Input curation

You didn't choose most of what you read today

Think about every piece of information you consumed in the last 24 hours. The news headlines that appeared on your lock screen. The tweets and posts that populated your feed. The emails from newsletters you vaguely remember subscribing to. The notifications from apps competing for your attention. The YouTube sidebar recommendations that pulled you from what you intended to watch into what the algorithm decided you'd click on.

Now ask: how much of that did you deliberately select?

For most people, the honest answer is almost none. The vast majority of information you consume on any given day was chosen for you — by algorithms optimizing for engagement, by default notification settings you never adjusted, by subscriptions that accumulated over years of casual sign-ups, by social norms that make you feel obligated to stay current on things that have no bearing on your actual life.

In the previous lesson, you learned that information flows through a five-stage pipeline: input, processing, storage, retrieval, and output. The pipeline model revealed that most people's systems break at specific, identifiable stages. This lesson is about the first stage — the one that determines everything downstream. If your inputs are noisy, unfocused, and unchosen, no amount of sophisticated processing can rescue them. Curation is not a nice-to-have at the front of your pipeline. It is the foundation that every other stage depends on.

The information environment has changed. Your habits haven't.

For most of human history, information was scarce. The constraint was access. If you wanted to learn something, you had to find a book, locate an expert, or travel to where the knowledge was. The challenge was getting enough signal, not filtering out noise.

That constraint inverted within a single generation. Herbert Simon, the Nobel laureate in economics and pioneer of decision-making research, identified the consequence as early as 1971: "A wealth of information creates a poverty of attention." When information is abundant, the scarce resource is no longer the information itself but the cognitive capacity to process it. Every additional input source competes for the same finite pool of attention, and attention that goes to low-value inputs is attention permanently unavailable for high-value ones.

The numbers are staggering. By some estimates, the average person encounters between 4,000 and 10,000 pieces of information daily — advertisements, social media posts, headlines, notifications, emails, messages. The human brain, however, has not upgraded its processing capacity since the Paleolithic. You have roughly the same working memory, the same attentional bandwidth, the same decision fatigue thresholds as your ancestors who needed to track the movements of a few dozen animals and the social dynamics of a tribe of 150 people.

The mismatch is not subtle. You are running hunter-gatherer cognitive hardware in an environment engineered by some of the world's most well-funded companies to capture and hold your attention. The information environment changed exponentially. Your information habits, if you haven't deliberately redesigned them, are still set to their defaults — and those defaults were designed by someone else, for someone else's benefit.

The information diet: a useful metaphor with real teeth

In 2012, Clay Johnson published The Information Diet, drawing an explicit parallel between information consumption and food consumption. The analogy is imperfect but instructive.

Just as the industrialization of food production created an environment where the cheapest, most accessible food is engineered to be maximally palatable but nutritionally empty, the industrialization of information production has created an environment where the cheapest, most accessible content is engineered to be maximally engaging but intellectually empty. Outrage generates clicks the way sugar generates cravings — not because it is good for you, but because your brain's reward circuitry cannot distinguish between stimulation and nutrition.

Johnson's core argument is that "information obesity" is a real condition with real consequences: impaired decision-making, chronic distraction, shallow thinking, and a persistent feeling of being overwhelmed despite consuming more information than any previous generation. The solution, he argues, is not to go on an information fast but to become a deliberate consumer — to treat every information source the way a nutritionist treats every ingredient: what does this contribute, and what does it cost?

The food metaphor reveals something important about input curation. Nobody argues that you should stop eating. The question is always what you eat, how much, and whether you chose it or whether it was pushed on you by the environment. The same is true for information. The goal of curation is not less information. It is better information, deliberately selected, consumed in quantities your cognitive system can actually process.

The filter bubble problem: why algorithms are not curators

If your information is being selected for you, you might argue that you already have curation — algorithmic curation. Every social media feed, every news aggregator, every recommendation engine is a curation system. The problem is whose interests that curation serves.

Eli Pariser coined the term "filter bubble" in 2011 to describe the invisible, personalized universe of information that algorithms construct around each user. Google shows you different search results than it shows your neighbor. Facebook surfaces different stories. YouTube recommends different videos. Each platform learns what you engage with and delivers more of it — creating a feedback loop where your existing preferences, biases, and emotional triggers are reinforced rather than challenged.

