The Long Tail was a profound cultural insight. When we created YouTube, Amazon, Roon, recipe websites and Netflix, the culture changed. When you give people a choice, they make a choice.
We went from the Top 40 to millions of songs. From Blockbuster to every movie ever made. From the local bookstore to all the books, all the time. Pick what you want instead of what other people picked.
There are still hits, but the majority of what we consume are titles that weren’t even available to us a few years ago.
It’s hard to believe this was a breakthrough idea only 21 years ago.
Now, with LLMs, it’s all going to change again.
That’s because it doesn’t matter how many cookbooks you own… Claude can invent a new recipe for you, one that’s never been seen before. You can do this today, the tech works.
Soon, it won’t matter if you have 400 Grateful Dead albums; AI systems will be able to generate live recordings from 1974 that never happened.
And eventually, Netflix will generate TV shows that no one has ever seen, for an audience of one.
The generative long tail will go on and on and on.
Most of the time, it’ll degenerate into slop and banality. The humanity of the movies, books and recordings it creates will fade, and we’ll be left with mindless airport music.
But that will evolve over time, and the arc will go in the other direction. Things that work, done more often. Combinations and evolving insights that move even faster than we can keep up with.
Slop and magic, in an eternal braid, a lot like the stuff humans have been producing for decades.
A golden road to who knows where.
January 31, 2026
Neither AI nor ads.
The problem started with search, and was weaponized by Amazon.
Display ads go back at least 100 years. A century ago, the idea was simple: Show up in a place where people are offering up attention and tell a story. The advertiser pays the media for the attention they’re using, and the consumer may find interest, amusement or, at the very least, tolerance for the fact that the media cost them less than it would without the ads.
Direct Marketing, named by the late Lester Wunderman, took the idea much further. This is action advertising. Measured. Direct marketing keeps score, and quickly evolves into something targeted and far more useful. Classified ads are direct, as is effective non-junk mail.
Google built an empire on this idea. When you do a search for left-handed widgets, the sellers of those widgets know that this is a great place to invest in running an ad. Google built an effective measurement tool, and soon, $1 ad buys became million-dollar ad buys. If it works, do it more.
But Amazon turned this into a tax. By incorporating barely concealed ads as search results, they corrupted the value of their shopping search engine. They confused consumers at the same time they stole margin from suppliers and ultimately increased the price that everyone pays for just about everything. Amazon sellers don’t want to buy more ads. In fact, they only do it because they have no choice.
When AI starts to incorporate ads, the corruption and lack of trust will only increase.
Permission marketing is the idea that we can deliver anticipated, personal and relevant ads to the people who want to get them. When I built this idea in 1992, it recognized that attention was precious and that trust was at the heart of every sustainable business. When you burn trust to get attention, you end up with nothing of value.
AI and search ads are about confusing and tricking users, burning trust and ultimately taxing any entity that is willing to pay money for attention–particularly those that feel they have no choice. The AI won’t recommend the best choice–it’ll point to the highest bidder. And the more they conceal that fact, the more money they’ll make. For a while.
AI companies say they are trying to be a trusted, independent tool. Ads are optimized when we benefit from the stories they tell. AI ads ruin both.
It’s unlikely that this will be regulated any time soon, but AI brands that want to earn the benefit of the doubt would be wise to do what early magazines, modern TV and even the Yellow Pages did–make it crystal clear that the ads are the ads. They work best when we know what they are and we want them there–and worst when we forget that search is a tool, not another chance to hustle customers.
Of course, with alternatives just a click away, they’re not going to have much luck clearly labeling the ads as ads. That’s why Google ultimately corrupted their clear division between the two–every time you make the line more blurred, revenue goes up.
We don’t need more hustle. We need more trust.
January 30, 2026
Every important medical innovation of the last five hundred years has been ridiculed, undermined and ignored by doctors and the medical establishment. Every single one. Hand-washing, antibiotics, and the dangers of smoking, just to pick a few.
This happens in any system where expertise is valued and where change might be risky.
Just because the establishment doesn’t like it doesn’t mean it’s not a good idea.
(On the other hand, there are plenty of bad ideas the establishment doesn’t like as well).
January 29, 2026
There have always been two sides: Hiring people to do tasks and jobs, or hoping to be hired to do those tasks and jobs.
The difference now is that it’s increasingly difficult to find a good job to get hired for, and easier than ever to be the person who hires an AI or a person to do a task.
Our understanding of ‘entrepreneur’ needs updating. It’s no longer mostly about raising money, buying a factory, taking financial risks or even out hustling for sales. Instead, it’s an emotional shift–to walk to the other side of the hiring line.
If you can hire an AI to do tasks that solve problems for other people, you can do it again. If you do it well enough, you can do it for a living.
Projects, not jobs. Hiring, not being hired.
