From information overload to personal intelligence

What is personal intelligence?

Personal intelligence is the practice of turning too many sources into a small, sourced, personal understanding of what changed. A personal intelligence system reads the sources you choose, keeps only the developments that matter to your questions, and tells you about them in a short account where every claim links back to its origin.

That definition is meant to be strict. It excludes a lot of software that people use to cope with information overload, and it explains why most of that software has not made the overload go away.

Why does information overload need a new category, not a better tool?

Information overload needs a new category because it is a symptom of a missing job. Most people already own tools for storing information, delivering it and retrieving it. None of those tools is responsible for deciding, every day, what changed in your subjects and whether it matters to you.

So the job falls to the reader. A product manager tracking three competitors, an investor holding a dozen positions or a researcher following one fast field ends up doing the same unpaid work: open everything, recognize repeats, guess what is significant, remember what was already known. The overload is the felt cost of doing that job by hand, across more sources than a person can read.

Adding a better reader app, a smarter inbox or a faster search engine improves one of the adjacent jobs. It leaves the missing one exactly where it was. A category earns its name when it describes a job nothing else does, and that is the case for personal intelligence.

How is personal intelligence different from PKM, news reading and search?

Personal intelligence differs from its neighbors in the question it answers. Each adjacent practice is useful; each answers a different question, at a different moment, with a different unit of output.

• Information management answers "where is it?" It files, tags and syncs documents and messages. Its unit is the file, and it acts when you put something in.
• Personal knowledge management (PKM) answers "what do I know?" Note systems and second-brain methods organize material you already read and chose to keep. Its unit is the note, and it acts after your attention has been spent.
• News reading answers "what did editors think was important?" A newspaper, an app or a newsletter selects for an audience. Its unit is the story, and the selection is someone else's.
• Search answers "what is the answer to the question I just asked?" Its unit is the result, and it acts only when you remember to ask.
• Personal intelligence answers "what changed in my subjects, and why does it matter to me?" Its unit is the development, and it acts continuously, whether or not you asked today.

Two features set the last item apart. It is proactive: it watches while you do other work. And it is judged by one person's mandate, not by an audience, a ranking or a query typed in the moment.

That is also why personal intelligence complements the others rather than replacing them. A brief that reports a development is worth filing in a knowledge base. A development that raises a new question sends you to search. The category fills the gap between them.

What can individuals borrow from professional intelligence work?

Individuals can borrow the shape of the work. Governments and large companies have run intelligence functions for decades, and the core method is simple enough to state on one page. The CIA describes it as a five-step intelligence cycle: planning and direction, collection, processing, analysis and production, and dissemination.

Three features of that cycle matter for personal use.

1. It starts with a question. Planning and direction comes first: what does the person who commissioned the work need to know? Collection without a question is hoarding.
2. Analysis asks forward-looking questions. The CIA's own description of the analysis step is to assess what is happening, why it is happening and what might occur next. That is a brief, not a list of links.
3. It ends with the person who asked. The finished analysis goes back to the same policymaker who started the cycle, and that reader's follow-up questions start the next round.

The President's Daily Brief, written each morning for a single reader, is the best-known product of this cycle. It is personal intelligence at the highest stakes: one reader, a defined mandate, finished judgments instead of raw cables.

What kept this model out of individual hands was cost. The cycle needs people to read everything, compare it with what was known, and write it up. Most professionals have never had an analyst. The interesting change of recent years is that the reading and drafting steps, the expensive ones, can now be done by software, while the question at the start and the judgment at the end stay with the reader.

What must a personal intelligence system do?

A personal intelligence system must perform five functions, and each one can be tested. A system that fails any of the tests turns back into one of the neighbors: a feed, an archive or a search box.

Consider a hypothetical reader, an investor following sodium-ion batteries as an alternative to lithium-ion for grid storage, to see what each test means in practice.

1. Source layer: does it cover where changes first appear? For this reader, that means cell makers' announcements, patent filings, utility procurement notices, a few research groups and the engineers who write about manufacturing yields. A system that reads only general news fails the coverage test, because those changes reach the press late or never.
2. Relevance filter: is selection tied to the reader's mandate? A new lithium price forecast matters to this thesis; a celebrity's electric car does not, even though both mention batteries. Keyword matching cannot make that distinction. A filter judged against the reader's stated questions can.
3. Deduplication: does one event produce one entry? A plant announcement will appear in a press release, a trade publication, a dozen posts and a newsletter. The test is whether the reader sees one development with its best sources attached, not fifteen items to reconcile.
4. The brief: does each entry say what changed, why it matters and what to watch, with sources? "Manufacturer opens line" is an item. "Manufacturer opens its first line at stated volume; it is the first supplier of that size in its region; watch the first utility contract" is intelligence, provided every clause links to where it came from.
5. Memory: does the next brief know what the last one said? When the utility contract is signed three months later, the system should report it as the sequel to the plant opening, not as news from nowhere. Without memory, every brief starts from zero and the reader supplies the history.

A sixth property follows from the five rather than adding to them: the system must be able to say nothing. If a week passes with no development that bears on the thesis, the honest output is an empty one.

Several products now attempt this combination. Kindal is one of them: the reader chooses the sources, and it writes a short brief only when something changes, with every line linked to the post, filing or video behind it. Whatever tool is used, the five tests are the fair way to judge it.

What is personal intelligence not?

Personal intelligence is not omniscience. A system can only report what its sources published, so a weak source list produces confident briefs about a partial picture. Reviewing the sources is part of the practice, not a setup step done once.

It is also not a verdict machine. A brief reports what a filing, a company or an expert said and attributes it; the reader weighs it. The value is in selection and context, not in telling the reader what is true.

And it is not a rebranded news digest. A digest is curated by an editor for many readers at once, so its relevance is an average. Personal intelligence is curated against one mandate, so two readers following the same company can rightly receive different briefs on the same day. A supplier and a short seller care about different paragraphs of the same quarterly report.

Finally, it does not remove judgment. The cycle starts and ends with the reader. Software can absorb the hours of reading and comparing; it cannot decide what you should be asking or what you should do with the answer.

What changes when you treat overload as an intelligence problem?

When overload is treated as an intelligence problem, the question changes from "how do I read all of this?" to "what do I need to know, and who is watching for it?" The first question has no good answer. The second one does.

The reader stops measuring the day by items processed and starts measuring it by developments understood. Sources are chosen for where change first appears, not for volume. Briefs accumulate into a record of how a subject moved, which makes each new development easier to read than the last.

Information overload described a world with more sources than readers. Personal intelligence describes what a reader needs in that world: a small, sourced account of what changed, written for one person, on the days when something did.