“In virtually all markets, there are far more niche goods than hits. That ratio is growing exponentially larger as the tools of production become cheaper and more ubiquitous.”
- Chris Anderson, The Long Tail, Chapter 4
In 2006, Chris Anderson wrote about the long tail of commerce, essentially how the internet suddenly made it profitable to sell niche products to tiny audiences. Amazon could stock books that Barnes & Noble wouldn't touch. Netflix could offer films your local Blockbuster never heard of. The reason was that the economics had shifted. Suddenly, serving the weird and wonderful made sense.
Twenty years later, I’d argue we're witnessing a similar cycle, with generative AI as the catalyst. The key difference is that today it's not just about selling to the long tail anymore, it's about solving for the long tail.
Let me explain…
The Feasibility Flip
Here's the interesting stat to start you off on this rabbit hole with me: 38% of startups are now founded by solo entrepreneurs, up from 22% just a decade ago. Then suddenly this year the famous Silicon Valley start-up accelerator YCombinator stated they think that the first $100 billion 10-man start-up is on the way. This isn’t just a growing trend line, it’s an accelerating one.
There’s a classic innovation framework that helped us frame new venture strategy risk at Accenture and in start-ups, DVF: Desirability, Viability, and Feasibility:
Desirability: Do people want it?
Feasibility: Is it possible to make it?
Viability: Can you even make money from it?
For decades, lean start-up methodology (made famous by Eric Ries and Steve Blank) would have you believe that desirability risk was the second act’s core villain. You see, feasibility and viability were the killers. You might have a brilliant idea that people wanted and would pay for, but could you actually build it? Maybe not… not without a team full of domain experts, serious funding, months of development and even by then you’d face the capricious wrath of lady luck if something went awry. But then maybe you could build it but the unit economics just didn’t make sense.
Now, AI’s torn down the wall.
What used to require a 10-person team and six months can now be prototyped by one person in a weekend - just look at lovable’s hackathon success. Solo founders are building businesses generating monthly revenue using AI tools that didn't exist two years ago. (e.g. Base44)
The ‘so what’: Return of the Artisan
I had to weave in a bit of a Tolkien-esque comeback story in here. This is what irks me on the current narrative I’m seeing: everyone's talking their hands about AI replacing jobs but it’s the straw man. The headlines are expectedly sensationalist, and god do they grab the attention (e.g. the job market shrinking for entry level roles) but they distract us from two key points I’d make:
Historically, the vast majority of people did NOT have a single, linear career
There are now so many more opportunities for the average person, not fewer
To address the first point:
“Anthropology shows, in most cultures, occupational identity was fluid. Only a minority of premodern societies had rigid caste/profession systems; even then, most laborers were generalists, shifting between farming, trade, and craft as required by season or necessity.” — American Anthropological Association, 2025
The linear, lifelong, employer specialisation of careers is extremely unusual. It was mostly reserved for the highly skilled in one department, only. This isn’t to say this was the cream of the crop focusing on one thing, far from it. A cursory scroll through the polymath on wikipedia would tell you this. From Imhotep to Aristotle, Da Vinci to Descartes, those who were given tools to succeed and had innate curiosity, could follow their line of questioning regardless of category boundaries and social objections.
However, notice for a second on that list of polymaths the obvious lack of names from industrial revolution on? Adam Smith’s rhetoric of specialisation was powerful, but it treated humans like cogs in a production line - something I’d happily argue with anyone that we’re not.
Yet, today if you say in a job interview that your career has spanned 3 industries, 4 hobbies, and that you still aspire to be an astronaut you’re likely to just be the latest case of ghosting.
This is how dreams die.
We find it difficult to consider intelligent people plying their hand at a multitude of different fields in their life - it signals a lack of focus, a lack of seriousness about the profession they’re applying for, etc.
However, we’re at a real inflection point.
ChatGPT, Claude, Gemini, Grok, Mistral, etc. just name your model - we may love to obsess ourselves with tips and tricks, ways to use LLMs better with better prompts, which model is better for this vs. that but ALL can now do an extremely decent job of giving the average person the intelligence of the entire library WITH librarian at their fingertips.
This is to say, the foundation of knowledge we now all have is ridiculously advanced. As Sam Altman has self-promoted, we have (or soon will have) a PhD researcher in our pockets. This means the base level of knowledge we can arm ourselves with is only a few questions into our phone away. Rather than worrying about the glass half empty, I do wonder why more don’t talk about the glass half full - the new possibilities that come with this.
The reason the narrative of job loss irks me so much is that most are focusing on loss aversion, but it’s not like that new AI productivity is just going to mean we lose our will to achieve, to do more. It’s instead going to become our new foundation to reach and strive for more in every direction.
The opportunity for an average person to let their curiosity take them like the polymaths of old is that much easier and only going to get more so. In other words, this is a welcome return to a more varied, more human approach to work in my eyes.
This isn't job replacement. It's job amplification.
And more importantly, it's enabling a return to something we lost in the industrial revolution: the artisan economy.
In software engineering, a front-end engineer is suddenly pursuing his full stack project as a hobby, while being able to write about it and make youtube videos, and making money from it!
Before factories, craftspeople solved specific problems for specific customers. The village blacksmith didn't make generic horseshoes, they made your horseshoes for your horse. But industrialisation changed this. The economic seduction of mass production meant one-size-fits-all solutions were in vogue.
