Why It's Important to Know Things

For most of human history, acquiring new information was expensive.

Books were scarce, experts were difficult to reach, and answering even a relatively simple question could require hours or days of research. Worse still, without mass communication technology, there would be no way to verify your conclusions by checking them against other people’s insights and information.

Today, we have the opposite problem.

Almost anyone can access enormous digital libraries, search billions of web pages, watch lectures from world-class experts, and, as of the last few years, ask an AI system to synthesize all of that information into a reasonably coherent answer in seconds.

Given that reality, it’s tempting to conclude that memorizing things is obsolete.

Why bother remembering dates, names, definitions, formulas, historical events, artistic movements, or anything else when you can simply look them up whenever you need to?

I understand why that line of reasoning seems plausible, but I’d argue the opposite is true: The easier it becomes to access information, the more important it becomes to maintain a substantial body of knowledge inside your own head.

I’m not saying that effective search engines are a bad thing or that AI is some civilizational mistake. In fact, I think they’re extraordinary tools, and they’re going to become dramatically more capable and more useful over time. But in an age where almost everyone will have access to powerful AI, advantages will increasingly accrue to the people who retain enough of their own knowledge to competently judge the answers they receive from those tools and to give them meaningful direction in the first place.

I’ll talk more about that at length in a moment, but there is also another reason to know things yourself that may initially seem like a contradiction:

No One Person Can Know Everything

By this, I do not mean that expertise isn’t real, or that it’s impossible to know enough about anything to make day-to-day decisions. What I mean is that any one person’s knowledge is inherently limited.

In his 1945 essay, The Use of Knowledge in Society (read it for yourself on the Lexandria platform), economist F.A. Hayek observed that the knowledge necessary to understand and navigate a complex society does not exist neatly assembled in one place. It is distributed among countless people, each with different experiences, specialties, interests, local information, observations, and perspectives.

And because we are each unique individuals who can only experience the world firsthand (i.e., we do not share a hive mind), our knowledge will necessarily be incomplete. No single person can possess all of it at once, and much of it will appear to conflict.

That means the objective of learning cannot be to make yourself intellectually “self-sufficient” — as that would be impossible.

Instead, the objective should be to know enough such that you can productively interact with the knowledge possessed by everyone else. Your own knowledge gives you the vocabulary to ask questions, the context to properly understand answers, and the ability to notice when somebody else says something that your model of the world cannot currently explain — either because they’re wrong about some aspect of their premises, facts, or conclusions; or because you are.

Paradoxically, knowing more yourself helps you develop epistemological humility and makes you better at learning across the board.

Without an internal knowledge-base to draw from, you’ll struggle to defend yourself against liars and scammers who selectively present only that information which supports their agenda. You’ll be unable to meaningfully participate in deeper thinking and human conversations about life and reality. You also won’t be effective at directing automated tools such as AI to do what you actually want.Simply put: You need to know things for yourself.

“Access” to Information Is Not the Same as Earned Knowledge

Psychologists have long known that easy access to external information changes how we remember things.

In 2011, researchers Betsy Sparrow, Jenny Liu, and Daniel Wegner published a series of experiments on what became popularly known as the “Google Effect.” They found that when people expected information to remain available later, they were less likely to remember the information itself and more likely to remember where they could find it.

This is part of a broader phenomenon known as cognitive offloading: using tools outside the brain to reduce the mental work we need to do internally.

We do this constantly. A grocery list is cognitive offloading. So is a calculator, a calendar reminder, GPS, notebooks, or spreadsheets.

There is nothing inherently wrong with any of this. Civilization itself depends on externalized memory. Writing allows knowledge to survive beyond one person’s lifetime. Libraries allow generations of people to build on discoveries made centuries earlier. Maps prevent every traveler from having to rediscover the landscape from scratch, and so on.

The problem begins when we confuse outsourcing information storage with outsourcing understanding.

Your phone can remember a fact for you, but it cannot make that fact part of your mental model of the world. And if your own model is weak enough, you have very little ability to recognize when the information being handed back to you is misleading, incomplete, or completely wrong.

LLMs can create the appearance that Hayek's dispersed knowledge has finally been concentrated in one place. You can ask one interface about physics, Roman history, camera lenses, monetary policy, Renaissance painting, or how to repair a dishwasher, and you’ll receive a confident answer to all of your questions.

But that confidence can be misleading.

The underlying human knowledge is still dispersed, contested, contextual, uneven in quality, and frequently contradictory. AI gives us an extraordinarily convenient way of navigating some of it. It does not, however, eliminate the original problem, which computer programmers have summarized as “GIGO,” or “Garbage In, Garbage Out.”

