I remember the exact moment Twitter died for me. Not the day Elon bought it. Not the rebrand to X. It was earlier than that - the day the timeline stopped being chronological. The day the feed became a guessing game, showing me things an algorithm decided I should see instead of the things the people I chose to follow actually said. That was the beginning of the end. Not just for Twitter, but for the entire era of algorithmic social media.
Between 2010 and 2015, Twitter was the greatest real-time information tool ever built. It wasn't perfect, but it was honest. You followed people. They posted things. You saw those things, in order. The simplicity was the genius. If your timeline was good, it was because you had curated it well - chosen the right people, muted the noise, built something that felt like your own personal newsroom. The quality of the feed was a direct reflection of the quality of your choices.
Then the algorithms arrived. And they broke everything.
What algorithms optimise for
Every algorithmic feed optimises for the same thing: engagement. Not quality. Not relevance. Not your wellbeing or intellectual growth. Engagement. Which means the algorithm has a very specific incentive: show you things that make you react. Outrage works. Fear works. Drama works. Calm, thoughtful, nuanced content doesn't work because calm, thoughtful, nuanced content doesn't generate clicks, replies, and shares at the rate the algorithm demands.
The result is that everyone's feed gradually converges on the same emotional register - high arousal, low substance. It doesn't matter whether you follow academics or comedians, activists or artists. The algorithm will find the most reactive version of each of them and push that to the top. Your carefully curated follow list becomes irrelevant. The machine decides what you see, and the machine has terrible taste.
I built CultureTerminal partly as a response to this. I wanted a news feed that reflected editorial judgment, not engagement metrics. One that surfaces stories because they matter to the intersection of culture, brands, and design - not because they generate the most clicks. It uses technology to aggregate and organise, but the editorial framework behind it's deliberately opinionated. It has a point of view. Algorithms don't have points of view. They have objectives.
The RSS renaissance
Something interesting is happening. Quietly, without fanfare, people are rediscovering RSS. The technology that everyone declared dead a decade ago - when Google killed Reader and social media promised to replace it - is making a comeback. Not as a mass-market product. As a quiet rebellion.
RSS is the purest form of the chronological feed. You subscribe to sources. Those sources publish things. You see those things, in order, with nothing filtered out and nothing algorithmically boosted. There's no engagement optimisation. No virality mechanics. No recommendation engine deciding that what your favourite writer published yesterday is less important than a viral argument you have no stake in.
The people returning to RSS aren't luddites. They're the most digitally literate people on the internet - developers, writers, designers, people who understand exactly how algorithmic feeds work and have decided they want something different. They want control. They want curation they perform themselves, not curation performed on them by a machine with misaligned incentives.
I find this deeply encouraging. Not because RSS will ever be mainstream again - it probably won't. But because it represents a philosophical shift. A growing number of people are rejecting the idea that a machine should decide what they pay attention to. They want human judgment back in the loop. They want the feed to be theirs again.
Nuzzel and what we lost
There was a product called Nuzzel that understood this perfectly. Nuzzel watched your Twitter network and surfaced the links your follows were sharing most. Not the algorithm's picks. Your network's picks. It was social curation at its best - human taste, aggregated intelligently. If three people you respected all shared the same article, Nuzzel would surface it. The signal came from the people you trusted, not from an engagement algorithm.
Nuzzel was acquired and shut down, because of course it was. The best curation tools always get killed. But the idea behind it - your network is the algorithm - is more relevant now than ever. It's the idea behind Curio, which I built to do something similar: let the people you trust surface what matters, rather than surrendering that power to a machine.
Chronological is a feature, not a limitation
The tech industry treated chronological feeds as a problem to solve. Users were "missing" content. Important posts were getting "buried." The solution, naturally, was an algorithm that would surface the "best" content for each user. This framing was always dishonest. The real problem wasn't that users were missing content. The real problem was that platforms needed users to spend more time scrolling, and algorithmic feeds - by mixing in viral content, recommended posts, and suggested follows - achieved that goal brilliantly.
But chronological isn't a limitation. It is a feature. It imposes a natural constraint on consumption. You open the app, you read what's new since you last checked, you close the app. There's no infinite well of "recommended for you" content to keep pulling you back. The finiteness is the point. It mirrors how human attention actually works - in bursts, not in bottomless scrolls.
Print magazines understood this instinctively. A magazine has a fixed number of pages. You read it, you finish it, you put it down. The editor made choices about what to include, and those choices created a coherent experience. Nobody ever complained that a magazine didn't show them everything. The constraint was the craft. The selection was the value.
Building the feed you want
I don't think the solution is to abandon social media entirely. I still check X first thing every morning, out of habit as much as anything. The real-time pulse of the internet still lives there, somewhere, underneath the noise. But I've learned to treat algorithmic feeds as one input among many, rather than as my primary information source.
The stack I've built for myself looks something like this: RSS for long-form writing and industry news. Newsletters from specific writers I trust. Slack channels for professional communities. X for real-time cultural moments, treated with scepticism. And my own products - CultureTerminal, Curio - to filter and organise what matters within my specific interests.
This takes more effort than letting an algorithm decide. Of course it does. Curation always takes more effort than automation. But the quality of information I consume is incomparably better. I read fewer things, but better things. I discover perspectives that no algorithm would ever surface, because they aren't optimised for engagement. I feel less overwhelmed and more informed. The feed is mine again.
The algorithmic feed isn't going to die overnight. It still has billions of users and trillions of dollars of advertising revenue behind it. But the cracks are showing. The best thinkers, the best writers, the best curators are increasingly building their own channels - newsletters, RSS, small communities - outside the algorithmic machine. They're choosing quality over reach, depth over virality, curation over automation.
The chronological feed isn't a step backwards. It's a step forwards. Back to a time when what you saw was a reflection of what you chose, not what a machine chose for you. Back to a time when curation was a human act, performed with taste and judgment and the willingness to say: this matters, and this doesn't.
The algorithm had its moment. Curation is next.
