When I was a child in the 1990s first learning about the internet, I often heard people call it the World Wide Web. If you’re like me, you might think the internet and the WWW are just different names for the same thing. But they technically aren’t. The internet is the physical network that connects our devices such as the one you’re using to read this newsletter – it’s undersea cables and racks of servers in data centers. The web is the thing we surfed after we connected our desktop computers’ dial-up modems to the internet in the 1990s – the web is those links that took us from one website to another.

While the internet is a series of tubes, the web is literally just a bunch of links. “Web” isn’t just a word that we thought sounded cool in the 1990s – it was meant as a metaphor for the links connecting every page on the internet to each other in an interconnected web. And the web is dying.

Nobody really “surfs the web” any more. We use the internet, but mostly through apps on our smartphones. Links are things we mainly click on to get work done, because most white collar work still happens on desktop computers – which nowadays are really just internet terminals. And even then, we spend most our internet time on the desktop inside a handful of internet-connected apps – a large fraction of desktop internet use these days is probably split across email, an instant messaging platform like Slack, a video conferencing platform like Zoom, and an office productivity suite like GSuite or Microsoft Office. More and more, we spend less and less of our internet time browsing on web browsers. We still use the internet to stay connected to other people – we’re just not surfing the web any more.

It might sound like I’m lamenting the slow death of the web, and it’s true that I am. But I also find it understandable. I also don’t think it’s necessarily permanent. I’ve been using the internet for just about three decades, which makes me old enough to be confident that how we use it is only going to keep evolving. Am I going to drop some kind of hot take about AI and the future? You’ll have to keep reading to find out.

A Brief History of the 1990s’ Internet

The World Wide Web was invented by a bunch of nerds to share academic documents with each other. The first webpages were all university websites and email was primarily a medium for academic communication. You might have heard that the Internet was created by the US government with the idea of having a communications network that would survive a nuclear war. That happened in the 1970s. Then in 1989 – coincidentally, at the end of the Cold War, and just before I was born – some guy named Tim Berners-Lee invented the idea of a worldwide web of pages linking to one another using the Internet. And thanks to the web, the Internet itself took off. The web was the first true reason that mass market consumers had to connect to and use the Internet.

To me, there were two things that the web really unlocked:

1. Links – the ability for an author or institution to point you to a completely unrelated thing they don’t own or control

2. Multimedia – the built-in ability to engage with images and videos, rather than just text

Like most other pre-WWW Internet technologies, email by itself doesn’t allow you to do either of these things. (Email as we use it nowadays does have linking and multimedia capabilities – but this isn’t true in the original way email systems were designed. Remember, email was created in the mid-20th century when text was basically the only thing that the Internet was capable of transmitting.) Without the WWW, you wouldn’t have any links to click on in your email.

In the early web, links were the only way you got around. Individual websites didn’t have much content on them so you would happily take referrals to other websites that covered topics you were interested in. You’d bookmark a handful of websites that you’d want to visit again, or memorize their domain names. Yahoo got its start as a website that was essentially a directory of interesting websites, so you could use it as a jumping off point to find other websites.

Eventually search engines came along, so people had a more scalable way to find websites they were interested in. Yahoo had a pretty popular search engine, but I remember using all kinds in the 1990s – for a certain generation the names Altavista, Lycos, and Ask Jeeves probably all ring a bell. All of these search engines sucked, so Google came along and made them history.

Google’s chief innovation that separated it from all the crappy search engines it replaced was that it treated links as votes. To identify the best webpages, Google just looked at who had the most links. I’m oversimplifying a complicated algorithm, of course. But I do think Google succeeded because it was the first search engine to truly understand the heart of the internet was links, and to take this seriously. Rather than just indexing each webpage as a bunch of text or a series of documents, Google built a search engine that actually indexed the world wide web of links, and in turn built a business empire on this.

This might be feeling like a boring history lesson, but remember, what I’m describing here is what a good number of people did for fun in the 1990s. To me, there is a very direct line from “surfing the web” thirty years ago to doomscrolling your feed on TikTok or Twitter today.

