Blog

The data wallet

You treat your money with a certain seriousness. You know where it is, roughly how much there is, and who has access to it. You keep it in institutions that are regulated, insured, and transparent about their terms. You'd notice if someone moved it without your permission, and you'd be alarmed if the institution holding it suddenly changed its business model in ways that affected your balance. When you switch banks, your money comes with you. It belongs to you, full stop, regardless of which institution happens to be holding it at any given moment.

Now think about your knowledge. The years of accumulated notes, research, annotations, saved articles, documented thinking, client insights, and hard-won understanding that you've built up over the course of your career. Do you know where all of it is? Could you move it to a different system tomorrow if you needed to? Do you know what the terms of service say about who can access it, who can train on it, and what happens to it if the company holding it changes direction?

For most people, the honest answer is no. Their knowledge is scattered across a dozen apps, each with its own terms, their own export limitations, and their own ideas about who that data ultimately serves. The asymmetry between how seriously people treat their financial assets and how casually they treat their knowledge assets is remarkable, especially given that for most knowledge workers, the accumulated knowledge is what makes the financial assets possible in the first place.

Knowledge as a financial asset

Jaron Lanier argued in Who Owns the Future that the data people generate should be treated as a form of labour, with the dignity and compensation that implies. His argument was economic: people create enormous value through their data, and the platforms that collect it capture that value while the creators receive nothing in return.

The same argument applies to personal knowledge, but with an important extension. Your financial data is valuable to a bank because it helps them sell you products. Your knowledge is valuable to an AI provider because it helps them build products that may compete with you. The knowledge asset is more strategically sensitive than the financial one, which makes the casual way most people treat it even more striking.

A consultant who has spent fifteen years building up frameworks, client case studies, and industry expertise has created an asset that's worth a career's salary if not more. That asset is currently stored, for the most part, in applications where the consultant has no meaningful control over how the data is used, no guarantee of portability, and no protection against the provider deciding to enter the consulting business themselves, which is exactly what's happening as frontier labs launch vertical products in professional services categories.

What a data wallet is

A data wallet is the infrastructure equivalent of a bank account for your knowledge: a system where your accumulated intellectual assets live under your control, portable, encrypted, and independent of any individual tool or provider.

The analogy to financial infrastructure is useful because it makes the requirements concrete.

Ownership. Your money in a bank account belongs to you, legally and practically. A data wallet means your knowledge lives in storage you control: your own cloud storage account, encrypted with your keys, governed by your policies. The application layer reads from and writes to your storage, but the data belongs to you in the same unambiguous way your money belongs to you even though a bank is holding it.

Portability. You can move your money between banks without losing any of it. A data wallet means your knowledge is stored in formats and accessed through protocols that make it portable across tools. If you switch from one knowledge system to another, the knowledge comes with you. If a tool shuts down, the knowledge is unaffected. The tools are services you choose; the knowledge is an asset you own.

Interoperability. You can use your bank account with any shop, any payment system, any financial service. A data wallet means your knowledge is accessible to any AI agent, any tool, any workflow through open protocols like MCP. Your knowledge library serves as the context layer for whatever AI you're using, without being locked into a specific provider's ecosystem.

Privacy. Your bank doesn't share your financial details with other customers or use your transaction history to build products that compete with your business. A data wallet means your knowledge stays private by architecture, with clear boundaries enforced by the system rather than by promises in a terms of service document. Open source models process your data without it flowing to a third party's training pipeline.

Compounding. Your money in a savings account earns interest. Your knowledge in a data wallet compounds through context: every article you capture, every note you write, every connection you draw makes the library more valuable and makes the AI that reads from it more useful. Unlike a savings account, the returns accelerate rather than remaining linear, because each new piece of information connects to more existing pieces.

The personal server

There's a broader vision here that goes beyond storage. Tim Berners-Lee, the inventor of the web, has spent the last several years working on a project called Solid, which proposes that individuals should have personal data stores (called "pods") that they control, with applications requesting access to specific data on the user's terms rather than absorbing the data into their own systems. The idea is that the web should work more like a library where you own your books and lend them to applications that want to read them, rather than a system where every application maintains its own copy of your information and you have to trust each one independently.

Solid hasn't achieved mass adoption, partly because the infrastructure wasn't ready and partly because the incentives for platforms to adopt it were weak (why would Facebook let you take your social graph somewhere else?). But the underlying thesis, that individuals should control their data and applications should request access rather than demanding ownership, is becoming more viable as the technical infrastructure catches up.

A knowledge management system built on bring-your-own-storage, open protocols, and model choice is a practical implementation of the personal server idea for the specific domain of knowledge work. Your notes, your research, your annotations, your saved articles, your voice memos, your web clips: all of it living in your own storage, accessible through open protocols, processed by models you choose, and compounding in value over time in a way that benefits you rather than the infrastructure provider.

Starting the wallet

The practical starting point doesn't require a radical change in how you work. It requires a change in where your work accumulates.

Connect your existing sources, Google Drive, Dropbox, email, reading highlights, into one searchable library that lives in storage you control. Continue using the tools you already use for creation. Let the library be the place where the outputs accumulate, searchable by meaning, connected by concept, and owned by you.

Over weeks and months, the wallet fills. The AI assistant that reads from it gets better, because there's more context to draw on. The search gets more useful, because there are more things to find and more connections between them. The value compounds, and because the infrastructure is yours, the compounding benefits you.

Your knowledge took years to build. It's the foundation of your career, your competitive advantage, and your capacity to do the work that matters to you. It's an asset, and it deserves the same seriousness, the same infrastructure, and the same protection as any other asset you own.

Frequently asked questions

How is this different from a second brain? A second brain is a methodology for capturing and organising knowledge. A data wallet is an ownership architecture: it's about where the knowledge lives and who controls it. You can practise second brain methodology in a data wallet architecture, and you should, but the data wallet adds the ownership layer that most second brain implementations lack. Your second brain in Notion is Notion's asset. Your second brain in a data wallet is yours.

Does this really matter for individuals, or is it mainly an enterprise concern? It matters for anyone whose accumulated knowledge is professionally valuable. An individual consultant, researcher, or freelancer has the same structural exposure as a large organisation: their knowledge is their primary asset, and the question of who controls it affects the value of their career in the same way it affects the value of an enterprise.

What happens to my wallet if Fabric disappears? This is the whole point of the architecture. Your data lives in your storage account (S3, R2, or similar), not in Fabric's systems. If Fabric disappeared tomorrow, your data would be exactly where you left it, in your storage, in standard formats, accessible to whatever tool you choose next. The data wallet outlasts any individual application, which is precisely what ownership means.

Is the data wallet concept new? The concept has been discussed under various names for decades: personal data stores, data pods (Solid), personal clouds, self-sovereign data. What's new is that the technical infrastructure, cheap cloud storage, open source AI models, open interoperability protocols like MCP, has matured to the point where the concept is practically implementable for ordinary users, not just technically sophisticated early adopters.