About
Most people are using AI like a slightly faster Google.
Most of them are wondering why they're not pulling away from anyone.
The gap isn't sharper prompts. It's that almost nobody is training the layer underneath, workflow design, judgment, taste, the parts of the work that compound while everyone else is still typing into a blank box.
That's what this publication is about.
James Tandy · Sheffield, UK
I'm James. Born and raised in Sheffield. I write James Tandy for ambitious digital builders, solo founders, content creators, agency owners, course makers, indie operators. People building income, products, or audiences online with AI as the unfair advantage that lets them compete with teams ten times their size.
I'd been building with AI for seven years before I started writing about it. I started in 2019 with GPT-2, when the model was barely usable and getting anything meaningful out of it took real effort. I've been a practitioner through every wave since: GPT-3, ChatGPT, Claude, agents, tool use.
Before AI, I spent a decade running SEO as a side practice alongside my full-time job as a Data Centre Operator. Ranked blogs internationally. Got one to 30,000 monthly visits at its peak. Most of my twenties were spent learning how to build digital assets that compound while you sleep, long before AI made that idea fashionable.
Two parallel tracks. Same underlying lesson: the people who win aren't the ones with the newest tools. They're the ones who design better systems around them.
The mission
Why I started this publication
I grew up obsessed with biology. Got an A* in it at school, then kept reading neuroscience and psychology in my own time long after the exams were over. Most of what I read kept landing on the same uncomfortable conclusion:
People aren't held back by what they don't know. They're held back by what they feel.
Education isn't the bottleneck for most lives. Emotions and meaning are. I've watched people who were, by every measure, capable, intelligent, hard-working, talented, stay stuck for years because something underneath the surface was in the way. I've been one of those people myself.
I've also seen what happens when that something shifts. People go from stuck to thriving in months, not decades. The talent didn't appear. It was always there. Something underneath got out of the way.
For most of human history there's been no real way to help with that part at scale. Therapy is expensive and slow. Books help if you're already self-aware. Most people aren't, and most of the people who would benefit most never get the help.
AI is the first technology powerful enough to actually do something about that. Not as a replacement for humans, but as something that can meet someone exactly where they are, every day, with no judgment, no waiting list, and no cost to ask another question.
That's the version of AI I care about most. AI for human potential.
The thread
How that connects to the publication
There's a thread running through Anthropic's founder, Dario Amodei, that hit me the first time I read about it.
He moved from biophysics and computational neuroscience into AI partly because the brain was too complex to crack with conventional methods. The same questions I'd been chasing since school, how the human system actually works, why people get stuck, what it would take to help them move, turned out to be questions some of the most ambitious labs in the world were now chasing too.
That's not a coincidence I want to sit on the sidelines of.
The publication is one expression of that. The audience is digital builders specifically, because that's the audience I know best and the one I can help most directly. But the mission underneath, AI for human potential, runs broader. More projects, more channels, more directions to come.
The work
What you'll find here
Four pillars, all in service of one promise: build unfair leverage with AI.
How to combine AI tools, agents, and models into systems that compound. Build guides you can copy this week, not theory.
Decisive comparisons. Verdicts. No 100-tool roundups, no affiliate-driven slop.
Frameworks for leverage, judgment, and where AI shouldn't touch the work.
Public testing. Failed runs. Real numbers. Receipts, not hype.
The angle is honesty. I'm becoming an AI power user in public, posting what works and what doesn't, sharing the systems that hold up and the ones that fall apart, so other people can compound from where I've already got to instead of starting from scratch.
The fit
Who this is for, who it isn't
It's for you if you're running a small or solo digital business and want signal in an ocean of AI noise. If you're trying to operate at a level your team size shouldn't allow. If you'd rather build one workflow that compounds for two years than copy ten prompts that work for ten minutes.
It's not for enterprise buyers, AI researchers building foundational tech, or anyone treating AI as a productivity hack. It's bigger than that.
Off the page
Outside the screens
Family. Nature. DIY projects. Boxing, because there's no better lesson in strategy than two people trying to outthink each other with everything on the line. Chess, because there's no luck in it. Sheffield is home and most likely always will be.
The Final
Two.
When AI does the work, only two skills hold their edge: creativity to direct it, and inner work to stay human with the people still here.
Written by James Tandy · For operators, builders, and thinkers who want unfair leverage with AI.
FREE PLAYBOOK · 38 PAGES
Get The Final Two. Plus a sharp essay every Sunday.
A 38-page playbook for solo founders, indie operators, and creators. The argument, the practice, and a 30-day plan you can start Monday.
Or read the latest → Workflows · Tools · Strategy · Experiments