Now we're seeing artisans return, but they're digital artisans armed with AI. Startups have demonstrated that rapid MVP development and laser-focused niche applications can swiftly scale Monthly Recurring Revenue (MRR) into six - sometimes seven - figures (Cactus: AI copilot for solopreneurs | Y Combinator shows that the market is being recognised as growing).
The Niche Explosion
What's genuinely mind-bending is the sheer number of personal problems that are now economically solvable.
Consider the explosion in micro-SaaS. For example, “Kattalog” - built by one person in two months while holding down a full-time job - Kattalog uses generative AI for product photography, letting small e-commerce sellers instantly create styled, high-quality images. The founder built the working MVP by leveraging AI APIs, no-code tools, and rapid prototyping platforms, then shipped on Product Hunt to reach thousands of potential users without a marketing team or VC budget. What’s striking is that this tool, which would have required a team of developers (and a studio) just three years ago, now went from idea to paying customers at micro-scale, serving the global long tail of Shopify and Etsy stores - all by a single founder, powered by AI and low-code. (read more here)
But the pattern is everywhere:
Legal document generation for small businesses
AI voice assistants for private chefs handling bookings
Transcription tools for podcasters
Industry-specific content calendars for real estate agents
Each of these represents a problem that always existed but was never worth solving at scale. The market was there, but fragmented. The need was real, but not concentrated enough to justify traditional development costs.
The Death of the Generic
Here's the kicker: as AI makes it easier to build solutions, generic products become less valuable, not more. Why use a generic CRM when you can have one perfectly tailored to your photography business? Why settle for broad marketing automation when you can have tools designed specifically for B2B SaaS companies selling to healthcare?
The shift is away from broad, general-purpose AI tools towards specialised applications that cater to specific market segments or address particular problems. The economics have flipped. Building for everyone is now harder than building for someone specific.
This is what many tech giants are yet to get. They're still playing the old game - trying to build the everything app, the universal solution. Moats will shift though, and when the incumbents move their resources are also far more plentiful, so be sure that the big opportunities will still be taken by the big players, eventually.
But when a solo founder can spin up a competitor that serves a specific niche better in a weekend, they can begin their personal relationship with their audience. They can become the independent artisan blacksmith.
The New DVF: Distribution, Voice, and Focus
The old innovation framework needs an update. In the AI era, Desirability, Viability, and Feasibility are table stakes. Everyone can build something people want that makes money. The new challenges are:
Distribution: How do you find your specific tribe in an ocean of solutions?
Voice: How do you speak their language better than a generic tool ever could?
Focus: How focused do you need to be to appeal to your audience?
The winners in this new economy won't be the ones who build for everyone. They'll be the ones who understand that companies are reaching as much as $10 million in revenue with teams of less than 10 people Y Combinator. The metric that matters isn't headcount or total addressable market. It's revenue per employee.
The Philosophical Shift
There's something deeper happening here that most miss. We're not just changing how we build businesses. We're changing what business is.
The industrial model was about standardisation, scale, efficiency. Find the common denominator and optimise for it. But that model only worked because customisation was expensive. When customisation becomes cheap - or free - the entire logic inverts.
Suddenly, the question isn't "How can we serve the most people with one solution?" It's "How many unique problems can we solve perfectly?"
This is why I think the "AI will destroy jobs" narrative is so backwards. AI doesn't destroy the need for human creativity and problem-solving. It amplifies it. It makes it economically viable to solve problems that were always there but never worth addressing.
A therapist can now build tools specifically for treating grief in teenagers. A nutritionist can create meal planning apps for people with rare allergies. A mechanic can develop diagnostic systems for vintage motorcycles. These aren't billion-dollar markets. But they don't need to be when you can serve them with near-zero marginal cost.
The Wayfinding Problem
Of course, this explosion of possibility creates its own challenge. When anyone can build anything for anyone, how does anyone find anything?
This is the wayfinding problem of our age. We're creating millions of perfect solutions for specific problems, but connecting problems to solutions becomes the bottleneck. It's not enough to build for the long tail - you need to help the long tail find you.
This is where I think the next wave of innovation will come. Not in building more AI tools, but in creating the navigation systems for an AI-saturated world. The marketplaces, the recommendation engines, the trust networks that help a small business owner in Dublin find the perfect inventory management system built by a solo founder in São Paulo.
The Bottom Line
Chris Anderson was right about the long tail, but what he couldn’t have predicted is how this trend would continue. It wasn't just about selling niche products. It was a preview of a world where every niche need could have a niche solution. Where the economics of personalisation beat the economics of scale.
The evidence is already here. Companies with more than $500 million in annual revenues are using gen AI throughout their organisations but often in spectacularly uninformed fashion - with a 95% failure rate. But the real innovation is happening at the edges. Solo founders are building million-dollar businesses solving problems that VCs would have laughed at five years ago.
We're not just serving the long tail anymore. We're solving for it. And that changes everything. Problems that were invisible because they were unsolvable are suddenly visible everywhere.
So here's my question to you: What problem have you always had that nobody's ever bothered to solve? Because I guarantee you, somewhere, a solo founder with an AI co-pilot is probably working on it right now. And if they're not? Well, maybe that's your cue.
Want more takes, tips and tricks for navigating the future of work? Subscribe to Wayfinder Notes. Thanks for reading! Would love to hear your thoughts in the comments.
AI is making old, overlooked problems worth solving not just creating new ones.
One thing I kept thinking about though it's the ongoing cost of keeping these niche products alive. Building is much cheaper now but maintenance, hosting support, and finding customers still take time and money. Maybe that's the next problem AI needs to solve.