AI Cameras

Abstractly, a camera can mean many things, but it's really a portal. In the early days of painting, our minds & eyes were the cameras. Hoping to capture the fleeting magic of a moment worth preserving and sharing, we translated what we uniquely saw into many styles like impressionism. It was often coarse, interpretive work.

Apprentices studied masters by copying their works over and over until they could reproduce not only the brushstrokes, colors and values, but the subtle feeling of sadness in eyes or hope from a burst of spring sunshine leaking through a window.

As modern mechanical & chemical cameras were developed, most people started to defer to a machine the work our brains and hands used to do. These small machines gave us more control and precision over reproducing what our eyes saw, until eventually the machines saw beyond what even our eyes could see, by swapping out various lenses, films, filters, and timespans.

These early camera machines had already transcended us as humans.

A lot of people fear machines transcending us, but this already happened over 200 years ago. It unlocked creativity and preservation of people's worlds and aesthetics in a way that paintings or drawings or stories weren't able to achieve.

Fast forward, cameras became so precise at reproducing what they saw that they also limited our creativity in ways. There was nothing left to interpret.

When Instagram & Hipstamatic launched, it was the first time we could easily add our own perspective onto a more exacting camera. Our chosen aesthetic offered a glimpse into our point-of-view.

We had a whisper of a voice again.

Zoom levels in life

But, a camera can mean many things. Our minds are a form of camera, empowering us to see a world that others may not, a vision that others can either buy into or opt out of.

Painters like Caravaggio would mentally work their way upwards from the skeletal layer, then to muscles and fat, and finally to the finished form. Only once they understood what was underneath could they accurately capture what was before their eyes.

To help us do the same, we've invented many kinds of cameras to explore many worlds, such as microscopes that can see much further down than our theoretical minds used to take us, X-rays and MRIs to see within the body, telescopes and satellites to observe other galaxies or even our own planet, and even first-person perspective in video games or CGI in movies to imagine universes with rich characters and unique physics.

Before these cameras, we had to hallucinate our way towards an understanding of those possible worlds. Now we have more clarity and escape, but this often comes with the same steep price tag as early painters: time until mastery & utility.

As models are built across the many zoom levels of life, we'll start to have simple controls to not only explore worlds in the way Google Earth first unlocked, but we'll be able to zoom in and out in infinite ways to explore the latent space of our own imaginations and curiosities.

You see, we're very early in the camera-building business, and every camera we know of today will be reinvented and reimagined. Previously, we built for precision and accuracy of what is whereas in the future it'll be a mixture of precision and prediction for what could be.

A world of text vs. a world of imagery

Half the world's population grew up in a world of text, and so they built tools around text: books, emails, text messaging, code editors, modern music players with all their lyrics and lists and even language models. Living with text as a primary communication modality, we've also had to live with its limitations.

The other half of the world's population is growing up in a world of imagery, and soon they will build new tools that transcend what text was capable of. We're already seeing this today with reels that immediately hook you in—almost too efficiently, in a zombie-like state. Imagine trying to do the same with text, and even micro-dosing on X/Twitter won't be able to compete.

Imagery & audio help us to feel something at a visceral level and they often have much shorter time-to-impact than text does. The easiest proof of this are the early internet "jump scare" videos where you'd watch a tranquil scene only to have a scary woman's face flashed on the screen, violently screaming at you.

Ever the "next-token generators", our minds would fill in the blank and instantly imagine something terrible happening—all within the span of milliseconds. It's in this way that imagery and audio are a form of text compression. They communicate in such rich, immediate ways that make text feel like the literal black-and-white world it often is.

So naturally, as the world evolves we'll build tools to communicate in richer, yet more compact ways. These will also all be new types of cameras, enabling a window into worlds that couldn't exist within the limitations of a text-heavy world.

(A)n (i)nfinite camera

When you grow up you tend to get told that the world is the way it is and you're job is just to live your life inside the world. Try not to bash into the walls too much. Try to have a nice family life, have fun, save a little money. That's a very limited life. Life can be much broader once you discover one simple fact: Everything around you that you call life was made up by people that were no smarter than you. And you can change it, you can influence it... Once you learn that, you'll never be the same again.

No one wants a limited life, and no one deserves one either when anything is possible. If we want to maximize the possibility of the impossible, we should all be working on building infinite cameras that build and preserve infinite worlds, perspectives, aesthetics, and change agents.

To many people, this means infinite hope.

In the mind of every child is an empty but infinitely vast canvas; we just need to build the cameras to project the possibility of the impossible onto it.

AI is the first time we've had the chance to build an infinite camera, one that understands the world and can reproduce it while predicting and imagining others far beyond it. We'll have infinite rabbit holes, some of which will be so incredibly compelling that we may even choose them over reality.

Looking back, the AI camera of today will look like a rough, impressionist painting. In the future, everything that exists (whether text, imagery or maybe even physics) will be a jumping-off point to remix into what could exist.

In this pursuit, perhaps we'll even find ourselves along the way.

-- Rob