Me, Myself & AI | A Rant on Publishing in the Agentic Age

The thatched farmhouses of Shirakawa-go in Gifu Prefecture sit beneath misty mountains and blazing autumn foliage, a quiet rural scene of the kind A Different Side of Japan has documented for a decade. The image opens Donny Kimball's 41st birthday rant on what it means to be a Japan travel publisher in the agentic age of AI.

Well folks, another year has come and gone, meaning that *checks notes* I am now 41. Yikes! What’s worse, the past 365 days have flown by in the blink of an eye. Just a year ago, I defiantly declared that I would be walking away from doing government gigs when I published this bombastic essay called Quitting OnlyFAMs, an obvious jab at the similarities between selling out and everyone’s favorite site with a blue logo. Little did I know that this decision would send me directly into a pelagic drift, floating without a map for making sense of why I was even still doing this. After all, if I am not going to sell something or take sponsored opportunities, what the heck am I even doing anymore when there’s a sea of other creators?

I mean, by all factors, there is little reason for me to keep investing a fifth of what I make into promoting Japan if there is no return, right? After all, travel costs, web hosting, and social media ads to promote the stuff add up quickly. While it’s not like I needed what I was making from FAM trips, those proceeds powered the pro bono consulting I was doing for a number of dying craftsmen and helped to cover the expenses I incurred along the way. Left with nothing but altruism on all sides due to not wanting to monetize the blog beyond reader-supported tips, I needed a reason to keep on creating. As I noted in the manifesto that I published back in May, the why was that the blog is a forcing function for all I do as a marketer in Japan.

Anyway, while derping about during that aforementioned dark night of the brand’s soul, trying to figure out what I was even doing, I also came to realize that I function far more like a one-man Condé Nast than a parasocial face-to-camera creator. Unlike all of my lovely Gram girlie friends, I do not appear on camera and twirl my skirt in front of seasonal allures like the cherry blossoms. Though I do, of course, post pretty pictures on social media, my feed is far closer to what you might expect from a glossy magazine. In retrospect, it’s comically easy to see why I struggled so hard on those government FAM trips that were built around performing for a camera, often in front of an entourage of chaperones.

On that note, I guess this is where I should pause for a brief moment and mention that, like with my prior trips around the sun, this article is another one of my birthday rants. While I hope to impart some categorical wisdom about what is happening with media, how information is being consumed, and how AI is changing all of that, this one is inevitably going to get a wee bit self-indulgent. If you’re just here for the Japan hidden gems, I suggest you go check out my prior work on all of the amazing off-the-beaten-path attractions this country has. So, with that disclaimer now out of the way, let’s get into what is happening to publishing in the agentic age.

Great Timing, Donny

The vermilion Bentendo Hall and arched bridge at Daigo-ji in Kyoto reflect in a pond ringed by fiery autumn maples. It is the kind of destination article travelers once found through Google search, before AI Overviews and the zero-click problem began answering their questions without ever sending them to a Japan travel blog.

Getting back to where we left off, on the one hand, understanding that I am more of a publisher and less of an “influencer” in the traditional sense brought a lot of clarity to why I struggled so hard with FAM trips. Alas, that realization could not have come at a worse time. You see, right now, AI is revolutionizing the way people interact with information. Under the old model of publishing, bloggers like myself might write a useful destination article, garner readers via search engines (and maybe do a little social media to promote the work), then move those eyeballs onto more permanent places like an email list or, at the very least, followers on the various social media platforms.

Nowadays though, the so-called “zero-click” problem is starting to raise its ugly head, whereby people get their answers without ever landing on a website. Between features like Google’s AI Overviews and users just outright asking ChatGPT or Claude, search traffic is taking a real beating. And lest you think it’s just a traffic problem for a niche blogger like myself, know that Roger Lynch, while still CEO of Condé Nast, told his teams to prepare for a world where search brings in no one. Not convinced? Well, how about the fact that Reddit, one of the intrawebz’s biggest discussion platforms, reportedly lost over a million search referrals per week on average in October 2026? Even the giants aren’t immune to this shift.

It’s rather funny when you look at dashboards like Google Search Console. On the one hand, my rankings in search engines are higher than they ever have been. People are seeing that blue link more than any time in A Different Side of Japan’s past. And yet, traffic continues to trend downwards month-over-month in a phenomenon that is comically called the “Alligator Graph.” Google is eating the click, and people are preferring to get the answer to their query without having to dig through the various sites themselves. Luckily, I have been able to prop things up with a bit more paid promotion, but the future of search looks closer to an LLM-powered chatbot by the day.

In a perverse way, publishers like myself, as well as bigger outlets, are now all producing bodies of text for machines. While the various AI models do come packed with a worryingly large amount of integrated knowledge, they still need references to cite as source material. So, in a weird way, writing now more and more exists so an LLM can regurgitate it back to someone looking for the answer to what to do in Japan during November. That, right there folks, is the zero-click problem that is wrecking a lot of businesses built on the back of an internet user base that still searched for stuff online. Increasingly, that is simply no longer the case.

