A gathering of thoughts on building an internet more you. At Crosshatch, we're exploring the meaning of a more personalized internet and imagining what's possible when we truly bring our full context and identity online.
Before AI agents, there were user agents. To get personalized services, have their agent call your agent.
Extending the Model-View-Controller architectural pattern with AI-Bindings to make generative UI react to You.
20 years ago Bill Gurley found the keys to capitalism: the Context Moat. For the age of AI, it's a local maximum.
AI makes personalization more valuable while delivering fixed- and variable-cost savings.
Personalization cold-starts and learning curves are over. Live cross-app context is now on tap.
For memory for AI, we don't just want her side of the story or his. Whatever you want to share, we want the whole story, unfiltered.
Privacy has been a tool of marginalization. It commodifies and constricts identity. Technologies of the self – rituals of self-actualization and truth – are the cure.
Personalization today is expensive and bad. It's not a skill issue: no one has the context. To get personalization from the movies, we need a radical shift.
Identity and the Other are mutually constructed. They pass into one another. How privacy absolutism could rob us of a rich identity.
Privacy has been a tool of marginalization. It commodifies and constricts identity. Technologies of the self – rituals of self-actualization and truth – are the cure.
Identity and the Other are mutually constructed. They pass into one another. How privacy absolutism could rob us of a rich identity.
A subject mindset has replaced the citizen mindset of the early internet. Duty will not fix this. Desire will.
Even in 2003, it felt like walls were going up on the internet. How a progressive self-sovereignty can make internet experiences richer at a reduced cost.
Open collaboration is one of America's greatest strengths. But its realization on the internet has felt constraining and icky. The pathway of AI to a more open and collaborative internet.
Model Context Protocol (MCP) allows users to bring tools and context to desktop AI clients. What about web clients?
The case for AI-native auth with flexible purpose- and time-limited scopes on context.
Simplicity-maxing Tim Berners-Lee’s Solid Project with opinionated design and AI.
Before AI agents, there were user agents. To get personalized services, have their agent call your agent.
Extending the Model-View-Controller architectural pattern with AI-Bindings to make generative UI react to You.
What could be an interoperable context layer for hyper-personalized AI? Any AI over a unified events representation of all interactions you've had with the world.
Hyper-personalization is the internet's version of rolling out the red carpet. Introducing a new personalization architecture that makes it possible.
The personalized internet needs a new data layer made from aggregating the aggregators. What Aggregation Theory predicts comes next.
Fat protocols are crumbling. Where does value accrue in AI, even with portable data? Thin Protocols. Fat "Thick" Apps.
20 years ago Bill Gurley found the keys to capitalism: the Context Moat. For the age of AI, it's a local maximum.
AI makes personalization more valuable while delivering fixed- and variable-cost savings.
For memory for AI, we don't just want her side of the story or his. Whatever you want to share, we want the whole story, unfiltered.
Personalization today is expensive and bad. It's not a skill issue: no one has the context. To get personalization from the movies, we need a radical shift.
AI unlocks malleable UI. It can be executed by the OS, browser or application-layers. Why "thick" applications resist OS or browser-led malleability and personalized AI will ship at the application layer.
Browsers and OS use identity to make data available across devices. The case for identity to make personal AI available across applications.
Supercharge search with personalized re-ranking that adapts to user context without needing to be told.
Expanding the app context window with serverless, user-owned context and AI. Introducing our personalized inference provider for Vercel's AI SDK.
Previewing Crosshatch Webhooks, extending our API to enable any service to take proactive action based on complete user context.
Personalization cold-starts and learning curves are over. Live cross-app context is now on tap.