Giant companies like Google and Facebook and Amazon dominate today’s internet, with hands in just about everything you can do online — from search, to advertising, to storage.

And love them or hate them, those companies’ websites get more than six hundred million visitors a year, with hundreds of millions more seeing their ads or using their apps. But it wasn’t always like this.

We last left our history of the internet series in the mid-nineties, when the U. S. government had just given up control of the internet that they’d been in charge of since 1969, and Tim Berners-Lee’s World-Wide Web was starting to get popular. The internet was finally open to the public, and it seemed like a new frontier where anything was possible.

Just twenty years later, the internet is an essential piece of the modern world, with the public side mostly commanded by a few powerful companies. Getting from then to now took a little bit of economics, a lot of fast computers, and a couple of fundamental changes in what we expect when we come online. With society still figuring out how it feels about that last one.

The Web of the mid- to late-nineties was a little bit like the Wild West. It was a huge new world of people and companies popping up everywhere you looked, and it seemed like there was plenty of room for everyone to have a piece of the pie. Rich investors started piling money into companies without worrying about whether they were making a profit — or even whether they had many customers.

They figured that on the internet, old rules for being cautious about investing in young companies didn’t apply any more, and that all anyone with a good idea needed was enough money to reach an audience. But the honeymoon didn’t last, and what’s known as the dot-com bubble started bursting in March of 2000. Stocks in tech companies plummeted for the next couple years, and over half of them declared bankruptcy, eventually losing trillions of dollars in total.

And while stocks were falling, a couple famous court cases against Microsoft and the music-downloading service Napster put new boundaries on what companies could do with the internet. Napster had been one of the fastest-growing businesses in history, but it went bankrupt paying back musicians for copyright infringement. And Microsoft narrowly avoided being broken up after violating antitrust laws.

The internet might’ve been a whole new world back then, but the turbulent early 2000s left a tamer, steadier Web in its wake, where a few well-run companies could quietly become empires. One of the dot-com bubble’s casualties was GeoCities, one of the first social networking sites. Today’s Web is flooded with social networking: Just about every website constantly encourages everyone to use their personal profiles and share everything they see with everyone they know.

But things were different back in 1994, when GeoCities first came online. The first websites mostly had a clear divide between creators and users. Creators — well, created.

They wrote the computer code and assembled the different documents and files and hyperlinks and pictures that made the website what it was. The users just visited the website and looked at whatever the creators had put on there. But GeoCities worked on a different model.

Anyone and everyone could make a GeoCities account and create their own website, with all their own stuff and formatting and backgrounds and interests. GeoCities mixed together users and creators, because everyone with an account was both at the same time. Users could also send each other messages and join communities — or “neighborhoods” 2:57— of pages with similar topics and interests, creating whole sections of the site that focused on just about everything imaginable.

Its nineteen million users made GeoCities the third-most popular site on the Web in 1999, behind AOL and Yahoo!. Yahoo! bought GeoCities that year for 3.6 billion dollars, but it fell on hard times after the dot-com bubble burst and never managed to regain its former glory. Yahoo! finally shut down most of GeoCities back in 2009, when it had long-since been surpassed by other sites that took the idea of social networking and ran with it.

Now, we’ve spent a lot of time in these last couple videos talking about things you might never have heard of unless you worked with them directly: ARPANET, TCP/IP, CompuServe — all that stuff. But GeoCities gave way to some sites that pretty much everyone recognizes, whether you used them or not. Friendster launched in 2002, giving each user their own profile and a way of seeing different networks of friends on the site.

It quickly became enormously popular, with three million accounts in its first three months. But it was plagued by technical troubles, and after only a couple of years, it had fewer users than MySpace, which was itself passed in 2008 by a blue-themed site started in a Harvard dorm. You … might’ve heard of it.

Facebook now has almost two billion active users, and that number just keeps going up. But it’s far from the only social networking site out there. There’s Reddit, Twitter, Tumblr, LinkedIn, YouTube — and that’s not even counting all the Disqus comments sections and Digg share buttons and Wordpress blogs sprinkled all over the Web.  All those sites, big or small, are descended from GeoCities: the first site to mix users and creators in a completely new way of using the internet.

Most people still used dial-up to get online in the heydays of GeoCities and Friendster. But by the time Myspace took over around 2005, there was something new on the market. Instead of using dial-up, most people had switched over to broadband. See, dial-up has a built-in speed limit. Because of how phone lines were made, the fastest dial-up connection could only receive or transmit about 56 kilobits per second. That’s 56,000 ones and zeros in or out of the computer every second. Which might sound like a lot, but on dial-up, downloading a single song off Napster would take about ten minutes, even at top speed. And downloading a whole movie could take days. Even just loading a site with a few images took forever.

