Context Is Your Capital.
The new (and only?) moat in the age of AI.
By Vincent Schalk
A few months ago, I watched Claude Code build a production-ready internal tool from a handful of prompts. Not a prototype. Not a proof of concept, but a real, deployable system - schema, backend, interface - assembled in the time it used to take a team just to finish the requirements document.
That moment shifted something for me - not because AI can write code, we already knew that, but because of what it implies about everything else. If complex software can already be generated on demand from context and compute, what exactly stops the same from happening to content, to shopping experiences, to learning, to entertainment?
The honest answer is: not much - and that is why we should care.
Given enough compute and the right context, anything you consume online can be generated for you, individually, in the moment you need it, assembled on the fly instead of selected from a pre-existing pool. And that leads to a more interesting question, the one that has been bothering me ever since:
When Everything Can Be Generated, What’s Left That Can’t Be?
I think the answer is personal context.
Your social graph, your identity, your behavior, your history, your taste, your patterns - the accumulated record of who you are and what you care about. That is the bedrock, and the race to control it is already underway.
Today, you open Netflix and pick from a catalog. You scroll Instagram and see what the algorithm decided to show you. You search Google and get results that someone, somewhere, optimized for. The content existed before you arrived; the platform simply shaped it around you.
You are consuming something pre-made.
Now imagine the opposite.
You arrive, and the content does not exist yet. It is built for you as you show up, from scratch, in real time - constructed from your context rather than selected from a library.
A documentary about a topic you have been circling for weeks, delivered at exactly the depth you can absorb right now, in the format you engage with best. A product that did not exist until your agent described what you needed and a system generated it. A learning experience calibrated precisely to where your understanding breaks down.
This is not science fiction. It is simply where the current trajectory leads.
It is no longer about who has the best content library, or even who controls distribution. It is about who understands you best. Who has the most accurate, most complete, most up-to-date model of your context - and can use it to generate something that actually fits.
Your context becomes the raw material everything else is built from, the last input that cannot itself be generated.
The Race for Context
Meta, Google, Microsoft - they see this clearly.
What looks like an AI arms race on the surface is, at a deeper level, a race for context. A race to become the system that knows you best, because in a generative world, that system becomes the one everything else flows through.
Every product decision starts to look different when you view it through that lens.
Meta is embedding AI across WhatsApp, Instagram, and Messenger - not just adding features, but capturing context across the most intimate communication layer people have. Google is threading Gemini into Search, Gmail, Drive, and Maps - not just improving usability, but building the most complete behavioral model ever assembled, now continuously updated and fed back into the system. Microsoft is placing Copilot across the enterprise stack - documents, emails, meetings, decisions - turning everyday work into a stream of structured context.
These are not just AI products. They are context acquisition systems.
The last generation captured your clicks and your social graph. This one is going further - it is trying to model how you think.
The Paradox Of Handing Over Your Most Valuable Asset
Here is where it gets uncomfortable.
I am German, and privacy is not an abstract position for people from my generation - it is a reflex. We were taught to guard our data, to resist surveillance, to be suspicious of handing personal information to large corporations.
And then I look at what we all actually do with AI.
I see many people uploading files, sharing private documents, and pasting conversations that were never meant to be public, feeding in personal context to make personal agents more useful, more efficient, better at the tasks people no longer want to do themselves.
In the past year, we might have given away more than we did in the decade before.
Every power user is doing the same thing - uploading context that was never on the internet. Private notes, meeting transcripts, personal workflows, idea concepts, not the curated version of themselves, but the real one.
AI companies have been saying they are running out of training data. That the public web has been consumed, and that further improvement requires something new.
So here is the uncomfortable question: Are we becoming the new source?
Are we - the people who cared most about privacy - now voluntarily handing over the most valuable data that has ever existed? Data that was never scraped, never indexed, never publicly available, data that describes not just what we do, but how we think. If that is true, the lock-in being built now goes far beyond anything we have seen before.
