How the internet's curation models evolved — and why it matters

Executive overview

The internet's content economy is built on distributed curation: picking what users see from a vast pool of user-generated content. Three successive models — links, networks, and loops — each solved the curation problem differently, with escalating effectiveness and escalating side effects.

Each jump in curation efficiency stripped out more human judgment, producing more engagement and more damage.

Links, networks, and loops: three curation models

1. Link model (blogosphere era): Trust-based hyperlinks between sites. Readers entered via pre-existing trusted sources; new sites earned trust slowly through repeated linking. Effective at filtering misinformation — content without credibility could not enter the trust web.
2. Network model (Facebook, Twitter): Homogenised interfaces + social graph. Facebook used declared friendships to curate feeds; Twitter added the retweet, enabling viral spread across the network at speed. Lowered the barrier to entry for creators and consumers alike.
3. Loop model (TikTok): Machine-learning replaces all human judgment. Content is embedded in a statistical space; watch-time signals weight future selections. No social graph, no explicit sharing — just a feedback loop tuned directly to the individual.

Advantages and disadvantages of each model

• Link model filtered out conspiratorial content naturally; hard to monetise, hard to break into as a creator
• Network model made content creation and consumption accessible to all; viral dynamics produced tribalism, outrage, and media self-censorship
• Loop model is the most effective curation ever built; removes any community or informational dimension — pure engagement stimulus

Why fixing these models from the inside doesn't work

• Moderation boards cannot counter the cybernetic dynamics of retweet virality or ML loops — they operate on different logics
• The only real lever is cultural: convincing people to spend less time on these platforms
• A revival of link-based, trust-grounded spaces (e.g. newsletters, podcasts) could partially restore what was lost

Multi-scale planning for long projects

• Quarterly/semester plan: set the project, milestones, and time heuristics (e.g. protect Tuesday and Thursday mornings)
• Weekly plan: translate quarterly intentions into actual blocked time for that week
• Daily time-block plan: execute against the weekly plan
• Task management systems (e.g. Trello) layer on top when a project generates discrete tasks or waiting-on items; uniform projects (like writing a book) don't need this

Managing email to eliminate inbox stress

• The real problem is not how often you check email — it is the volume of messages that arrive unscheduled and require a response
• Redesign work processes to minimise those messages; batching alone does not solve the underlying issue
• Email is suited for delivering information, files, or single-reply questions — not back-and-forth conversation
• Route ongoing discussions to office hours, docket-clearing meetings, or hallway conversations instead

Building career capital before pushing back

• Early-career workers should deliver fast and at high quality — reliability is the fastest path to leverage
• Career capital creates the autonomy to dictate terms later; demanding it before earning it fails
• Know where you want to go before you accumulate capital, so you can deploy it intentionally

Lifestyle-centric career planning: four lessons

1. What it is: work backwards from a concrete image of your ideal lifestyle — physical environment, social environment, pace, mental and spiritual life, time outside work — before deciding on career moves
2. How to find the vision: expose yourself to documentaries, profiles, biographies, and people you know; trust your gut on specific concrete details, not on abstract job categories
3. Case study — two visions for a data engineer:
   • Vision A (meadow, autonomy, writing): pursue bespoke freelance-ready skills, location independence, lower cost of living, part-time project model
   • Vision B (city, startup energy, leadership): pursue team leadership roles, move to high-growth companies, build investor and founder networks
4. Why it works: daily lived experience — environment, pace, relationships — is what directly produces wellbeing; career decisions made in isolation and hoping life adjusts is "rolling the dice"