That distinction will matter more as AI becomes more persuasive.

AI Is Useful, But It’s Not a Truth Machine

We tend to treat search engines (both traditional and LLM) as neutral windows into reality, but that isn’t actually what they are.

Search systems try to identify information that appears relevant, useful, authoritative, new, popular, and also connected to other sources that seem similarly trustworthy. Those are all reasonable signals for truthfulness, but they are still proxies rather than truth itself.

That would be less concerning if false information behaved like obviously false information, but of course it often doesn’t.

A major 2018 study published in Science examined roughly 126,000 cascades of true and false stories shared by millions of Twitter users. The researchers found that false information tended to spread farther, faster, deeper, and more broadly than true information. Novelty and emotional reactions helped give false stories an advantage over more mundane truths.

There is also the well-established “illusory truth effect,” in which repetition alone can make a claim seem more believable. Familiarity makes statements easier for our brains to process, and that ease of processing can be mistaken for evidence that something is true. Some experiments have found this effect persists even when people possess knowledge that contradicts the false claim.

That last point is important because having knowledge does not make anyone invincible. A person can know the right answer and still be manipulated by repetition, tribal loyalty, wishful thinking, or simply by hearing a false claim from a persuasive presenter (recall that the term “conman” is a shortening of “confidence man”).

But having knowledge gives you something to compare a new claim against. It gives you a reason to hesitate when something doesn’t fit.

It also gives you the ability to compare your conclusions against other people who know different things than you do. Sometimes your gut check will be right, and the article, search result, or AI answer will be wrong. Other times your gut check will expose a hole in your own understanding, and someone with different expertise will be able to show you why. Both outcomes are potentially valuable.

The important thing is knowing enough real information to recognize that there is a disagreement worth investigating. If you don’t have that knowledge, you’re far more prone to believing things that aren’t true and being duped by persuasive writing.

In a speech he gave in 2002 called “Why Speculate?” writer Michael Crichton described a version of this problem as “Gell-Mann Amnesia,” named after a conversation with noted physicist Murray Gell-Mann:

“Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well. In Murray's case, physics. In mine, show business. You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. I call these the "wet streets cause rain" stories. Paper's full of them.

In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.”

It’s obviously not a scientific “law,” but the observation is painfully recognizable. Most of us have encountered a news story, documentary, viral post, or online argument about something we understand well and thought, “That’s not how any of this works!”

The problem is that when we lack any personal knowledge of the next subject, we lose the ability to perform the same gut check.

AI can make this problem even harder to overcome, because it is extraordinarily good at presentation.

A bad answer no longer looks like a rambling forum post written by somebody with poor grammar in ALL CAPS. It can arrive instantly as a polished, organized explanation with bullet points, seemingly relevant citations, written in the tone of an expert.

That doesn’t mean AI is uniquely deceptive, but it does mean that presentation quality is becoming less useful as a signal of underlying accuracy. A confident, beautifully written answer can still be wrong, and your own knowledge is one of the first tools available to warn you that something seems off.

And on a practical level, misinformation isn’t the only issue to worry about.

You also need to retain as much of your own knowledge as you can, because it’s going to be increasingly crucial in order to navigate the world of tomorrow.

The Human Role in Creativity Will Change, But It Won’t Disappear

As AI becomes better at executing tasks and more of our lives become automated, the human role increasingly shifts from technician to director — like how American factories used to require hundreds of individual workers physically welding and fastening parts together on an assembly line, but now largely employ human workers to sit at computers while overseeing the robots doing all that laborious physical work instead.

This is already happening in creative professions with AI.

Someone using generative tools can produce images, rough animations, music, written material, visual effects, storyboards, and other assets much faster than before. As those systems improve, the bottleneck becomes less about whether you can physically produce every element yourself and more about whether you know what the finished artwork is supposed to be.

That sounds alternatively liberating or terrifying to different people, depending on whether or not they are good at and enjoy doing the technical work themselves. But being the director is not necessarily any easier than being the technician.

Say you tell an AI image system that you want something to look “more cinematic.” What does that actually mean?

Do you want low-key lighting? Deep focus? Wide-angle distortion? A naturalistic Roger Deakins palette? Gordon Willis-style darkness? German Expressionist shadows? A symmetrical Wes Anderson frame? Michael Bay’s aggressively kinetic camera language?

The more you know about the medium, the more precisely you can describe what you want, and the more effectively you can discern whether the result is any good (i.e., how closely it matches your vision).