Taking time out of your day to sit at a desk and read random strangers’ personal websites doesn’t feel like it has much relation to swiping through a feed of videos. But I think more than half the fun of the early internet was not knowing what you might stumble on. You might stumble on a new online game to play, a new internet forum to join, or a new website with a trove of fun facts. The closest equivalent experience I can think of that still exists is going on a binge of Wikipedia articles, rabbitholing through a series of links learning about a whole area of knowledge newly fascinating to you which you had no idea even existed before today.

I recognize this is objectively quite a different – and much rarer – experience compared to the passive act of swiping through our TikTok feeds. But they share one thing in common – we are bored, and to pass the time we decide to see what other human beings out there have put out onto the Internet. The main difference is this used to be an arduous task of following links that often led to nowhere interesting, or trawling through low-quality search results – and all to just read a wall of text. Nowadays you can just lie in bed and occasionally lift a finger to surf a never-ending wave of short videos. But both of these activities are motivated by a curiosity about what’s going on in the world, and a desire to be entertained by the act of discovering.

Social Media and Web 2.0

The Internet of the 2000s and early 2010s to me was all about the rise of social media. We started this period with internet forums being the main form of social media, and this period ends with the utter domination of Facebook and its other platforms like Instagram. This is the time when links and the open web began to die, though we didn’t know it yet.

The first harbinger of this for me was Friendster, the first social networking site I ever used. For the first time, it was possible to learn about my friends, and my friends’ friends, by browsing profiles they had curated for themselves on the Internet. But the main thing to do on Friendster was to view other people’s profiles, or to chat with them. You weren’t really clicking on links to other things living outside of the Friendster ecosystem – that wasn’t the point. And really, it turns out that’s not the point of any social media platform.

This period also corresponds with the rise of blogging. This newsletter owes its name, Infernal Ramblings, to the blog I started in high school. (That’s also how I conveniently already owned the domain name infernalramblings.com.) Blogging was the total opposite of modern social media – linking was very much the point. Much of blogging was about being in conversation with other bloggers who wrote about the same topics as you – so of course you’d link to what they had to say, and vice-versa. Most blogs maintained extensive blogrolls – lists of other blogs they recommended to their readers. 

It’s hard for me not to look back fondly on the blog era. One of my oldest still-active friendships is with a fellow Malaysian blogger. As a young professional living in Washington, DC in the early 2010s, I spent quite a bit of time blogging at Open Borders: The Case, which introduced me to the fringes of the think tank community in DC that took open immigration seriously. Many of my friendships, and so many of the things I spent my free time on in my early 20s, stem in some way from the circles of bloggers I joined in the late 2000s and early 2010s.

But as most blogs have, all these efforts eventually sputtered as Facebook and other social media platforms began truly eating the world. A lot of the discourse that might previously have happened in blogs or blog comments moved into Facebook posts and groups.

The more time I spent on Facebook, the more apparent it was to me that this was breaking the link-friendly paradigm of the internet I’d been accustomed to. This wasn’t obvious at first, of course: a lot of the people I was friends with or followed would share interesting things they’d read or seen elsewhere on the internet on Facebook. I did the same; for much of the 2010s, I’d often hear from friends that my posts were some of the most interesting things they’d seen on Facebook, just because nearly all of my posts were links to things I’d read.

But I rarely shared things that had been posted on Facebook. After all, most Facebook posts were not meant to be read or viewed by a broad audience, if they could be seen by non-friends at all. Facebook encouraged us to spend our time creating content for circles of people we already know. Facebook Notes, the one feature I saw used occasionally for sharing content intended for a broader audience, never actually got used much and unsurprisingly was killed pretty much as soon as the decade ended.

Even if I saw something interesting on Facebook – say an interesting discussion in the comments on a post – it was somewhere between difficult and impossible to share a link to the conversation. The easiest way to share this was to just take a screenshot and post it as an image. If I wanted to publish my own ideas as a Facebook Note, linking to this from outside the Facebook ecosystem was also difficult – people would need to login to a Facebook account to see what I’d written.

