My Screenshots Know About the Life I Almost Lived

I have taken an absurd number of screenshots across my life.

Most of them were never organized. Many were never opened again. They sit inside old phones, cloud backups, forgotten folders and photo libraries mixed between actual photographs. Some are useful: tickets, addresses, receipts, error messages and directions. Others are harder to explain. A sentence from an article. A strange interface. Someone's profile. A job I considered applying for. A restaurant in a city I might visit. A product I might buy. A person I might contact. An apartment I could imagine living in.

At the moment I took each screenshot, something felt important enough to preserve. Then I moved on.

Photographs usually document what happened around us. Screenshots document what happened inside our attention. They capture what interrupted us, confused us, made us laugh, looked beautiful, felt suspicious, seemed useful or appeared to contain some possibility we did not want to lose.

A camera roll shows where I went. My screenshots might show where my mind was going.

If I gave an AI access to every screenshot I had ever taken, it could probably reconstruct the changing structure of my interests. It could identify the months when I was obsessed with filmmaking, Taiwan, investing, health, design, AI tools, travel, job applications or some historical rabbit hole that briefly took over my life. It might find that some interests appeared once and disappeared, while others kept returning under different names.

There are platforms designed to capture taste. Pinterest boards, bookmarks, playlists, mood boards and saved posts. But those collections are deliberate. They are curated versions of who we think we are.

Screenshots are messier.

I might screenshot a beautiful interface without knowing why I like it. I might capture a hotel room, a jacket, a film still, a typeface, a person, a map or a product without placing it into any formal collection. Over thousands of screenshots, patterns would begin to emerge. The same kinds of rooms. The same visual density. The same faces. The same interface structures. The same kinds of cities.

An AI could build a model of my taste from decisions I never consciously described. It could show me the visual patterns I have repeated for years and help me understand why certain things continue to feel like me.

Screenshots also capture problems. We screenshot error messages because we want to fix them later. We screenshot confusing booking pages because we are unsure what to choose. We screenshot medical information because we want to remember a symptom or compare an explanation. We screenshot products because we are trying to decide between them. We screenshot conversations because something needs to be proven, remembered, translated or discussed.

Across an entire life, these are no longer isolated incidents. They become a database of friction.

AI could identify the problems I repeatedly encounter. It might notice that I struggle with the same kinds of government forms, travel bookings, payment systems, account settings or health questions. It could reveal that many of my supposedly new problems are recurring versions of old ones.

For a designer, this could be especially valuable. My screenshots may contain years of informal user research. Every broken interface, confusing workflow, strange confirmation screen and badly worded message I encountered is still sitting there. I have been collecting evidence without calling it research.

Screenshots could also reveal how I make decisions. How many products do I compare before buying one? How long does a city appear in my screenshots before I decide to visit? How many job listings do I save before submitting an application? Which companies keep returning to my attention? What do I screenshot immediately, and what do I return to weeks later?

AI could identify things I nearly did. The jobs I almost applied for. The people I nearly contacted. The products I almost bought. The cities I almost moved to. The projects I nearly started. The relationships I considered pursuing.

Our official biographies are built from decisions that became visible. Screenshots preserve the much larger field of possibilities surrounding those decisions. They contain the branches we never followed.

Many screenshots are small promises made to a future self.

Read this.

Visit this place.

Contact this person.

Try this tool.

Buy this book.

Build something like this.

Remember this idea.

Most of those promises are forgotten.

AI could turn a screenshot archive into a map of unfinished intentions. It could separate abandoned ideas from ideas that still feel alive. It could notice that I have screenshotted variations of the same project for eight years without ever starting it.

It could ask uncomfortable questions. Why do I keep saving information about this subject? Why have I returned to this city so many times? Why do I repeatedly screenshot companies working on the same problem? What am I waiting for?

A normal task manager only knows about the intentions I consciously entered. Screenshots contain intentions that appeared before I was ready to name them.

Screenshots may also document the changing shape of my social world. Profiles, group conversations, event pages, contact cards, comments, emails, dating apps and people appearing in the background of other information.

An AI could reconstruct communities I passed through. It might find people who appeared constantly for several years and then suddenly vanished. It might identify someone I intended to reconnect with but never did. It could show how my professional and personal networks changed as I moved between countries, industries and versions of myself.

This would be powerful. It would also be deeply invasive.

A screenshot archive contains other people's lives as well as mine. Any system capable of understanding it would need strong privacy boundaries, local processing and careful controls over what is indexed, remembered or shared. The more useful the system becomes, the more dangerous careless access becomes.

The most unsettling possibility is that screenshots contain emotional patterns. What do I screenshot late at night? What do I save during periods of uncertainty? What appears when I am lonely, ambitious, anxious, attracted to someone or planning a major change?

Perhaps every time I begin wanting to leave a country, I start screenshotting apartments, universities, local events, flight routes and people in another city months before I admit that I want to move.

AI might notice the beginning of a personal transition before I do.

A screenshot is rarely labelled with an emotion. But emotion may still be visible through repetition, timing, subject matter and context. The pattern could be clearer than the individual image.

I used to think screenshotting was a crude form of bookmarking. Now I think it may be closer to external memory.

A bookmark preserves a destination. A screenshot preserves a momentary state. The exact sentence. The exact price. The exact profile. The exact arrangement of information. The exact thing I was looking at when my attention decided: keep this.

That difference matters.

A future AI could learn why I take screenshots. It could distinguish between evidence, inspiration, reminders, references, jokes, attractions, ideas and emotionally meaningful moments. It could stop treating every screenshot as an image and start treating it as a cognitive event.

The obvious product is AI search for screenshots. Find the screenshot with the hotel address. Find the error message from last year. Find the shoes I considered buying.

Useful, but limited.

The more interesting product would interpret the collection as a whole. It could generate chapters such as "The six months you nearly became a filmmaker," "Every company you considered working for," "The interface patterns you keep collecting," "People you intended to contact again," "Cities that appeared before major life changes," or "Ideas you screenshotted before they became mainstream."

The result would be part archive, part autobiography and part psychological mirror.

My photographs show the life I lived. My screenshots may contain the life I was considering.

They record thousands of tiny moments when another future became briefly imaginable. Somewhere inside that archive are abandoned projects, forgotten attractions, unresolved questions, emerging obsessions and versions of myself that never fully arrived.

AI could help me search them.

The more interesting possibility is that it could help me understand what they were trying to tell me.

Connected reading

• Eleven Terabytes Are Not a Story
• Sixty Bets Against Forgetting
• The Real Bottleneck Is Human Attention
• The Interface That Wants You to Touch It
• The Copy Button Is Missing

AI authorship note

This essay began with Kris Haamer asking what AI could learn from the screenshots he had taken across his life. AI helped expand the idea into a personal essay about attention, taste, recurring friction, decision-making, unfinished intentions, social history and emotional patterns. Kris asked for the piece to be written as a Raw Drafts article, then removed the sectioned structure in favor of continuous paragraphs. He selected this version and made the final publication decision.