Interface Alignment
Practices and interfaces for fortifying one's attention from attacks by AI, especially as it gets stronger.
Ask anyone to picture AI. Most will picture a chat window. Our conception of AI grounds in the interfaces we use. So far those interfaces have been only occasionally predatory. That is going to change.
The attacks will be personal, en masse, and far more intelligent than chat. They are arriving as subtle frames that determine everything.
The threat model
Interfaces are where AI risk enters daily life. Some software extracts. Some simply does not fit you. Both are forms of indifference.
You are pulled in further than you would choose. Not in ways you can easily refuse. The software is rewarded for taking your attention, your time and your data.
You pull back from capacities you used to exercise. Attention, judgement and taste get outsourced. Then they atrophy. The harm is unintentional. It is still real.
The deepest version. Intelligent software shapes the frames your choices are made inside. It decides what counts as a problem, and what is worth attending to at all. Fixes at the level of content miss this entirely.
An email client used to sort by time. Now it writes in your voice, decides what is urgent, and nudges you to reply. AI makes this category of harm intelligent.
Once habituated, technology becomes the ontology we think inside. The patterns forming now are the ones stronger systems will inherit.
What we do
The frame that filters attention matters more than the content within. This is where we come in.
We work with individuals personally. We design tailormade interfaces to match the threat, and give back choice. The user and their flows are right there, as we work.
Co-agency is a relationship. Relationships cannot be shipped. They can be practiced, studied, written down and taught. So we hold the relationship and the judgement ourselves.
Building software for one person used to be far too expensive. AI-assisted coding collapsed that. A practitioner can now build a working interface inside a single session, then iterate on it.
The mindset has helped me be more creative about how I'd like my computer to be.
I always blamed myself for task-hopping. It was not willpower. My work interfaces themselves were chaotic.
Interface Alignment Slowathon
Three weeks. You will learn the frames and the design methodology, then work live with people.
The focus: spot where a user tunes out, or gets pulled in against their will. Then design one small intervention for it.