UX for the post-smartphone era.

Like everyone else, I am already part of an experiment. Social media, smartphones, notifications, chatbots, feeds, and cloud tools are constantly reshaping how I remember, communicate, decide, read, write, and move through the world. The difference is that I want to set some of the conditions myself.

So I have been testing constraints. Can I travel in Europe for two months without my computer? Can I leave home without my phone and still remain reachable, oriented, and capable? Can an Apple Watch, voice, paper, files, and a few well-designed loops preserve the important benefits of technology without pulling me back into the smartphone as the default surface? Can handwriting, Kindle, reMarkable, tablets, and printed packets become serious parts of how I think, design, and build software?

These are not lifestyle experiments on the side of the work. They are the work. They are ways of asking what kind of digital environment makes a person more capable without making them more dependent; more connected without making them more interruptible; better supported without losing the ability to steer.

Once the experiment is visible, the question changes. I am not primarily testing whether AI is useful, whether a chatbot is smarter, or whether one device can replace another. I am testing the shape of the environment itself.

AI is one material. Voice is one material. Files are one material. Paper, e-ink, handwriting, agents, watch interactions, terminal commands, SQL views, PDFs, and human review are all materials. What matters is the composition: what the environment remembers, what it reveals, what it hides, what it lets me repair, and whether I can find my way back after work has passed through many tools, agents, files, and moments of attention.

That is why this is not a better-assistant project. A better assistant can still leave the rest of the environment untouched: notifications still pull, files still disappear, provenance still gets lost, decisions still become untraceable, and the phone remains the gravitational center. The object here is the experience of computing itself.

The test is not whether the system automates more. The test is whether it leaves me more capable of steering my life and work.

An environment amplifies agency when it helps me act with intention, keep context, understand what happened, and recover after interruption. It erodes agency when it remembers instead of me, answers without sources, hides actions behind glass, or turns hard decisions into feed-shaped drift.

The important thing underneath is not a chatbot. It is a filesystem-native knowledge substrate: a way for life and work signals to become durable, queryable, inspectable, and revisable without disappearing into a black box.

The substrate matters because surfaces come and go. A web page, a terminal command, a voice interaction, a paper export, an e-ink reading packet, or a dashboard should be a projection. The durable value is underneath: owned files, identities, traces, provenance, read models, lints, evals, and repair loops.

This is where the technical work becomes philosophical in practice. Provenance is not metadata for auditors. It is what lets a human remain responsible. A view is not just a query. It is a way to ask: what did this conclusion depend on, and what happens if the source changes?

One architecture has to support both tempos: fast and efficient when the system needs to capture, route, and act, and patient enough to adapt to the naturally slower tempo of human thinking. The hard part is not choosing between speed and depth. It is making each loop honest about what kind of attention it serves.

The design implication is concrete: voice and capture deserve production-critical latency budgets; research and synthesis can happen as patient background work. The point is not one universal assistant. It is an environment with multiple loops, each honest about its tempo.

The experiment is no longer imaginary. Voice memos become transcripts. Notes become briefs. Files become reports. PDFs and reMarkable packets support slow reading. Agents can research, synthesize, route work, and produce artifacts. But the system still has gaps where agency depends on visibility, latency, and repair.

The personal experiment is not the ceiling. It is the smallest environment where the full problem can be observed: attention, memory, handoffs, trust, work, interruption, recovery, and action.

Organizations have the same problem at larger scale. They need to know what happened, who owns what, which sources are trusted, how decisions were made, what changed since the last review, and whether AI-assisted work can be audited and repaired. The vocabulary shifts from personal agency to organizational capability, but the design pattern is similar.

The goal is not to replace the phone with another device, or the app with another assistant. The goal is to make computing less dependent on one dominant surface and more compatible with how attention, memory, movement, conversation, reading, and judgment actually work.

This is also a wager about the next design frontier. The phone-and-chatbot stack will improve, and that will matter. But the more interesting frontier is the environment around it: computing that stops demanding constant visual attention and becomes infrastructure for thought.