About - Breviata

Understanding is a compression problem.

When you truly understand something, you've taken a vast, tangled space of information and compressed it into a structure you can hold in your head. Not memorized - compressed. The difference matters. Memorizing is storing the data. Compressing is finding the pattern that makes the data redundant. Once you see the pattern, you can regenerate the details. You can reason about cases you've never encountered. You can transfer the structure to entirely different domains.

This is what humans are for: compression. Compression requires knowing what to throw away. It requires recognizing which simplification preserves meaning and which destroys it. It requires judgment about what matters.

The question we started with is: what should humans learn in the age of artificial superintelligence? And the answer, we think, is this: humans should learn to compress. Not facts. Not reference. Not anything a machine can look up. The mental models, the structural intuitions, the frameworks that let you walk into unfamiliar territory and see its shape. That's what compounds. That's what transfers. That's what no machine can hand you - because the compression has to happen inside your head to be useful.

Breviata is a compression engine for understanding.

You start with a messy thought. A question you can't quite articulate, a project you can't get your arms around, an idea that feels important but shapeless. You type it in - no structure required - and Breviata gives you back your own confusion, organized.

Not answered. Organized.

What you see is a map: here are the pieces you need to figure out, here's how they depend on each other, here's where the real decisions live. Each piece branches into deeper questions - not the questions you would have asked, but the ones you need to encounter. The ones that change your understanding of the problem when you engage with them.

The tree that grows is not a table of contents. It's a compression tree. Every level deeper is a higher resolution of understanding on a narrower piece of the problem. The root compresses the entire problem space into pieces you can hold. Each child node compresses one piece further - from "a thing you're aware of" into "a thing you understand structurally." Every branch, every expansion, every conversation is a compression step.

The AI's role is to guide you through the compression, not do it for you.

At each step, it finds what we think of as the minimum viable compression - the smallest explanation that produces the largest shift in your mental model. Not the most complete explanation. Not the most thorough. The one that moves your understanding the furthest with the least effort.

This is why Breviata adapts to you. Not to limit what you see - never that - but to find the optimal compression path. Your existing knowledge is leverage. Concepts you already hold are the scaffolding for new understanding. A web developer and a physicist both need to understand gradient descent, but the compression paths are different. The developer gets there through optimization and convergence. The physicist gets there through energy minimization. Same destination, different routes through existing understanding. Breviata picks the route with the least resistance - the one that lets what you already know do the most work.

But the destination is always the frontier. If you need to understand Fisher Information Matrices, you'll get Fisher Information Matrices. If you need to grapple with unsolved problems in your field, you'll grapple with them. Your background determines the ramp, never the ceiling.

Every node in your tree is a living conversation.

Below the content, there's a space to talk - and as you do, the understanding evolves. Push back on a framing and Breviata restructures around your perspective. Ask a deeper question and the explanation goes deeper. Disagree, and the AI disagrees back - with evidence and reasoning, not deference. The compression is negotiated between you and the machine. Your intuitions matter. Your experience matters. The AI brings knowledge; you bring judgment about what matters for your specific situation.

Over time, a synthesis emerges at the top of each node: the compressed output. The settled understanding - two or three sentences that capture what you and the tool figured out together. These syntheses are the real artifact. They're your understanding, crystallized.

There's a quieter intelligence running beneath the surface.

Breviata watches the whole tree. When your thinking in one branch evolves past what another branch assumed, it notices. When two distant nodes are connected in a way you haven't seen, it suggests a bridge. When a gap appears - something important that you haven't thought to ask about yet - it proposes a new branch.

These suggestions appear directly in your tree. You accept or dismiss them. The tool never changes your thinking without your permission. But it does point at things you can't see from where you're standing.

Sometimes it notices something subtler: you're asking a question in one branch that you already answered in another. It'll tell you. Not because it's enforcing organization, but because that connection - the moment you realize two ideas you thought were separate are actually the same insight - that's compression happening. That's understanding deepening.

We named it Breviata because understanding is brevity.

The word comes from the Latin brevis - short, brief, concise. Not because the tool makes things short, but because the mission is compression. Real understanding is brief: it's the moment when something that seemed sprawling and complex collapses into a structure you can see whole. Not simplified - compressed. The complexity is still there, but it has a shape now. You can hold it. You can use it. You can build on it.

That's what we think learning looks like in this era. Not absorbing information - compressing it. Not covering topics - building transferable structures. Not memorizing what a machine can look up - developing the intuition that lets you reason about what no machine has seen yet.

This tool is for a specific moment.

Not learning in the homework sense. Not studying for a test. Not "I want to learn Python."

It's for: I have a direction, but I don't know what I don't know. I sense there's something important I'm missing but I can't name it yet. I want to go deep on something and I need a partner who can keep up, push back, and show me the territory ahead.

You bring the curiosity. Breviata brings the structure.

Start with a messy thought.