▸ Learnings // evergreen, by theme
What I've learned, shared.
Not thought leadership. Just what I've figured out, generalised from a life. No dates here, because good ideas don't expire.
Start here · AI in practicePart 1: Two AI tiers on one Mac Studio, and the routing between them
First in a series on running local and cloud AI together. The economics piece answered whether to own your AI. This one is the wiring: a 93.7 GB Flash Next model, Kev decision models, a hosted frontier model, and Hermes deciding which one answers. Measured on the machine, with the parts that are still rough named.
12 min read →Is it feasible to own your own AI models?
Everyone is preaching 'run your own model.' In the middle of a memory-price crisis, is owning your AI actually feasible? The honest math, and when it makes sense.
6 min read → AI in practicePractical AI: what to build first
A non-technical starting point for putting agents to work this week.
4 min read → AI in practiceAutomation that clears the operational noise
How small teams use AI to kill the busywork that used to need a whole ops hire.
4 min read → For foundersBootstrapping is the new default. Building alone is not.
Every founder I work with arrives excited that AI made building cheap. They are right about the economics, and usually wrong about who they hire next.
6 min read → For foundersHow VCs are starting to look at AI startups
What's actually changed in how investors size up an AI company, and what it means if you're raising now.
5 min read → For foundersAI gave you your time back. Where did it go?
AI promised us our hours back. Most of us are working more. The productivity paradox is mostly self-inflicted, and the fix is in your hands.
5 min read → For foundersWhy founders fear the build
It isn't the tech. It's not having the controls.
5 min read → PerspectiveIn the AI age, the generalist wins
The world says niche down. I think that's backwards. AI is the great amplifier of breadth.
4 min read → PerspectiveEveryone wants personalised learning. Almost nobody is talking about the data.
Personalisation is a data problem, and education's data is a mess: scattered, disconnected from real recruitment data, and mapped for humans in a world that is about to be automated.
6 min read →