signal.brettmcm

Disposable apps, durable data

Fri, 14 Aug 2026 00:00:00 GMT

I keep building apps for problems I have technically already solved.

My reminders are in Apple Reminders. My time is in Apple Calendar. Both sync everywhere, survive new devices, understand recurrence, and have years of quiet infrastructure behind them. The responsible thing would be to stop there.

But I do not want to interact with a general-purpose reminder list every time I need to understand my day. I want a small thing in my menu bar that tells me what is left, puts events and tasks in one sequence, and then gets out of the way.

So I made Nearterm.

View the interactive capsule on Signal

Nearterm does not own anything. It is a 360-point opinion laid over services I already trust. Complete a task there and it is completed in Reminders. Reschedule it and the date changes in Reminders. Calendar remains Calendar. If I stop maintaining Nearterm tomorrow, I lose an interface. I do not lose my life.

That separation feels intoxicating.

For most of software history, making a new app also meant accepting responsibility for accounts, databases, sync, recovery, permissions, and every strange little edge where durable software earns the name. It made an interface expensive. Once the interface is personal enough to have an audience of one, the math stops working entirely.

AI changes that math, but the stable layer is what makes it useful. I can spend my time deciding that the type should be 13 points, that today’s events belong above today’s tasks, and that the whole thing should be black, white, and quiet. Apple can keep worrying about whether a recurring reminder appears on my phone after I edit it on my Mac (thank god).

The app becomes less like a permanent container and more like a lens. I can change it when my habits change. I can make another lens for another context. I can throw one away without planning a migration.

I used to think personal software meant owning the whole stack. Now I think it might mean knowing which parts I have no interest in owning.

Minimal by subtraction

Mon, 03 Aug 2026 00:00:00 GMT

I started MCMDS by documenting everything I knew about my own taste. Color, type, spacing, imagery, product, print. I wanted an agent to have enough material to make the same decisions I would.

Instead, I made a skill that could explain my design system better than it could use it.

The structure was right: universal rules in DESIGN.md, judgment in BRAND.md, production constraints by channel, and reusable assets for output.

View the interactive capsule on Signal

The mistake was requiring all of it. More guidance produced more visible design: panels, labels, metadata, technical decoration. It knew the ingredients. It did not know when to stop cooking (sorry).

The first pivot was from copying a visual vocabulary to naming the judgment inside it. Dark fields and document grids mattered less than the behavior beneath them: show evidence, use contrast decisively, let one thing lead. The technical signal layer became conditional. A proposal does not need to cosplay as a control panel because I like exposed electronics.

Scope caused the second pivot. I had treated it as a product reference when the useful part was restraint: one quiet field, compact type, direct content, controls only when needed.

MCMDS became less about applying traits and more about removing defaults. The top-level skill now reads the universal rules first. Taste, creative direction, and channel guidance enter only when the work actually needs them.

View the interactive capsule on Signal

The example changed more than another hundred rules would have. An asset review needs the work, the current selection, and the next action. Everything else can wait.

The main lesson is that specificity is not volume. “Use restrained typography” is vague. “Use size as the last hierarchy lever” changes a decision. References work the same way: useful as evidence, dangerous as templates.

There is still a contradiction. Codifying taste makes it repeatable, and repetition is how taste becomes wallpaper. I want the skill to recognize my decisions without freezing them.

Sometimes that means deleting the rule.

A layer called Scope

Sat, 01 Aug 2026 00:00:00 GMT

For years I’ve wanted a place for projects to exist without also needing to move every task and event into a new app. I like Reminders. I like Calendar. They hold the actual stuff just fine.

But neither knows that a few lists and calendars are all part of the same thing. They don’t know there are phases, or a decision that changed the work, or four dates that actually matter. I could put all of that into project management software, but then I have another inbox to check and another little administrative life to maintain (exactly what I need).

I wanted a layer for projects, not just dates and tasks.

View the interactive capsule on Signal

Something thin enough to sit on top of the systems I already use. Associate a few iCloud Reminders lists and Calendars, then add just enough shape to understand the work: phases, milestones, a timeline, and the decisions made along the way. Reminders still owns the tasks. Calendar still owns time. Project owns the relationship between them.

I was finally able to make it in a web app called Scope. The timeline above is from it: three months of a branding project, overlapping the way projects actually overlap, without asking the underlying systems to become something they aren’t.

Maybe the project management app I’ve wanted all this time was barely a project management app at all.

The dopamine hits of an LLM

Wed, 29 Jul 2026 00:00:00 GMT

The part I didn’t realize is that I am becoming addicted to Codex. I’m tinkering and building and generating materially interesting and exciting things! But I’m doing it so much. I’m not sure I’m actually using or enjoying what I output more than the act of making it.

I think that has a lot to do with the fact that I can tell Codex what I want and it will get started. I go off and “experience life” for 30 seconds just to come back to check in to see if it’s done. When it’s not… back to life for a fake minute. When it is.. look at what I have made! Dopamine. The trick is going to be to remember to leverage AI to automate so that I don’t have the performative 30 seconds of life experience but ongoing time away from screens and agents.

Is this a standalone, always-on Mac Mini? Or is that just another dopamine drip fang sunk into my neck?

Rejecting the machine

Wed, 29 Jul 2026 00:00:00 GMT

I’m feeling a heavy rejection of technology in interesting places. Instead of synthetic athleisure clothing, i’m seeking out natural, cotton fibers again. Instead of my tech bags and backpacks, i’m opting for simple totes. I feel myself searching for ways to differentiate myself from machines. And I welcome it.

I think where previously it felt efficient and good to align myself to the benefits of technology, i’m now turning my attention to the benefits of being a human animal.

Introducing Signal

Wed, 29 Jul 2026 00:00:00 GMT

Signal is a public working record: a place to publish thoughts while they still have some heat in them.

Some signals will be essays. Others will be fragments, interfaces, references, or small HTML experiments that are more useful to experience than explain. The format will follow the idea, not the other way around.

View the interactive capsule on Signal

Why this exists

Most finished work begins as something smaller and less certain: a question, an observation, a visual pattern, or a tool built to test a hunch. Signal gives those intermediate artifacts a durable URL without asking them to pretend they’re complete.

The archive will grow unevenly. That is part of the point.