How AI shifted our relationship with knowledge from distraction to cognition.
The second screen was a story about divided attention.
We never solved it. We outgrew it.
Circa 2010: smartphones everywhere, broadcasters excited. The pitch was complementary engagement: watch the show, tweet the reaction, Google the cast, check the scores. A richer TV experience.
The reality: we just looked at our phones instead of the TV. Parallel distraction, not enrichment.
Attention is finite. Fragmentation was the feature, until it became the bug.
The second screen assumed live television would always anchor the experience. It didn't. The algorithm replaced appointment viewing. TikTok doesn't tolerate a divided screen: it is engineered for total, undivided attention.
Without a shared live moment to orbit, the second screen had nothing to orbit. The concept quietly dissolved into just being on your phone.
80% of knowledge workers report information overload. The average worker spends 9.3 hours a week just searching for information they already have.
Personal Knowledge Management emerged as the response: externalise your thinking. Build a store of notes, ideas, and connections outside your head so you can actually reason across them.
The tool stack arrived: Obsidian. Notion. Roam. Bear. The concept was older than the software. But this time the tools caught up with the idea.
Andrej Karpathy describes the second brain in its purest form: capture everything worth keeping. An ever-growing log. Obsidian as the vessel.
What separates this from a notes app: the graph of connections. The ability to query your own thinking. Not a filing cabinet. A mind you can interrogate.
Before LLMs, the second brain was a better filing cabinet. More searchable, more connected than a Word doc, but still passive. You had to read it, synthesise it, connect the dots yourself.
LLMs changed the contract. Now the system can reason across everything you've captured: find the contradiction between two projects, brief you in your own voice, surface the pattern you missed three months ago. The second brain stopped being a repository and became a thinking partner.
Notion has 100 million users. Obsidian has around 1.5 million. Both are knowledge tools. That gap is the second brain problem in a single number. Put another way: 98.5% of the potential market hasn't adopted a true second brain system. The infrastructure exists. The habit doesn't.
The Karpathy version (Obsidian plus a custom LLM pipeline) is nerd territory. The mainstream wants knowledge tools, not knowledge systems.
But second screen went mainstream. Broadcasters built for it. Twitter was built for it. Stories was built for it.
What is the mainstream version of the second brain? Who is building that?
Zettelkasten methodology. Atomic notes. PARA framework. Graph thinking. Tagging discipline. There are whole books, courses, and communities dedicated to getting this right.
But the deeper problem isn't the tools. The second screen required no behaviour change: people were already on their phones. The second brain demands an entirely new way of working. That's not a UX problem. That's a mindset shift. And mindset shifts don't go mainstream until someone makes them invisible.
The ambient products are closer. Claude Projects and ChatGPT Memory build the brain around you: no behaviour change required. The capture tools (Limitless, Mem.ai) still ask you to opt in.
Mine is Obsidian and Slack. The vault holds the strategy; the channels hold the cultural feed. Claude connects both. That's the system.
Passive capture. Active surfacing. Ambient, not intentional.
The mainstream second brain won't feel like a brain. It will feel like things working slightly better. You mention something: it is logged. You read an article: it connects to three things you already knew. You start a brief: the relevant context arrives before you ask.
"Second brain" is nerd vocabulary. Like RSS, or markdown. It signals the enthusiast class.
The most successful infrastructure disappears. Nobody says "I use HTTPS." Nobody calls their timeline "the algorithm." The second brain will win when it stops being a thing you set up and becomes something that just happens. No name. No friction. No moment of adoption. It arrives the way spell-check did: silently, then everywhere.
By 2030, the second brain will be as ubiquitous as the smartphone was for second screen.
Not because everyone learns Obsidian. Because AI agents are beginning to handle capture, organisation, and retrieval automatically. The setup cost is the only remaining barrier. It won't last long.
The second screen split your attention.
The second brain extends it.
mikelitman.me · hello@mikelitman.me
With thanks to Ollie Glass, whose conversation this morning sparked the whole thing.