Pariser's insight was not that algorithmic filtering is inherently bad. Some filtering is necessary — you cannot process the entire internet. The problem is that algorithmic curation optimizes for engagement, not for your epistemic growth. An algorithm that shows you content that makes you angry, anxious, or outraged is succeeding at its job — you are engaged. But you are not better informed, you are not thinking more clearly, and you are not making better decisions. You are simply more activated.

This is the critical distinction: algorithmic curation serves the platform. Deliberate curation serves you. When an algorithm selects your inputs, the optimization target is your attention — how long you stay, how often you return, how many ads you see. When you select your own inputs, the optimization target is your cognition — how clearly you think, how well you decide, how effectively you act.

Relinquishing input curation to algorithms is like letting a fast-food company design your diet. The company's incentive is to make you eat more. Your incentive is to eat well. These objectives are fundamentally misaligned.

Signal-to-noise ratio: the metric that matters

Engineers use a concept called signal-to-noise ratio (SNR) to describe the proportion of useful information in a transmission relative to the irrelevant background. A high SNR means the signal is clear. A low SNR means it is buried in noise.

Your information pipeline has an SNR, and for most people it is astonishingly low. Consider a typical Twitter feed: for every genuinely insightful thread that changes how you think about a problem, there are hundreds of hot takes, performative arguments, engagement bait, and context-free fragments. The signal exists, but the noise is overwhelming. The cost of extracting that signal — scrolling, filtering, resisting clickbait, recovering your focus — often exceeds the value of the signal itself.

Nicholas Carr made a related argument in The Shallows (2010), drawing on neuroscience to show that the internet's architecture — hyperlinks, notifications, infinite scroll, auto-playing media — actively trains your brain toward shallow, skimming attention and away from deep, sustained focus. Every low-quality input source does not just waste the time you spend consuming it. It degrades your capacity to process the high-quality inputs that follow. Noise is not neutral. It is corrosive.

Input curation, framed as an SNR problem, becomes straightforward: for every information source in your life, ask whether it raises or lowers your overall signal-to-noise ratio. A carefully written newsletter from a domain expert you trust raises it. A social media feed tuned by an engagement algorithm lowers it. A curated RSS feed of primary sources raises it. A 24-hour news channel running the same five stories on loop lowers it. The math is simple even if the execution requires discipline.

The deliberate input stack

Cal Newport, in Digital Minimalism (2019), argues that the solution to information overload is not moderation but intentional rebuilding from the ground up. His protocol is radical: a 30-day "digital declutter" where you remove all optional technologies from your life, then reintroduce them one at a time, each earning its place through demonstrated value. The process forces a shift from the default — everything is allowed unless you explicitly remove it — to the deliberate — nothing is allowed unless you explicitly choose it.

Tim Ferriss proposed a similar concept in The 4-Hour Workweek (2007), which he called the "low-information diet." Ferriss's version is more extreme: stop reading news entirely, ask friends to tell you about anything truly important, and consume only information directly relevant to the task you are working on right now. The principle behind both approaches is the same: your default information environment is not designed for your benefit, and repairing it requires starting from zero, not trimming from the edges.

Here is what a deliberately curated input stack looks like in practice:

Tier 1 — Primary sources. These are the information inputs most directly relevant to your active work, projects, and decisions. If you are building a company, this includes your industry's primary data sources, your customers' communications, and the small number of thinkers who consistently produce insight about your specific domain. Five to eight sources, consulted daily or as needed.

Tier 2 — Growth sources. These are inputs that expand your thinking beyond your immediate domain — adjacent fields, contrarian perspectives, long-form writing that challenges your current models. Three to five sources, consulted weekly. These prevent your information diet from becoming an echo chamber while keeping consumption manageable.

Tier 3 — Ambient sources. These are low-frequency, high-quality inputs that keep you connected to the broader world: a curated news digest (not a 24-hour feed), a monthly magazine, a quarterly journal. One to three sources, consumed on a set schedule. These prevent the "monk problem" — becoming so isolated in your curation that you miss genuinely important developments.

Tier 4 — Serendipity sources. One or two sources that you keep specifically because they are unpredictable — a friend who sends interesting things, a bookstore you browse without agenda, a subreddit or community that consistently surfaces things you would not have found on your own. Curation does not mean eliminating surprise. It means ensuring that surprise is a feature of your system, not the only feature.

Everything that does not fit into these tiers is overflow. It can exist, but it does not get your primary attention, and it does not get to interrupt you.

The subscription audit: practical methodology

The most immediate action you can take to curate your inputs is an audit of your existing subscriptions — every newsletter, podcast, YouTube channel, social media follow, app notification, Slack workspace, and group chat that currently delivers information to your attention.