This is part of what expertise has always meant. Experts are not simply people who have memorized more isolated facts than everyone else. Their knowledge allows them to recognize patterns. Stored knowledge gives new information somewhere to go. It lets us organize details into larger conceptual units and make connections without stopping to research every component individually.

That kind of synthesis of different reference points is also a foundation of creativity.

Quentin Tarantino is an obvious example because his influences are unusually visible. His movies invariably combine Sergio Leone’s spaghetti westerns, Hong Kong action cinema, exploitation films, crime movies, pop music, Hollywood history, and countless other sources into a single cohesive production. The result is recognizable as “a Tarantino film” because he recombines those influences through his idiosyncratic sensibilities.

But every artist is doing some version of this.

George Lucas did not create Star Wars in a vacuum. Those films draw on Akira Kurosawa, old Flash Gordon serials, World War II aviation movies, westerns, mythological structures filtered through Joseph Campbell, Fritz Lang’s Metropolis, Kubrick’s 2001: A Space Odyssey, and many other influences.

Wes Anderson’s influences pull from older European cinema, theatrical staging, illustration, literature, mid-century design, and filmmakers such as François Truffaut and Ernst Lubitsch.

This applies to every artform.

Beethoven did not emerge independently from the musical world of Mozart and Haydn, and later composers such as Wagner or Chopin built on Beethoven’s chromaticisms and complex borrowed harmonies in their own work.

Jazz musicians build vocabularies by absorbing previous players. Rock guitarists learn from blues musicians, who learned from earlier folk traditions. Hip-hop is almost uniquely explicit about this process because sampling turns the recombination of existing musical language into part of the medium itself.

Visual artists, architects, designers, photographers, writers, animators, and game developers all do the same thing — synthesizing their favorite art and other influences into their own work. Originality does not mean having no influences. In many cases it means having so many influences, and understanding them deeply enough, that the final combination becomes difficult to trace back to any one source.

A generative AI model may have access to far more creative references than any individual person could memorize, but that doesn’t help much if the human operating it cannot identify which references are interesting, compatible, contradictory, or worth pursuing.

In other words, you cannot deliberately synthesize ideas that you cannot even recall or summon into your mind.

And remember, artistry is only one form of creativity. Scientific research, engineering, entrepreneurship, technological development, and business strategy all involve the same basic process of combining existing knowledge into new arrangements.

We’re All Standing on the Shoulders of Giants

A scientist needs enough familiarity with previous research to ask a worthwhile experimental question. An engineer needs enough knowledge of prior solutions to avoid unknowingly recreating a design that failed decades earlier. Entrepreneurs who understand nothing about markets, incentives, technology, regulation, or history can ask AI for ten thousand business ideas and still have no reliable way to know which one is viable.

This is what people mean when they talk about “standing on the shoulders of giants.”

The phrase is sometimes treated like a generic celebration of progress, but there is an important prerequisite we never talk about:

You have to know where (and who) the giants are.

If you don’t know what has already been tried, what has already failed, which arguments have already been answered, or which discoveries changed the field, you can spend enormous amounts of time reinventing things that do not need to be reinvented.

Even worse, without a broad base of knowledge, you may have difficulty challenging your own assumptions. It is hard to recognize that you’re missing something if you have no idea what the competing evidence or alternative explanations might be.

The engineer may understand a technical constraint the entrepreneur has never considered. The historian may recognize that a supposedly unprecedented social problem has several precedents. The camera operator may understand why the director’s theoretically perfect shot will be physically impossible to execute. The customer may know something about using the product that nobody who designed it anticipated.

A knowledgeable person can absorb those perspectives and update his model. An ignorant person often cannot even recognize their significance. Relying on external technology to handle all your thinking and fact-checking is a recipe for disaster — especially when that tool is merely regurgitating information it’s found online.

For instance, consider how a claim might seem to be accurate simply because several websites repeat it.

What if those websites are all only citing a single original source such as Reuters or the Associated Press (who may have gotten the initial story wrong)? What if they’re just recursively citing each other in a self-reinforcing loop of nonsense?

We’ve also all gotten somewhat used to judging the quality of information online by its packaging — dismissing some content as untrustworthy based on the amateurishness of the claimant’s prose writing and website design — but what happens when you never see the original source of the information in the first place? What if you only get new information from an AI that accidentally summarizes false claims which had been poorly expressed on an embarrassing website into clean, well-written paragraphs presented in a professionally designed interface?