So once the internet started getting more popular, companies came up with better ways of getting online that weren’t so limited. Like DSL, which transmits digital data along phone lines instead of analog signals like dial-up does; and cable, which uses the wires for a cable box to connect to the internet. These and a bunch of others all came to be known as broadband, which is really just a broad term — get it? — for all the ways of getting online that aren’t dial-up.

Depending on the type of broadband, the connections can be tens or hundreds or even thousands of times faster than dial-up is. People started using broadband in the early 2000s, and in 2005 it overtook dial-up for the most popular way Americans got online. With more people on faster connections, it wasn’t as big a deal for sites to have a bunch of images, or even video. But all the data on those sites needed to be stored somewhere.

Another legacy of the dot-com bubble is the place of data centers in today’s internet, where hundreds of computers work together for a single company to give users a better, faster experience. The earliest computers could be big enough to take up entire rooms, but those dedicated computer rooms stuck around in a lot of places even as computers got smaller and faster. Eventually, they’d have something like ten computers in them, all connected together to act like one big computer with the combined speed and memory of all ten. These computer rooms or even computer buildings became known as data centers, and lots of tech companies had them in the eighties. But a lot of companies would also rent computers in a data center nearby. And that model became really popular as the dot-com bubble inflated.

Lots of those startups needed computer space to store all their data, and computer speed to handle all the users on their websites — at least, the lucky ones who had lots of data and users. And data centers were perfect for the job. Instead of owning, powering, cooling, and maintaining your own computers, and reliably connecting all of them together and to the internet, you could just pay a data center to do it all for you. And even after a lot of those startups went bankrupt, data centers held onto their role in handling a lot of the traffic for the big Web companies.

Today’s websites have more than a hundred and fifty times as much data on them as they did in 1995, and a lot of that information comes from rooms or buildings or whole complexes full of computers. After all, Mark Zuckerberg doesn’t keep everyone’s Facebook status on his own personal computer. They’re on an army of Facebook’s computers that are all working together — and the same goes for Google, or Amazon, or your bank, or your email provider, or even your college. Today’s Web relies on these huge collections of computers to work like we all expect it to.

And with all that speed and storage available, today’s websites can do things they couldn’t dream of in 1995. Just think about going on a site like YouTube. Even if you’re not signed in, you’re the only one in the world who sees that exact YouTube homepage, with those exact recommended videos. Because no one’s watched the same videos you have, in the same order, for the same amount of time, from the same places in the world — all stuff that YouTube uses to choose what to recommend. Or think about Facebook, where it seems like you only ever see pictures of weddings and babies and puppies instead of the hundred other status updates people have posted. Or when you see banner ads everywhere for products you were just looking at on Amazon. Just about everywhere on the Web has this sort of algorithmic filtering, where the website decides what you’re probably interested in and shows you more stuff like that.

This is somewhere else those data centers come in. They don’t just store all the website’s data; they run special programs to look through that data and use information that they have saved about you to decide what you want to see next. Sometimes, that might be a video that has four views but is exactly the sort of video you’d like to see. And sometimes, it’s the kind of Facebook status that gets lots of likes and comments, like wedding and baby photos. And in a lot of ways, this is great. I mean, who hasn’t binge-watched YouTube when those recommended videos are just what you’re looking for, and who doesn’t love pictures of babies and puppies?

But just like any technology, there are trade-offs to this new responsive Web. Facebook ran a study on some of its users and found that what people see on their News Feeds can strongly affect their moods, which means that algorithms are helping determine how we feel about the world without us even knowing it. And figuring out what you want to see can involve tracking you around the internet.

So a lot of people are worried about their privacy — especially because some companies aren’t shy about selling what they learn about people to the highest bidder. And it can seem like every day there’s another news story about governments tracking people around the internet in a similar way. Or maybe it seems like there’s never a story about government tracking. Whatever the algorithm decides to show you.

…. Which just about brings us to the present. In fifty years, the internet’s grown from four computers to billions, and from the western United States onto every continent and even into outer space. But it’s still pretty new, and it’s still evolving. No one knows what pieces of today’s internet are going to keep being central to the story when we look back in a few decades — or in a few centuries. Will it be remembered for all the amazing communities who do things like the Project for Awesome, where millions of dollars are raised by people who just want to decrease worldsuck? Or will today’s communities and social networks fade into the footnotes to bigger stories about privacy concerns and government influence? Ultimately, it’s up to us. The citizens of the internet get to decide what our priorities are, and we get to choose what kind of internet we want to be a part of — now and in the future. Because it might not be the Wild West anymore, but the largest collection of knowledge and ability in human history is still just getting going.

Thanks for joining us on our journey through the history of the internet! This mini-series wouldn’t have been possible without our patrons on Patreon. If you want to help support this show, just go to patreon.com/scishow. And don’t forget to go to youtube.com/scishow and subscribe!