhone spurred the use of Internet-enabled smart phones.
As the masses immigrated to the Internet, the average character of the users changed. Early settlers were very focused on preserving anonymity and privacy. Recent arrivals seem more concerned with getting noticed. Although early settlers were intrigued by entertainment on the Internet, for the most part they valued its practical uses more highly. Recent arrivals demand much more entertainment. Early settlers wanted to be active participants in building the World Wide Web and to explore its various strands. Recent arrivals are more passive users of sites like Google and Wikipedia.
Hal Varian, a keen observer of technology who became the chief economist at Google, once wrote a paper that contrasted software that is easy to learn with software that is easy to use. Sometimes, software that is a bit harder to learn can be more powerful. But catering to the mass market can lead software developers to focus on making the software easy to learn rather than easy to use. This distinction may be useful for understanding how Facebook triumphed over blogging.
Back in the 1990s, many of us thought that since everyone could have their own web site, all web sites were created approximately equal. In Free Agent Nation, Dan Pink exuberantly proclaimed that the Internet fulfilled Marx’s vision of workers owning the means of production. We thought that the “means of production” were computers connected to the Internet, and they were accessible to individuals.
Instead, enormous advantages accrued to large companies that could amass vast stores of user data and then mine that data using artificial intelligence. If the “means of production” today are Big Data and the algorithms to exploit it, then the means of production are much more accessible to Facebook, Apple, Amazon, and Google than they are to ordinary individuals.
Although America Online was a powerful franchise in the mid-1990s, its glory soon faded. We thought that the reason for this was that AOL was a “walled garden,” as opposed to the open Internet. The pattern that we noticed was that closed systems tended to lose out. This was the explanation for the near-demise of Apple Computer, which was much less friendly to outside developers than its competitor, Microsoft.
Today, the iPhone is much closer to a walled garden than smart phones that use the Android operating system. Yet the iPhone has maintained a powerful market position.
Facebook is much closer to a walled garden than is the world of blogs. But Facebook grew rapidly in recent years, and blogs are getting less attention.
Traditional mass media was “pushed” to the users. If you wanted to watch a TV program in 1970, you could not record it or stream it. You had to turn your set to the right channel at the right time.
The World Wide Web was designed as a “pull” technology. You would make the choice to visit a web site, often by following links from other web sites.
Big corporations and advertisers are more comfortable with “push” than with “pull.” But in the 1990s, it looked like “pull” was going to win. One of the first efforts at “push technology,” Pointcast Network, famously flopped.
Today, “push technology” is everywhere, in the form of “notifications.” 21st-century consumers, especially smart phone owners, seem to welcome it.
The traditional telephone system put a lot of intelligence in the middle of the network. Central switchboards did a lot of the connecting work. Sound pulses traveled over wires, and your phone, sitting on the edge of the network, did not have to be intelligent to make sound pulses intelligible. But by the same token, your phone could only respond to sound pulses, not to text or video.
With the Internet, all forms of content are reduced to small digital packets, and the routers in the middle of the network do not know what is in those packets. Only when the packets reach their destination are they re-assembled and then converted to text, sound, or video by an intelligent device located on the edge.
Hence, the Internet was described as a dumb network with intelligence on the edge. One of the characteristics of such a network is that it is difficult to censor. If you do not know the content of packets until they reach the edge, by then it is too late to censor them.
Today, governments are better able to meet the challenge of censoring the Internet. Part of the reason is that the Internet is less de-centralized than it once was. It turns out that in order to process today’s volume of content efficiently, the Internet needs more intelligence in the network itself.
The advent of “cloud computing” also changes the relationship between the edge and the network. The “cloud” is an intelligent center, and the many devices that rely on the “cloud” are in that respect somewhat less intelligent than the computers that used the Internet in the 1990s.
Another factor is the importance of major service providers, such as Google and Facebook. These mega-sites give government officials targets to attack when they are not pleased with what they see.
One of the aspects of the Internet that intrigued me the most in 1993 was its governance mechanism. You can get the flavor of it by reading this brief history of the Internet, written twenty years ago. In particular, note the role of Requests for Comments (RFCs) and Internet Engineering Task Force Working Groups, which I will refer to as IETFs.
I compare IETFs with government agencies this way:
— IETFs are staffed by part-time or limited-term volunteers, whose compensation comes from their regular employers (universities, corporations, government agencies). Agencies are staffed by full-time permanent employees, using taxpayer dollars.
— IETFs solve the problems that they work on. Agencies perpetuate the problems that they work on.
— A particular group of engineers in an IETF disbands once it has solved its problem. An agency never disbands.
When I hear calls for government regulation of the Internet, to me that sounds like a step backward. The IETF approach to regulation seems much better than the agency approach.
Call me a snob or an old fogy, but I am not happy with where the Internet is today. I believe that things could change. I think that a lot of people are unhappy with the current state of the Internet. But I suspect that the enemy is us.
I am not sure what the solution will look like. I don’t think that regulating Facebook is the answer, especially if the main driver of regulation is that people are upset that Donald Trump won the 2016 election.
I don’t think that blockchain is the answer, even though it has some of the characteristics of the 1990s Internet. I have little confidence that blockchain can scale gracefully, given what we have seen so far and given the way that the Internet has evolved. And even if blockchain is able to overcome scaling problems, I think that the lesson of the last 25 years is that culture pushes on technology harder than technology pushes on culture.
I think that the challenge that we face on the Internet is the challenge that we face in society in general. In our modern world, we thrive by doing less ourselves and getting more from the services provided by others. But we seem tempted to become passive and careless in ceding power to governments and other large organizations.
In short, how can we sustain an ethic of individual responsibility while enjoying the benefits of extreme interdependence?