Developer's Notebook

We updated AIOS so a missed commitment can be moved to another day, reminders wait for a sleeping phone, and AI usage cards link to your provider dashboard

We expanded AIOS Skills to level 50 with public profiles, unified login across four clients to a single password, and added yearly task recurrence.

Two weeks in which the family grew to five products, payments went live, and the AI Council opened to anyone.

Three months, more than a hundred releases, and one idea that never changed.

Two products instead of one, the first standalone app, and the fifth major version.

Eighty releases in eighteen days, almost all of them from using the product every day and fixing what got in the way.

Six weeks of foundations: accounts, privacy, a grammar for plain sentences, a real calendar and a test suite — ending with the open beta on 13 August.

The first two weeks: an idea written down on 17 June, and a system running in production on 2 July.

This is a course, not a list of links.

In one line: the AI you talk to is a program that is extremely good at guessing what comes next, and everything it can and cannot do follows from that one

In one line: there are about five names worth knowing, they leapfrog each other every few months, and the useful thing to learn is what KIND of choice you

In one line: ask what it costs you to be wrong, and whether the text is allowed to leave your machine.

In one line: most disappointing answers are not the model being stupid, they are a request that left out something the model had no way to know.

In one line: four things actually go wrong.

In one line: you can download a model and run it on your own PC — free, private, no internet — and it is genuinely good enough for a great deal of everyday

In one line: give a model the ability to search, read files or run commands and it stops being a thing that answers and becomes a thing that does — which i

In one line: you now know enough to use these tools well.

Start here whatever your job is.

This is the track where the tooling changes fastest and where the gap between people who use it well and people who use it badly is widest.

The honest summary of this track: AI is very good at two specific parts of planning and actively harmful at a third.

This is the track with the sharpest trade-off in it.

The creative tools are the most visible part of AI and the most oversold.

The least interesting track and probably the one that gives you back the most hours.

This is the last track and the one the others lead to.

Apache can accept a perfectly valid security configuration and deliver none of it.

A contact form that silently stopped sending, what actually broke, and the setup that survives a rebuild.

Notes from tuning a static site, including the changes that measured as nothing.

A stylesheet was renamed.

Front-end development is the part of a site a person actually touches.

A working environment, a codebase you will keep, and the commands that get Linux from a fresh install to ready.

A short reference for what HTML is actually doing, written for the point where you have opened a text editor and want to know what you are typing.

The commands, in the order you actually need them, for a box running Apache in front of a Python app.

A working reference, not a tutorial.

Notes on what the job actually is, why I started with web development, and how to pick the next language rather than guessing at it.

Programming is about getting to a minimum viable solution quickly and then refining it.

How a single repeated action turns into a capability you can point at a problem you have never seen before.

Most learning plans are written on a good day and then tested on a bad one.

A prompt is an ordered chain of facts the model reasons from, not a question with background attached.

The setup steps for the Google Cloud Natural Language API, from creating the project through to calling it from a FlutterFlow app.

A ten-day plan for getting a working AI-backed mobile app in front of users, using FlutterFlow for the interface and a hosted model API for the intelligenc