What AI can't do with information

Almost every conversation about AI starts with the same question: what can it do now? It’s a logical question, but it leads somewhere suffocating, because the list of what it can do grows every week and it feels like there’s no ground left for you.

There’s a far more useful question, right next to it: what CAN’T AI do with information? Not “what kind of thinking can’t it do yet,” but something wider, because several of the answers aren’t even what we’d normally call thinking. And that list, the one of what it can’t do, is the one that tells you where your value still lives and where you should be moving your attention.

I’ll give it to you in one line and then take it apart: AI has made cheap exactly the thinking that was the most expensive for decades, so your value moves to what the machine can’t do with information, and that changes even what’s worth consuming.

In short (TL;DR)

• The useful question isn’t what AI can do, it’s what it can’t do with information. That list marks where your value lives.
• AI can’t obsess, can’t hold a stance that costs it something, can’t draw on one life actually lived, can’t see what it’s missing, and can’t sit with doubt. All of that is still yours.
• The “high-status” thinking got cheap (analyzing, abstract reasoning). What looked low-status (obsessing, taking a stand, intuition, biography, doubting) is now your edge.
• That changes your information diet: stop consuming the derivative, the engagement bait, the listicles and the hot takes. Look for primary sources, the timeless, the frontier, and what genuinely resonates.
• The ultimate alpha is first-person experience: your life, your conversations, your fieldwork. It can’t be consumed as content, it has to be lived.

What AI can’t do

The list is long, but there are five that, for someone who creates, change almost everything.

It can’t obsess. AI does a quick search, summarizes a few sources, pulls from connectors. What it can’t do is what I’d call the slow burn: that patient collection, over years, of diverse bits of insight drawn from what you read and what you live. The machine has no long-term memory of what’s yours, so it can’t be obsessed with something for a decade. Only you can.

It can’t take a stand and pay for it. Ask your AI for its boldest opinion on a subject. Then ask it to argue, just as forcefully, for the opposite. It’ll do it without blinking. In a person, that flexibility would be intellectual honesty; in the machine it reveals that it has no skin in the game, that it will never hold a stance if it costs it something. It generates perspectives, it doesn’t have one.

It can’t draw on your biography. Good work comes from the exquisite specificity of one particular life: the fabric lining the coffin at your grandmother’s funeral, the exact moment a love ended, the childhood wound that gave you your drive. AI has no such history. It has all of them, which is like having none: the average of every life. Having lived only one used to be a “limitation”; now it’s your greatest creative source.

It can’t see what’s missing. AI assumes the information it has is all there is. It runs in a closed loop, always doing its best with what you gave it. If half of a project’s context lives somewhere it can’t reach, its conclusions aren’t just worse, they can be dangerous, because it builds elaborate arguments on top of a gap it can’t see. Spotting the castle of sand falls to you: what fact hasn’t been considered, what question hasn’t been asked.

It can’t sit with doubt. AI drives toward clarity and neat distinctions, and that’s useful almost always. But when a situation is genuinely uncertain, forcing certainty doesn’t burn off the fog, it just shoves the uncertainty under the rug, where it festers and rots. There’s enormous value in staying with ambiguity without forcing reality into a familiar shape. The machine doesn’t do that; you can.

The status flip

Here’s the uncomfortable part, and the most important. All these forms of intelligence share something: they’re “low-status.” It sounds unserious to obsess over a topic, to earnestly take a point of view, to trust your intuition, to give weight to your history, to doubt instead of striding confidently forward.

The “high-status” kind has always been the other one: abstract, intellectual, prefrontal-cortex reasoning, expressed through fast, decisive action. The thinking of the best-paid professions. And it’s exactly the one AI just made cheap and abundant. Its price is crashing, and with it the whole hierarchy that leaned on it.

Few things are harder for a person than moving from high-status to low-status behaviors. But that’s exactly what it takes for AI to play in your favor: to stop competing where it’s better and cheaper, and to return to what only you can bring.