January 28, 2026
When energy is cheap, people build buildings that are poorly insulated. It’s faster and cheaper in the short run.
In the long run, though, insulation always wins. You invest in it once and get the rewards forever.
And of course, this is true for all things, not just buildings.
Learning to do it right is cheaper than using effort to overcome our laziness every single time.
January 27, 2026
Most of us are terrible at some things.
Lack of skill, focus, practice, care or just temperament means that we don’t do the task as well as we might. This might be anything from promptness to conflict to high-stakes negotiation. It could include filling in forms, taking notes or brainstorming innovative ideas. Perhaps it’s living with uncertainty…
Once you realize your areas of terrible, choices arise:
- We can choose to put in the effort to become not-terrible.
- We can avoid the tasks, automate or delegate and simply avoid our terrible areas.
- When asked, we can announce we’re terrible, setting expectations so we don’t let folks down.
The one that’s probably worth avoiding is: Accepting tasks and making promises and then quietly doing a terrible job.
January 26, 2026
One friend recently opened a bookstore instead of a bookmobile.
Another is investing two years of his life to open a restaurant instead of a series of pop up dinners.
And a third is buying a boat instead of chartering one.
It’s easy to see why. A real bookstore has a lease. They post their hours. It’s solid.
And a real restaurant, the kind we’ve all been to, looks, feels and smells like a restaurant.
Leaving the ridiculous economics of boat ownership aside, it’s worth taking a look at the first two.
The key questions are:
- what asset are you building?
- what is the purpose of the thing you’re building?
- who is taking a risk and who is getting the upside?
If the asset of the future is trust and attention, then it’s easy to do a new calculation. The purpose of getting a lease and a place on Main Street was to momentarily get the attention of passers by–and to earn their trust. After all, you’ve got a building.
But when we shift to a permission asset, when we cherish attention and use it to build trust over time, being on Main Street might not be the best way to achieve this.
The landlord gets paid regardless of whether the space does the job, and the upside is limited by the size of the space.
What happens to the life of a chef when they have 6,000 people who eagerly read their popup updates? When each person comes to two or three dinners a year, each held at a fascinating location, rented just for the night…
Instead of finding diners for their restaurant, they could create dining experiences for their diners.
The bookstore? Imagine a route, a series of partnerships with 50 or 100 organizations that value information and exploration. A curated selection of 400 books, ready to roll out at a community center, church or school…
In both cases, the hard work is earning attention and trust. It’s easy to avoid that when you’re passively sitting in a leased space waiting for people to come to you.
This applies to musicians, filmmakers or anyone who seeks to create magic.
Earn attention and trust. Spend time and money on that, and the rest will take care of itself.
January 25, 2026
That’s easy advice and a fine goal.
Except… if you look at the last hundred years, we haven’t seen many useful advances in mousetraps, despite the number of people who have tried. It feels like an infinite market, so it attracts a lot of entrants.
You probably won’t come up with a better mousetrap. But you might find the empathy and focus to find a small group of people with a more specific problem and solve it for them in a way that earns you trust, traction and word of mouth.
That’s enough.
January 24, 2026
Who is Nicole Bennet and why does she keep calling me?
A few times a day, a voice pretending to be someone named Nicole rings my cell, and in a petulant, entitled voice, insists she’s calling me about a loan that I never applied for. I’ve never interacted, I block each number, but the calls keep coming.
AT&T certainly has the technology to block calls like this, but they don’t have an incentive to do so.
At the same time, many subscribers to this blog don’t receive their emails because Google has a clear incentive to move the emails to the promotions folder. Google benefits by forcing marketers and writers to pay them for access to folks’ attention.
Merged medical practices have an incentive to charge patients more, push doctors to work even more unreasonable hours and cut corners on medical outcomes.
Instagram has an incentive to make people feel as though they’re falling behind unless they adhere to the algorithm. And Amazon has an incentive to denature the business model of most of their merchants by charging for advertising, even though they know the ads serve neither merchants nor customers.
Tim Wu is our best explainer of how, without boundaries, networks always spiral out of control. His new book is twenty years too late or exactly what we need right now.
But we don’t need any more Nicole Bennet, thanks.
January 23, 2026
42 years ago, Apple’s 1984 ad ran on the Super Bowl. Once. It’s generally considered the most effective ad of its kind, creating a legend and also a trap.
Was this ad the reason the Mac is still around?
Or was it Regis McKenna’s work in getting Steve on the cover of more than 20 magazines the month it launched?
After all, they say that getting the word out is the key to marketing.
That’s not what did it.
It was Guy Kawasaki’s tireless year of evangelizing the platform to software developers, creating an ecosystem that made the Mac useful from the start.
And it was Susan Kare’s and Bill Atkinson’s unreasonable standards that made the experience of using the Mac unlike its alternatives–an advantage that has lasted half a century.
Hype is a trap. Better is better.
January 22, 2026