Like Friendster, Facebook is basically a walled garden. The content on Facebook isn’t meant to be shared or discoverable outside Facebook. The point of spending time on Facebook is to keep spending time on Facebook, exploring its feed. At one point during the infancy of mobile gaming, the point of spending time on Facebook was to play games – it was the only way to play FarmVille.

At some point, likely in the late 2010s, Facebook began suppressing the distribution of posts with links. I saw fewer and fewer interesting links in my Facebook feed, and fewer of my friends saw my posts linking to interesting things I’d seen elsewhere on the web. Today, hardly any of the content you’ll see on Facebook is a link to something outside Meta’s walled gardens. Ironically, in 2016, hoping to some day make the interesting links I’d curated more visible outside the walls of Facebook, a friend registered the domain johnleesharedalink.com. The idea never took off – in hindsight, that year was probably the peak of link-sharing on Facebook.

Of course, there were alternative social media platforms that tied themselves more closely to the open web. Pinterest, my former employer, is a notable one. I’m old enough to remember using Digg and Reddit early in their history – both platforms purely meant for sharing interesting links. (Digg users loved making fun of Reddit’s ugly Web 1.0 interface compared to Digg’s more modern design, but somehow Reddit is still around today while Digg is not.) While today in 2025 both Pinterest and Reddit de-emphasize links compared to their earlier designs, both platforms still owe their fundamental architecture to an original purpose of making it easy to share links to interesting things you’ve seen elsewhere on the internet – that’s the core value proposition of their product. 

Unfortunately, when you search for almost anything on Pinterest today, it also often seems like much of the content is also stuck in the 2010s. Without blogs, it turned out there wasn’t very much else worth linking to – Pinterest is still a great way to find a blog post about planning a travel itinerary or a wedding, but you’ll be constrained by the reality that very few people (as opposed to companies or LLMs) are still blogging about planning trips or weddings. Peruse the content of any blog you find on Pinterest or scroll to the comments and most of the time you’ll be met with evidence that the post has been live for years, belying the seemingly recent publication data in the header of the post. Today, the most novel and popular content on both Reddit and Pinterest tends to be content natively posted to these platforms – Gen Zs love using Pinterest, not to find links to click on, but to curate boards of aesthetic pictures.

The most notable attempt to buck social media’s rotting of the web I think came from Google Plus. True to the ethos of its search engine, Google Plus made posts globally visible by default, and easy to share and link to. You could always opt your posts into more private degrees of visibility – the infamous “Circles” – but by default new users’ posts were for the world. And people absolutely hated this. Google Plus never got any meaningful traction independent from Google attempting to shove it down its users’ throats.

I personally found the paradigm of open posting exciting – but I never took to the platform, because nobody I knew did. Many people were incredibly scared of having their posts visible to everyone by default, even though Google made it easy to limit sharing to just your circle of friends. Today, I’d consider LinkedIn the successor of Google Plus – its posts are relatively easy to share off-platform, and it has successfully defaulted its users into the paradigm of public posting, while still allowing you to limit posts’ visibility to a closer circle as you require. Google Plus was just ahead of its time.

How the Smartphone, Short Form Video, and the Feed Killed What’s Left of the Open Web

At some point in the mid-2010s, I heard of an app called Musical.ly. Reportedly popular with teenagers, it was a short-form video platform that basically let you post brief clips of yourself dancing or lip syncing to popular music. I recall getting quite a few emails from recruiters trying to sell me on a role there, which was the main way I remembered this product existed. I didn’t get the big deal about it, though I knew from the numbers it was very popular with young people; it didn’t sound that interesting to me, so I never followed up about any job opportunities there. At some point the recruiter outreach from Musical.ly stopped, and the company got sold to Bytedance. The Musical.ly app was rebranded TikTok in 2018; it’s one of the biggest apps in the world today.

TikTok exploded in popularity during the early days of the covid pandemic, and that was the first time I remember having the impression that it was no longer just an app that only teenagers used. In early 2021 I decided to download TikTok, knowing that based on everything I’d heard, it’d likely become a huge timesuck. I’ve been an almost-daily active user of the app since.