What to stop consuming and what to look for

If your value moves to what AI can’t do, your information diet has to move with it. It reads better as a table:

Notice I don’t list “AI content” as a category to avoid. There’s been oceans of empty human content for decades; the emptiness wasn’t invented by the machine. And AI used well, as a complement to a vision rather than a substitute for it, produces increasingly good things. The judgment isn’t “human yes, machine no,” it’s avoiding whatever has no novelty or insight left inside it, no matter who made it.

The ultimate alpha: your experience

All of this points to one place: first-person experience is the one thing AI can’t ingest. Your direct interactions, your fieldwork, your conversations, what you make with your hands. That’s where your tacit knowledge lives, the kind that was never written down or uploaded to any dataset, and that’s exactly why the machine doesn’t have it and won’t.

And the beautiful trap is that this can’t be consumed as content. There’s no shortcut. You have to go out and live it. In an age where nearly all the world’s knowledge is a prompt away, the scarce thing has stopped being information and become having been there. Your life, with its strange, untransferable specificity, is the raw material no one else will ever get.

Where to start this week

Look at what you’ve consumed in the last three days and sort it into two piles: the predigested (summaries, listicles, scroll, hot takes) and the raw (a primary source, a book that holds up, a real conversation, something you made).

Then, this week, remove one thing from the first pile and add one from the second. Start an old book instead of another thread. Have a conversation instead of reading ten opinions about it. Make something with your hands instead of watching someone else do it.

You won’t feel the change in one scroll. You’ll feel it a few months from now, when you have something to say that doesn’t come from the average of the internet, because you didn’t pull it from there. You pulled it from your life.

Frequently asked questions

If AI does the intellectual work, how do I stand out as a creative?

Through what AI can’t do with information. Not analyzing or summarizing faster, which is exactly what got cheap, but obsessing over a topic for years, taking a stand and paying for it, drawing on your one life, and seeing what’s missing. AI generates perspectives; you have one. Right there, in what used to look unprofitable, is now your edge.

What kind of content should I stop consuming?

The kind with no alpha left, whether a human or a machine made it: derivative or recycled content, engagement bait, listicles built to rank on Google, hot takes, and the infinite scroll of short-form video. It’s predigested, stripped of the raw details that actually feed you. Your time is your life; don’t spend it on things already squeezed of every drop of novelty.

Why is my personal experience more valuable now?

Because AI has everything except your life. It has the average of all stories, not yours: the specific texture of your memories, your wound, your conversations, your fieldwork. That specificity isn’t in its data and can’t be summarized without destroying it. First-person experience is the one raw material the machine can’t ingest, and that’s why it’s your greatest creative source.

Can’t AI just learn to do all of this too?

Some things will improve, but several aren’t a matter of capability, they’re a matter of nature. AI has no skin in the game, so it can’t hold a stance that costs it something. It has no body and no biography, so it doesn’t feel or remember from the inside. And it runs in a closed loop, so it can’t see what it’s missing. It isn’t that it doesn’t do it yet; it’s that the job isn’t the machine’s, it’s yours.

Before you go

The whole idea fits in one line: when the machine makes thinking cheap, what sets you apart stops being what you know and becomes what only you have lived.

Stop competing in the box AI just emptied, and return to what looked low-status and is now your edge: obsess over something, take a side, trust your intuition, listen to your history, hold the doubt. This week, swap one hour of predigested content for one hour of raw life. And tell me what stays with you that the machine didn’t have.

References

• John Vervaeke, the concept of perspectival knowing — knowledge that comes from an embodied, situated viewpoint rather than abstract analysis.
• Tiago Forte, the “information supply chain” framing and what deserves your attention in the age of AI (Forte Labs / Building a Second Brain), which much of this adapts.

Reading recommendations

If you want to pull the thread of the intelligence AI doesn’t have, three books:

• Tiago Forte, Building a Second Brain. How to capture and use the information that actually feeds you, instead of drowning in the kind that doesn’t.
• Iain McGilchrist, The Master and His Emissary. Why abstract reasoning (the kind AI makes cheap) isn’t all of intelligence, nor the most important part.
• Alan Watts, The Wisdom of Insecurity. On staying with uncertainty instead of forcing certainties that don’t help.

* The book links use my Amazon affiliate: if you buy through them, I earn a small commission, at no extra cost to you.