Blog

Using AI for real work: my experience, my take

The posts below are sorted by where you are, not by topic: still deciding if this is for you, building your own AI agent or assistant, or running one already. Start at whatever’s nagging you.

New here? The homepage is the two-minute version of who I am and what this is about.

how-these-posts-are-made.md

## Vibe writingEvery post here starts as me talking. I know the story, so I say it out loud, and my AI shapes the ramble into a draft. I fact-check it, fix what’s off, and delete the stupid sentences before anything ships. No effort to beat AI detectors – the goal is a useful, true story, fast. read more

Phase 1

Is this even for me?

This part is about using AI at all, before you build anything. The doubts I hear most, and my take on each, from my own time with it.

Will it just make things up? Can I trust it?

AI hallucinations: why your AI lies to you and how I catch it

read it

Is ChatGPT accurate? What it gets wrong and how I check it

read it
Is using it cheating? Will I get flagged?

Is using AI cheating? The six fears I hear and what I think

read it

AI writing detectors and false positives

read it

These posts were drafted by AI. Here’s why they’re still mine.

read it

Vibe writing: how I draft blog posts by talking

read it
What does it cost, and which tool?

What I actually pay for AI (and what I canceled)

read it

I tried Claude Cowork and switched to Claude Code – an honest comparison for non-coders

read it
Can someone like me do this?

What learning AI at 51 actually looked like

read it

AI for scientists and researchers

read it

ChatGPT for content creation: what I learned drafting a whole blog with it

read it
What is she actually doing with hers?

What working with AI agents actually looks like: a real day of client work

read it

Everything I do, and the one system underneath it all

read it

AI in Obsidian: how I actually use it

read it

Diary of an AI Operator: the day-by-day record of the system getting built

the diary

Phase 2

You want an AI agent or assistant?

This is how you can build one – and the first questions that come up after you’ve built it.

Okay, I want to build one.

Build your own AI second brain – the whole build, free, in order

the build path

How I built my AI wiki (wrong turns included)

read it
It won’t remember me.

How to give Claude lasting memory of your work

read it

Your AI forgets. Here’s why.

drafted · in review
How do I get my own stuff into it?

How I built an AI that actually knows my work

read it

I talk to my AI all day. It never talks back.

read it

Why I don’t dump all my files into Claude

drafted · in review
Connect it to things, safely: passwords, API keys, your work.

What’s connected to my AI, what isn’t, and what died

read it

Client work and AI: how I keep proprietary information out of chatbots

drafted · in review

Did you accidentally agree to let AI train on your work?

drafted · in review

Phase 3

Your AI assistant is running. Now make it better.

Some ideas for improving it, and what to do when things go wrong.

Stop babysitting it.

A skill is just a markdown file (and most of them are garbage)

read it

Claude Code subagents – the honest use case

read it

Your context window: compact vs. handoff for long sessions

read it

/loop, /schedule: stop babysitting Claude

coming soon
It keeps making the same mistake.

Your frustration is data: stop your AI repeating the same mistakes

drafted · in review

Also worth a look once yours is running: a real day of client work and everything I do, one system – the same two posts from phase 1, because “what could mine do next?” is this phase’s version of “is this real?”

the-newsletter.md

The newsletter

Subscribers hear first

Every week or so, one letter: what I built, what broke, and what I’d tell a smart, busy person to try next. New posts land there before they land here.

Double opt-in · unsubscribe anytime · GDPR-compliant