Instagram and TikTok are the only products I’ve mentioned so far that were not born as part of the web. Everything else started as a website; the expectation at inception was that you’d use it on a desktop computer. Everything else also has a smartphone app these days, but social multimedia platforms like Instagram and TikTok were undoubtedly first built as apps and later had websites built on top. A social video platform could not exist without the smartphone putting an internet-connected camera in basically every human being’s pocket.

A different way of talking about the history of the internet in the 2010s is that this is the decade when smartphone apps took over computing and the internet. At the dawn of the decade, everything internet-related you did was through a web browser or email client – even if you had a smartphone, which the vast majority of humanity did not. By the end of 2019, mobile devices had overtaken computers as the main platform we use to spend time on the internet.

For the first time, internet users had access to a plethora of apps – Uber, DoorDash, Instacart, Grab, TikTok, Snapchat – that could only ever exist on a smartphone. This generation of apps could take advantage of the smartphone’s capabilities natively in a way that older apps first built for the web trapped inside of a desktop machine could never have.

One of smartphones’ native capabilities is that they are all connected to the internet by default. The desktop computer was invented as a standalone device; with a smartphone, connectivity is the point.

But in breaking free of the desktop’s constraints, smartphones have also introduced new ones: links don’t have a place in how smartphones work. Clicking a link in an app that takes you to another app is rare, and it’s such an unpleasant experience that most major apps have built web browsers inside of their apps so you don’t ever have to leave their app. If you click the few links you can ever find inside an app like Facebook, Instagram, or TikTok, they will open by default in an in-app browser rather than your phone’s default web browser, because the experience of clicking a link to something outside the app you’re using is so jarring on the modern smartphone.

In the old web, links made it possible for you to engage with content published by dozens of different people and institutions in a single session. You could read what a random blogger had to say about something and juxtapose it alongside what a major newspaper had to say. In the smartphone era, we are limited to engaging with the products and content published by a handful of companies. From choosing which websites to visit or links to click, our choice has now dwindled to which multi-billion corporation’s app to open.

Of course, there is plenty of diverse content to see on the apps – every person will see content from many different people in their feed when they open Facebook, Instagram, or TikTok. But the rise of the feed as the navigation paradigm means that the publishing company has decided which creator’s content we’ll see. We only see what the algorithms allow us to see.

Going straight to the feed is actually a fantastic shortcut for the classic Web 1.0 behavior of firing up Google to search for something. Rather than having to type in what to search for and waiting for the algorithm’s response, why not just have an immediate firehose of all the stuff that the collective wisdom of big internet publishing companies (curated based on the people we are friends with or following on social media) think we want to see? If we actually do want to search for something specific, every social media app these days has a search function.

But all of this is still a poor replacement for what the traditional web enabled. Navigating to find a specific video is way harder than finding a specific webpage – and this is even after accounting for advances in technology that make videos searchable at all (thanks to advances in automatic transcription and image recognition). The information density of video cannot compare with text: if I want to find out what people think of a given restaurant, I can get the opinion of hundreds of people from scrolling reviews on Google Maps or Yelp in the same amount of time it would take me to watch a handful of Tiktoks.

Where video and the feed converge is catering to human laziness. It takes deliberate effort to use the legacy web to learn things. Even passively engaging as an idle consumer of entertainment content on the legacy web required active thinking: surfing the web assumes someone capable of choosing a link to click, or thinking of a search to run. Video is easy to consume and process, and a convenient feed serving it to us in short-form chunks makes it so we’re never far away from something new to watch.

If the Facebook feed made links unworkable by downranking, modern apps like Instagram and TikTok make links unworkable by not having a place for them. This is not necessarily inherent to the format, though it is not obvious to me how you can ever seamlessly insert a hyperlink into a video. It’s also a business choice: the only way most of us will ever click a link on Instagram or TikTok is through a link in someone’s profile, the one place that these platforms still make it convenient to link to the web.

Even organically, I think most of us would not care to click on links presented to us in the feed. It’s jarring enough to switch apps, but to switch from a feed of videos to a static webpage and from a mindset of passive consumption to active navigation is a lot to ask. If there’s one thing I’ve learned from years of working on internet products, it’s that you can hardly ever go wrong by betting that people will be lazy and take the path of least resistance.