Meta knew your social graph. Google knew what you searched for. These systems are learning how you reason, how you speak, and once that model exists inside a platform, leaving it behind becomes almost impossible.
Is It Priced In?
Amazon took more than seven years to reach $100 million in annual revenue. Google took roughly three. That made sense. Building distribution and trust takes time.
OpenAI reportedly reached $100 million in annualized ad revenue within six weeks of introducing ads into ChatGPT. The obvious explanation is novelty or scale. But novelty explains attention. It does not explain monetization at that speed.
What advertisers were paying for was something more valuable - relevance at the moment of intent. Not a demographic proxy. Not a keyword match. A system sitting inside the user’s context while the need is being expressed.
If that is true, then the speed of monetization is not just impressive. It is a signal. The return on that context must be extraordinarily high.
Look at TikTok versus YouTube Shorts. Same format. Similar creators. A completely different experience. TikTok feels like it knows you. YouTube Shorts feels like it is guessing. That difference is context depth. The system that understands you best wins.
Relevance at the level of actual context is the most valuable thing in the attention economy.
The Solution: Identity Liquidity
So here is the problem. Right now, the context that defines your online experience - your viewing history, your taste, your behavior, your reputation - belongs to the platforms that captured it.
You generated it, but you do not own it. You cannot move it, deploy it elsewhere, or use what one system knows about you to improve another. You start from zero every time you cross a platform boundary.
This is the identity liquidity problem, and in a generative world, it becomes central. Because when everything is generated from context, fragmented context means no system actually knows you. Everyone has a piece, but no one has the whole. You end up with partial versions of yourself, locked across platforms, none complete enough to generate the experience you actually want.
Identity liquidity is the answer.
Your identity - your preferences, your history, your behavior - should be portable. Yours to carry, yours to deploy. A context layer that moves with you across everything you use, not owned by the platform that captured it.
In practice, that means rethinking identity from the ground up, not as a login, but as a system that accumulates context over time and can be reused anywhere. In the work we are doing with FanPass at OneFootball, this starts with something simple: your history as a fan should not reset every time you change platforms, but travel with you as a continuous, portable record of what you care about. Your taste should belong to you. Your history should belong to you. Your reputation should belong to you.
The technology already exists - decentralized identifiers, verifiable credentials, zero-knowledge proofs.
What has been missing is a forcing function.
That forcing function is now here with personal agents.
The Compounding Nature Of Context
You created the value, yet you do not control it. Fragmented context produces worse outcomes, as systems guess instead of knowing, and both content and products miss the mark.
Lock-in compounds over time, because the better the model gets, the harder it becomes to leave. At the same time, the privacy paradox accelerates everything, as we feed these systems the most intimate data ever created - voluntarily. And in an agentic world, this becomes decisive.
An agent without your context is a stranger acting in your name, while an agent with your context can actually represent you.
The Context-Native Internet
When context is portable and user-owned, the internet changes.
Instead of feeds optimized for engagement, you get experiences built around what you actually care about. Instead of your attention being sold, your preferences become the signal that determines what gets built for you. Instead of starting from zero everywhere, your identity travels with you, and every system begins with who you are.
In domains like sports, that shift is especially visible. A fan’s history - matches watched, communities joined, moments experienced - is inherently cumulative, and once that history becomes portable, it stops being just data and starts functioning as identity.
Most of today’s agentic systems are reactive. They execute tasks - booking flights, summarizing documents, drafting emails. Useful, but limited. The more ambitious version is different.
A continuous, evolving stream of context that you own and can deploy anywhere, not a static profile but a living representation of who you are. When that exists, the generative future becomes something else entirely. Not a system optimized to extract attention, but one that works from your context, on your behalf. Your experience, shaped for you. Your identity, portable. Your context, working for you - not on you.
That is what identity liquidity enables. The race is already underway, and in some corners it has already started to take shape as products begin to treat identity not as a login, but as a layer.
The question is whether the system that wins is owned by platforms - or by you.
This article is part of a series on identity liquidity published on Proof of Passion.