Before there was a feed, you had to build one. I don't mean that metaphorically. I mean you literally had to construct, by hand, the stream of information that arrived on your screen each morning. There was no default. No "For You" page. No algorithm deciding what you should see first. You opened your browser and you saw whatever you'd chosen to see. If you hadn't chosen anything, you saw nothing.
I've been thinking about this a lot, partly because I'm building tools like CultureTerminal that try to recreate some of that intentionality, and partly because my kid is growing up in a world where the feed is just there, fully formed, waiting. They'll never know what it felt like to assemble one from scratch. And I think that matters more than most people realise.
The assembly required era
My first proper internet setup, around 2004-2005, looked something like this. I had a Bloglines account -- later Google Reader -- where I subscribed to maybe fifty RSS feeds. Design blogs like Kottke, Swiss Miss, and Brand New. Advertising sites. A few music blogs. Some personal blogs written by people I'd never met but whose taste I trusted. Each morning, I'd open Reader and scroll through everything that had been published since I last looked. Chronological. Unfiltered. Every item got the same visual weight. A post from a one-person blog in Tokyo sat alongside a post from a major publication. The only hierarchy was time.
Then I had my Delicious bookmarks, organised by tags. Hundreds of tags, built up over years. "Branding." "Architecture." "Japanese-design." "Things-to-read-later." When I found something good on the internet, I'd bookmark it and tag it, and over time these tags became a personal taxonomy of everything I cared about. I could go to my "typography" tag and find three years' worth of articles, tools, and examples that I'd hand-selected. It was a library I'd built one link at a time.
And then there was my Twitter follow list. In the early days, this was the most important curatorial decision I made online. Every person I followed was a conscious choice. I'd find someone through a blog post, or through someone else's retweet, and I'd look through their last twenty tweets before deciding whether to follow. It sounds obsessive. It was. But the result was a timeline that felt like mine -- a roomful of people I'd personally invited, all talking at once about things I cared about.
The work was the point
Here's what people forget when they talk about how inconvenient the old internet was: the inconvenience was a feature. Finding a new blog to follow required effort. You had to stumble onto it, or follow a link from a blog you already read, or find it through someone's blogroll. That effort created investment. Because you'd worked to find your sources, you paid attention to them. You didn't skim. You read.
A blogroll -- for anyone under thirty who's never seen one -- was a list of links on the sidebar of a blog. Other blogs the author read and recommended. It was the most primitive form of curation possible: just a list of names with links. And it was enormously powerful. If I trusted a writer's taste, I'd go through their entire blogroll and check every link. Sometimes I'd find one blog through another blog through another blog, three levels deep, ending up somewhere I never would have found through any search engine. The chain of human recommendation was the discovery mechanism.
The intentional information diet
What I'm describing is something I've started calling an intentional information diet. Not in a self-helpy, digital-wellness way. In a practical, structural way. Before algorithmic feeds, your information intake was the direct result of choices you'd made. You chose your RSS subscriptions. You chose your Twitter follows. You chose your bookmarking tags. You chose which blogs to check daily and which to check weekly. Every source in your information diet was there because you'd put it there.
Now think about what your information diet looks like today. Open Instagram, TikTok, even LinkedIn. How much of what you see is from sources you actively chose? Half? A quarter? On TikTok it's essentially zero -- the entire experience is algorithmically determined from the first scroll. You're not choosing your diet. Someone else's business model is choosing it for you, optimised not for your growth or your taste but for your continued attention.
The shift happened so gradually that most people didn't notice. Twitter introduced the algorithmic timeline in 2016. Facebook had been doing it for years before that. Instagram followed. Each time, the pitch was the same: we'll show you the best stuff first. What they meant was: we'll show you the most engaging stuff first. And "most engaging" is not the same as "most valuable" or "most interesting" or "most aligned with your actual taste." It's what makes you stay on the app. Those are very different optimisation targets.
What CultureTerminal is trying to do
I built CultureTerminal because I missed this. Not the specific tools -- RSS readers still exist, and I still use one -- but the philosophy behind them. The idea that your information feed should be something you construct deliberately, from sources you've evaluated and chosen, without an intermediary deciding what's important.
CultureTerminal pulls from RSS feeds. Real ones. From publications and blogs that I've selected because they consistently produce interesting things. There's no algorithm sorting them. No engagement score determining what rises to the top. Things appear in the order they were published, the same way Google Reader showed them in 2009. If you miss something, you miss it. There's no "top stories" rescue net. The experience demands a bit of your attention, and that demand is the whole point.
It's not a nostalgia project, though. I'm not trying to rebuild 2009. I'm trying to take the principle that worked -- human-curated, chronological, intentional -- and see if it still holds up when the entire internet has moved in the opposite direction. I think it does. Not for everyone. But for the kind of person who used to maintain a blogroll and tag their Delicious bookmarks, there's still an appetite for this. For information that arrives because you chose it, not because a machine decided you'd engage with it.
The magazine rack at WHSmith
I keep coming back to a physical analogy. When I was a teenager in Nottingham, I'd go to WHSmith and stand at the magazine rack for ages. The Face, i-D, Dazed and Confused, Wallpaper, Creative Review. I couldn't afford them all, so I'd pick one. That choice -- which magazine to buy this month -- was an act of taste. I was deciding what information, what aesthetic, what perspective to spend time with. The constraint (money) forced intentionality. I couldn't have everything, so I had to choose what mattered most.
The internet removed that constraint. Suddenly everything was free, everything was available, everything was competing for the same screen. And into that overwhelming abundance stepped the algorithm, saying: don't worry about choosing. I'll choose for you. And we were so relieved by the offer that we didn't think about what we were giving up.
We gave up the choosing. And the choosing was the thing that built our taste in the first place.
Building it back
I'm not suggesting everyone go back to RSS readers and blogrolls. That ship has sailed for most people. But I do think there's something worth recovering from that era. The principle that your information diet is yours to design. That the sources you choose to pay attention to are a reflection of who you are and who you want to become. That the effort of finding and selecting those sources isn't a bug to be automated away -- it's a fundamental part of developing taste.
Every product I'm building right now is informed by this idea. Trove is a personal collection tool where you choose what to save and the patterns emerge from your choices, not from an algorithm's guesses. CultureTerminal is a curated feed built on RSS, not engagement metrics. Even the pub guide is, at its core, a hand-picked list -- my taste, my choices, presented without ranking or recommendation.
The feed before the feed was messy, effortful, and entirely your own. The feed after the feed is smooth, effortless, and entirely someone else's. I know which one I prefer. And I suspect, if you've read this far, you do too.