But even so, making any sort of connection possible to the traditional web is inimical to the business model of the apps with feeds. Traditional web search engines made money by getting us to search more; they wanted us to spend the least amount of time possible searching and for us to click on links as fast as we could so we’d get to the right webpage, so that we’d then be able to think of the next thing we wanted to search for and run another search that they could put a new set of ads against. With feeds, it’s the total opposite: every feed-based platform wants to maximize the amount of time you spend in their specific walled garden, so that they can show you as many ads as possible. The more you can be lulled into a mindless scrolling session, the better.

Trapped in the Apps, Will We Ever Have a Web Again?

Today I see the internet as less an interconnected web of independent publishers where traffic flows freely across boundaries, and more a handful of concentrated fiefdoms jealously guarding their turf. The Facebook feed’s downranking of links was a foregone conclusion, especially once most of us began using the company’s properties on our phones instead of our computers.

Twitter for the longest time was an interesting counter to my view that an app designed around a feed has to be incompatible with an open web of links. For the longest time, it operated similarly to Pinterest and Reddit, as a social media platform that had clearly been designed by founders living in the 2000s era of early Web 2.0 – when blogs and links were still a thing.

I think Twitter was to people who love reading and writing as TikTok is to most people today: an easy way to mindlessly scroll through content. The beauty of Twitter though was that links were a native part of the format. While over the years people invented threads as a way to post longer chains of thought on Twitter, microblogging has always been incompatible with deeper written work. Despite Elon Musk’s best efforts, nobody goes to Twitter for longreads even today, long after he removed the technical barriers preventing users from making Twitter posts in the thousands of words. A feed of microblogged thoughts inevitably leads to a need for links to more polished, deeper exploration of the pithy posts casually tossed off into the ether.

Some may remember the era when blogs had comments. The comments section died out as a forum in the late 2000s because without extremely active moderation, most comments sections would dwindle into a cesspool of spam and trolling. Twitter helped replace individually-hosted comments sections: if you saw something interesting, you could post a link to it on Twitter and start a conversation about it. If you found something interesting linked on Twitter, you could read the link and then start a dialogue about it with the person who’d shared it on Twitter.

Twitter had a symbiotic relationship with the remaining legacy bits of the web: the few remaining blogs, and of course the print media. For a few years, Twitter was how I would find new things to read; even when there weren’t links, there were interesting discussions to read on the platform, courtesy of the people I’d chosen to follow. Unlike most other platforms, Twitter for a long time primarily invested in its Following feed, and that feed was the default for most heavy users of the product. There was no direct algorithmic intervention controlling what you saw.

All of this changed when Elon Musk bought Twitter. Musk forced a lot of investment in the algorithmic For You feed, which I do think now is genuinely a better experience for browsing Twitter in its current incarnation – assuming you’ve successfully blocked the trolls and Nazis that Musk has promoted into the For You feed. Concurrent with this, Musk heavily penalized all tweets with links, with the explicit intention of forcing users to stay within the Twitter walled garden. 

There is no escaping that Musk as the private owner of the Twitter platform now has a lot of heavy influence over what all Twitter users see by default. The best a user can do is minimize his influence by blocking the users he promotes, a tedious process. And none of this will bring back what Twitter used to be – the community of writers and academics that previously invested heavily in their Twitter presence because of its driving traffic to their platforms has been hollowed out, with many migrating to other platforms like Bluesky and Threads.

To me, what Musk did to Twitter is the worst case scenario for what the modern internet has become. The future I fear for the internet is similar to what China’s internet has become: a handful of walled gardens. All we need are a handful of apps: between WeChat (what Musk wants his X to be), Xiaohongshu (Instagram/Pinterest), Douyin (TikTok) and a handful of more utilitarian apps for banking and productivity, the average internet user doesn’t actually need much more.

When young people today are looking for a restaurant recommendation or planning a trip, they’re as likely – if not more likely – to start by searching on TikTok rather than Google. To a young person, TikTok might as well be the internet.