Suddenly false information and illogical conclusions may seem stronger, more settled, and more significant than they would have a few years ago.

At some point, you need something against which to compare the shadows. Part of that “something” should be learning from the knowledge of other actual human beings whose experiences and expertise differ from yours, but a lot of it should be about retaining a strong base of true information by yourself.

Building a Library to Call Your Own

Facts without context are not wisdom. Memorization alone is not education. Your goal here should not be simply to remember “everything” you can.

Your goal is to build an internal library, featuring:

• Key concepts in philosophy and logic
• An understanding of the many scientific principles, arguments, and counterarguments shaping your worldview
• Hard skills and deep knowledge of techniques and crafts within your field
• An expansive vocabulary
• A strong understanding of historical events
• A broad range of artistic references
• Real-world examples of whatever you focus on that you can draw on any time

All that information should gradually connect into a larger model of the world.

That means reading primary sources instead of exclusively relying on summaries and someone else’s interpretation of them. It means learning enough history to understand what came before what, and which ideas developed in response to which events. It means learning the strongest arguments against your own positions instead of knowing only the version presented by people who already agree with you.

And it means talking to people who know things you don’t.

Seek out genuine expertise. Compare independent sources. Ask people to explain why they disagree with you. Spend enough time around people from different professions, disciplines, generations, and backgrounds that you are repeatedly forced to encounter information your normal intellectual environment would never have supplied.

Rote memorization can be useful, too!

Research on retrieval practice has repeatedly shown that actively recalling information strengthens long-term retention more effectively than simply rereading the same material over and over. Spaced repetition similarly improves retention by forcing us to revisit ideas over longer periods instead of cramming them into one short session.

Your internal library should be substantial, but it should never be closed. The whole point of possessing a strong model of reality is that you can compare it against actual reality — and against other people's models — and revise it when the evidence demands it.

The great news is that none of this has to be complicated.

After finishing a book, close it and try to write down its major arguments from memory. After watching a great film, challenge yourself to define why particular scenes worked or didn’t work before looking up someone else’s analysis. Keep lists of artists, books, films, historical events, techniques, and ideas you want to understand better, then gradually turn those names into actual knowledge. Even LLMs like ChatGPT can help, if you use them right. For instance, you can easily upload a passage of text you’ve recently read and ask it to quiz you on it, or you could write out your opinion on a subject and then deliberately ask it to challenge your perspective with counterarguments and facts that don’t fit your conclusions.

The point is not to reject external tools. It is to make yourself capable enough to use them well.

AI may become the most extraordinary knowledge-access tool human beings have ever created, but it does not abolish the problem Hayek identified.

Knowledge remains dispersed. Context remains local. Experts continue to disagree. New information continues to emerge. And no single interface, no matter how sophisticated, should fool us into believing that all relevant human knowledge has finally been gathered into one infallible mind.

Even in an age of infinite information, it’s always going to be up to you to decide which information is important, which answers deserve to be trusted, when your own conclusions need to change, and which ideas are worth turning into something new.

That makes having your own knowledge, judgment, taste, curiosity, and imagination — and the ability to learn from other people who know things you don’t — all the more important.

Citations:

• Sparrow B, Liu J, Wegner DM. Google effects on memory: cognitive consequences of having information at our fingertips. Science. 2011 Aug 5;333(6043):776-8. doi: 10.1126/science.1207745. Epub 2011 Jul 14. PMID: 21764755.
• Morrison AB, Richmond LL. Offloading items from memory: individual differences in cognitive offloading in a short-term memory task. Cogn Res Princ Implic. 2020 Jan 3;5(1):1. doi: 10.1186/s41235-019-0201-4. PMID: 31900685; PMCID: PMC6942100.
• Vosoughi S, Roy D, Aral S. The spread of true and false news online. Science. 2018 Mar 9;359(6380):1146-1151. doi: 10.1126/science.aap9559. PMID: 29590045.
• Udry J, Barber SJ. The illusory truth effect: A review of how repetition increases belief in misinformation. Curr Opin Psychol. 2024 Apr;56:101736. doi: 10.1016/j.copsyc.2023.101736. Epub 2023 Nov 14. PMID: 38113667.
• Lisa K. Fazio; Repetition Increases Perceived Truth Even for Known Falsehoods. Collabra: Psychology 1 January 2020; 6 (1): 38. doi: https://doi.org/10.1525/collabra.347
• Roediger, H. L., & Karpicke, J. D. (2006). Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention: Taking Memory Tests Improves Long-Term Retention. Psychological Science, 17(3), 249-255.