People look at my project list and see chaos. A pub guide. A wearable tech aggregator. A Nottingham Forest quiz. A Japanese restaurant directory. An AI art project that reimagines brands in 1970s retail aesthetics. A culture news terminal. A taste-scoring system. A tube exit guide. What connects them? At first glance, absolutely nothing.
And that's exactly the point.
For years I felt slightly embarrassed about this. The conventional wisdom in both advertising and tech is to specialise. Pick a lane. Go deep. Become the person who's known for one thing. Every LinkedIn thought leader will tell you the same: niche down. Find your thing. Own it.
I tried. I genuinely tried. But my brain doesn't work that way. I get fascinated by a problem, build something to solve it, and then my attention drifts to the next fascinating problem. The pub guide led to the restaurant directory. The culture aggregator led to the taste engine. The Forest quiz was born from a Saturday afternoon conversation about whether anyone could name Forest's 1979 European Cup squad. Everything sparks something else.
The generalist advantage
David Epstein wrote a book called Range that changed how I think about this. His argument, backed by extensive research, is that generalists outperform specialists in complex, unpredictable domains. The specialists win in narrow, well-defined environments - think chess, or golf, where the rules are fixed and repetition builds mastery. But in messy, real-world domains where the rules keep changing and the problems are ambiguous? The generalists dominate.
Building products is one of those messy domains. There's no fixed rulebook. Every project brings different users, different constraints, different aesthetics, different problems to solve. The person who has only ever built one type of product sees every problem through the same lens. They reach for the same solutions. They optimise within a narrow band.
The person who has built across categories - pubs AND wearable tech AND culture AND football AND Japanese food - brings connections that specialists can't see. They have a wider pattern library. They have more analogies to draw from. They've experienced more types of failure and more types of success.
One thread, many fabrics
It took me a while to see the common thread. For months I thought there wasn't one - that I was just a generalist with a short attention span, jumping between interests like a browser with too many tabs open. But then I started looking at the methodology rather than the subject matter, and the pattern became obvious.
Every project follows the same approach: find a gap, apply taste, curate the experience, ship it fast.
The London Pub Guide exists because every existing pub resource is either an exhaustive database with no editorial point of view, or a listicle written by someone who has never actually been to a proper local. The gap was taste-led curation in a space drowning in data. So I built a pub guide that reflects how I actually think about pubs - the atmosphere, the neighbourhood feel, the kind of pint you're going to get.
Oishii London exists because Japanese food in London is extraordinary but poorly mapped. The gap was a curated directory that treats Japanese dining with the specificity it deserves - not just "Japanese restaurants" but ramen versus izakaya versus omakase versus kissaten. Same methodology. Different content.
CultureTerminal exists because the culture news landscape is a firehose of noise. The gap was editorial filtering - someone deciding what actually matters at the intersection of advertising, design, fashion, and brands. Same methodology. Different content again.
The subject changes. The standard doesn't.
Cross-pollination
Here's where range becomes genuinely powerful: the learning compounds across projects in ways you can't predict.
Building CultureTerminal taught me about content scoring - how to weight and rank articles by relevance, recency, and source quality. I applied that thinking directly to the pub guide, creating a system for surfacing the most relevant recommendations based on what a user is actually looking for. That cross-pollination would never have happened if I had only built culture products.
Building Little London, the family activities directory, taught me about local SEO and location-based content organisation. I applied those lessons to Oishii London, structuring the Japanese restaurant guide in a way that actually helps people find what they need by area and cuisine type. The parenting directory made the restaurant guide better. Nobody would have predicted that connection.
Building Modern Retro, the AI art project, taught me about visual identity and brand aesthetics in a way that fifteen years of advertising strategy never quite did. Generating eighty-six brand images forced me to think deeply about what makes each brand's visual language distinctive. That understanding now informs every design decision across every project.
Range creates compound learning. Each project makes every other project better.
The T-shaped builder
There's a concept in hiring called the "T-shaped person" - someone with broad knowledge across many areas and deep expertise in one. I think about it slightly differently. For me, the breadth comes from curiosity. The depth comes from caring about details.
Every single project, regardless of subject, gets the same obsessive attention to typography, spacing, colour, and user experience. The Forest Quiz looks as considered as CultureTerminal. The pub guide has the same level of design craft as the taste engine. The surface varies wildly - football, food, culture, wearables, art - but the standard of execution doesn't.
That consistency is the depth. Not depth in one subject, but depth in craft. In the care and attention that turns a side project into something that makes people say, "Wait, you built this?"
The breadth is what makes the work interesting. The depth is what makes it good.
Why this matters now
We're entering a world where AI makes specialised execution cheap. If you need a specific type of app built, or a specific type of content created, or a specific type of analysis run, an AI tool can do it competently. The specialists who built their careers on executing one thing well are going to face real pressure.
But the generalist who can connect ideas across domains? The person who sees that a content-scoring system designed for culture news could also improve a pub recommendation engine? The builder who understands audiences in parenting AND food AND fashion AND sport AND technology? That person is irreplaceable. Not because they do one thing better than anyone else, but because they see connections that nobody else can see.
In a world that has always rewarded specialists, range feels risky. It isn't. It's the most powerful position you can have - especially now, especially as tools make the execution layer accessible to everyone. When everyone can build, the advantage goes to the person who knows what to build and why. That knowledge comes from range. From having seen enough different problems, in enough different domains, that you develop an instinct for what matters.
Stop apologising for being interested in everything. Start building across everything. The connections you make will be the ones nobody else can.