It might feel like this just a natural replacement of the mainly-text Google search algorithm with a mainly-video TikTok feed and search algorithms. But I think it matters that Google crawled the open web, while TikTok crawls its proprietary videos. While challenging Google’s effective monopoly over web search is difficult, there have always been alternative search engines for those who wanted to see what a web crawled and ranked by someone else would look like. With TikTok, there is no alternative. 

In theory, a different media platform like Instagram is an alternative – but the fact is that the content within Instagram’s and TikTok’s platforms are totally walled off from another. If you want to find something on TikTok, you are at the mercy of the tools and the algorithms Bytedance chooses to make available to you; you can’t search TikTok’s content using Instagram’s search ranking algorithms. You’ll never stumble on something on TikTok without the platform allowing you to. And that’s a big difference between the modern internet and the open web. If the only microblogging service you use is Twitter, you’ll only see what Elon Musk wants you to be allowed to see there.

I don’t blame people for turning away from traditional search or the legacy web. The highest-quality media sites are largely gated behind paywalls these days. Most traditional search engine results are festooned with ads, and overrun by websites hosting slop that was either written by an LLM, or a content mill that will soon be replaced by LLMs. A very reasonable theory for why young people have turned away from the legacy web is that it has as much to do with the unreliability of the modern search engine and print media experience as it has to do with the appeal of videos and images. Half the reason some people prefer to use LLMs like ChatGPT or Claude for basic search queries could easily be that you get a direct answer with no ads, no paywalls, and no obvious slop results.

I hope the modern internet we have today is not the final evolution of what the web started over three decades ago. I don’t know that we’ll ever get back an interconnected web of links. But I do want the spirit of discovery and exploration and good faith dialogue that characterized so much of the early internet I grew up with to survive. I acknowledge it’s hard to be optimistic, seeing internet trolls and Nazis from the dark corners of 4chan and Twitter promoted into prominence not just in one walled garden but in the US government.

One bright shoot I think is how platforms like Bluesky and Threads have opened feed selection as a choice for their users. The modern internet is built around the feed, and for most of us most of the time we have to live with the defaults that the corporate overlords who run the media platforms we all use give us. But it’s nice to have options.

Yet overall I find it difficult to be optimistic about the future of the internet. The technology is here to stay, but the original ethos of the open web does not seem compatible with where we are taking it. One bit of hope I cling to is reflecting on Francis Fukuyama’s famous prediction about the seeming end of history. In 1989, the same year the web was invented and as the Cold War came to a close, Fukuyama penned a famous essay titled The End of History – by which he meant that after the collapse of communism, there would be no further geopolitical conflict about the ideal political system, as everybody would agree on some form of liberal democracy as the ideal government. Fukuyama ended his 1989 essay musing: “Perhaps this very prospect of centuries of boredom at the end of history will serve to get history started once again.”

My interpretation of the 2024 US presidential election is that in one of the world’s oldest democracies, a slight plurality of voters decided they want history to start again; that 49.8% of American voters were bored of liberal democracy. Geopolitically it’s hard to think of a scarier time since the end of the Cold War, but also certainly hard to think of a less boring time either, which is what Americans voted for.

The reason I cling to hope that the internet won’t collapse into a never-ending feed of AI slop in walled gardens is that I want to believe the same human spirit that would revolt against the prospect of centuries of boredom under liberal democracy is one that will also eventually tire of the walled gardens of the modern internet. If the American voter tired of liberal democracy just three decades after the end of the Cold War, perhaps we are closer than we think to a revolt of internet users against a handful of publishers feeding them an infinite scroll of slop.

Perhaps this is being too optimistic, but just as some Americans evidently harbor nostalgia for being ruled by a king, I expect there will be young internet users who eventually idealize the old open web. Just as my generation idealized vinyl record players, there will be people who idealize surfing the web on 56k modems. Somewhere out there the Gen Alpha hipsters who’ll regret they were born too late to write a blog are taking their first steps.

As long as those of us who were there for it can keep the memory of an open, decentralized web of ideas and individuals alive, there’ll always be a positive ethos for new generations of internet users and innovators to aspire to.