The audit works in three passes:

Pass 1 — Catalog everything. Open your email and search for "unsubscribe" to surface every newsletter. Open each social platform and export or scroll through your follow list. Check your podcast app subscriptions. List every Slack workspace and group chat. Check your browser bookmarks. The goal is a comprehensive inventory. Most people are stunned by the length of this list — it is common to discover 50 to 100 recurring information sources that you barely remember subscribing to.

Pass 2 — Evaluate each source. For each item on the list, answer: When was the last time this source produced an insight I acted on? If the answer is "I can't remember" or "more than 30 days ago," it has failed the test. Be honest. The question is not "is this interesting?" — interesting is cheap and infinite. The question is "does this make me think better, decide better, or act better?"

Pass 3 — Decide: keep, probation, or cut. Sources that clearly earn their place stay. Sources that clearly do not get unsubscribed, unfollowed, or muted immediately. Sources in the middle get a 30-day probation — you keep them, but you pay attention to whether they deliver value. At the end of 30 days, they earn a permanent spot or they go.

This audit is not a one-time event. Your information environment changes, your needs change, and new sources appear. Schedule the audit quarterly. Fifteen minutes of pruning every three months prevents years of accumulated noise from degrading your pipeline.

The cost of every input

There is a hidden cost to every information source in your life that goes beyond the time spent consuming it. Every source carries at least three costs:

Attention cost. Every notification, every email, every unread item in a feed creates a small cognitive pull — what researchers call an "attention residue." Even if you don't open the notification, knowing it exists splits a fraction of your attention. Sophie Leroy's research on attention residue showed that when people switch between tasks, their performance on the new task suffers because cognitive resources remain anchored to the previous one. Every subscription is a background task your brain is partially tracking.

Decision cost. Every input that arrives demands a micro-decision: read or skip? Open or ignore? Respond or defer? These decisions feel trivial individually, but they compound. If you make 50 micro-decisions about incoming information before lunch, that is cognitive budget that did not go toward your actual work.

Opportunity cost. The most insidious cost. Every minute spent on a low-quality input is a minute not spent on a high-quality one. The newsletter that wastes five minutes of your morning is not just five minutes lost — it is five minutes that could have gone to reading something that genuinely changed how you think. Over a year, a single mediocre daily newsletter costs you more than 30 hours. What could you have read instead?

These costs are why input curation is not optional for anyone serious about information processing. You are paying for every source whether you realize it or not. Curation is the practice of ensuring you are paying for sources that pay you back.

Your Third Brain: AI as curation partner

AI systems introduce a new capability in input curation: the ability to pre-filter, summarize, and evaluate information sources at a scale no human can match.

Here is how this works in practice. You can configure an AI assistant to monitor a broad set of sources — news feeds, research databases, industry publications — and surface only the items that meet your explicitly defined criteria. Instead of subscribing to 30 newsletters and scanning each one, you subscribe to 30 newsletters, route them to your AI system, and receive a daily digest of the five items most relevant to your current projects and decision domains. The AI does the scanning. You do the thinking.

This is not a replacement for deliberate curation. It is a force multiplier. You still define the criteria: what topics matter, what quality threshold applies, what counts as relevant. The AI executes the filtering at machine speed. Your role shifts from scanner to evaluator — you review what the AI surfaces rather than wading through everything raw.

A second application: using AI to evaluate your existing input sources. Feed your AI system the last 30 days of a newsletter you are unsure about and ask it to identify the actionable insights. If the answer is zero or one, that is data for your subscription audit. The AI is not making the curation decision — you are. But it is giving you evidence that makes the decision clearer.

The constraint remains the same as in every Third Brain application: the AI amplifies your judgment, but it cannot replace it. An AI that curates your inputs based on engagement metrics will reproduce the same filter-bubble problem that algorithmic feeds create. An AI that curates based on criteria you defined in a moment of clarity — relevance to active projects, epistemic quality, diversity of perspective — becomes the most powerful curation tool available.

The bridge to processing

Input curation determines what enters your pipeline. But curated inputs still require decisions. Every article, every email, every piece of information that passes your curation filter arrives with a question attached: what do I do with this?

That question — act on it, store it, or discard it — is the processing stage of your information pipeline, and it is where we turn next. Curation ensures that you are only asking that question about information worth processing. Processing ensures that you answer the question systematically rather than letting items pile up in an undifferentiated heap.

You have built the gate. The next lesson teaches you what to do with everything that passes through it.