Finding Readers Who Weren’t Looking

The sun rises over a sea of clouds surrounding Mt. Oe in northern Kyoto Prefecture, the legendary stronghold of the demon Shuten Doji. Readers searching for the yokai used to discover Kyoto by the Sea through this article, a route to the hidden gems of Japan that AI Overviews are now cutting off as paid social takes over the job of finding new readers.

One of the silver linings in all of this is that I’ve had to lean a lot more into doing paid social amplifications (largely on 𝕏 due to super cheap CPMs, so if you’re here from there, hi!) to offset the loss of search traffic from Google. I mean, my system relies on a river of new people encountering the work, be that on the blog or on social media, so the upside of all of this is that new entries to my ever-growing corpus of area guides and travel tips now get in front of more people who WEREN’T actively looking for them than before. Though I cannot fully offset the thousands of clicks lost to search’s decline out of my own pocket, it’s at least net new discovery of hidden gems in Japan that people might otherwise not have encountered.

One of the downsides of search and SEO for someone on a mission to introduce people to the other sides of Japan is that, by definition, someone needs to actually type the query into Google for me to show up. So, while I may get those clicks and people might find my body of work that way, I am not really doing much to overcome the awareness problem that the blog exists to tackle. Instead, I am just recycling the same existing intent and satisfying it via my own writing. In a strange twist, AI Overviews and the change in consumer behavior are actually helping to achieve my goals of bringing more people out into the countryside.

There is one notable exception, though, where search actually helps people discover places they never knew existed. My article on Shuten Doji, the “Drunken Demon” and one of Japan’s three most fearsome yokai, ranks on the first page of Google for the fiend. Thanks to appearances in pop culture like TYPE-MOON’s Fate series, there is actually ample search demand for an otherwise obscure keyword, and people searching for the mythological backstory end up discovering Mt. Oe and “Kyoto by the Sea.” A similar phenomenon happens with my article on Miyamoto Musashi’s birthplace. People come for the swordsman and leave knowing Mimasaka.

These days, even that unexpected discovery is becoming a casualty of AI. People just get the TL;DR on the mythological origins of a character in a popular anime series like Shuten Doji without needing to click any blue text. As a result, they leave with what they came for but never uncover anything else about an unknown region of Japan. What’s worse, AI Overviews often straight-up pilfer my writing word for word. Nowhere is this more obvious than in my series of seasonal articles like What to Do in Japan During May. I guess such is life in the zero-click age, where anyone who types that into Google is just going to get an AI Overview quoting me.

The Ghost in the Machine

The terraced mountainside hamlet of Ochiai in the Iya Valley of Tokushima Prefecture glows with autumn colors beneath cloud-covered peaks. One trip through the Iya Valley can now feed an area guide plus focused articles on Mt. Tsurugi and the vine bridges, which is how Donny Kimball uses AI as a writing tool without letting it invent the travel itself.

Given that I am a writer and that AI is ruining the craft in favor of plagiarized summaries, you’d think I’d be more against OpenAI, Anthropic, and their ilk, like many other artists who can be found bemoaning the technology all over the internet. To be honest with you, though, I don’t think I could be more bullish on AI. Sure, it’s completely disrupting the old ways that people discovered information, but it’s also the most transformative technology we humans have made since the invention of fire. And, seeing as anything that potent WILL be developed, I don’t see any reason to fight the currents of fate.

And let’s not pretend that AI is anything more than a tool. Just as digital cameras and Adobe Photoshop replaced the darkroom, ChatGPT, Claude, and whatever comes next are simply new ways of getting the work done. Give them to any random person on the street and the result will be unexceptional slop. Put them in the hands of a gifted writer with taste and years of lived experience with the topic they’re collaborating with AI on, and the result can be something else entirely. The software can help with the actual act of getting words onto the page, but it cannot magically supply the reason those words are worth reading.

For me, that distinction starts long before I open a chatbot. I still have to go somewhere, take my own photographs, learn about the place, and figure out what I actually think about it. Only then does AI enter the picture. I’ve even put together a rather ridiculous voice profile to help the models work with my writing rather than flatten it into generic travel-blog mush. Still, that doesn’t make them the source of the ideas. They’re helping me express what I want to say and do more with the material I’ve gathered, not inventing trips I never took or deciding which corner of Japan deserves another thousand words.

Take my recent travels through the Iya Valley, for example. There’s enough material there for a broader area guide that helps someone make sense of the region, but the same trip can also feed more focused articles on places like Mt. Tsurugi and the village of Ochiai. In the past, turning all of that into separate pieces would have been impossible, so what got written was the pillar piece, not the ancillary articles. With AI helping on the production side though, I can make better use of what I’ve already experienced and documented, not to churn out more AI-enabled slop for the sake of it but to give each destination the attention it deserves.

Then there are pieces where the original take is the whole damn point. My essay on Somei Yoshino cherry blossoms, for example, uses their uncanny uniformity as a lens for examining Japan’s imperialistic past. Sure, a chatbot can explain what these trees are, when they bloom, and where people go to photograph them. What it cannot do without being handed the premise is know why I chose to connect those dots in the first place. And if AI lets me produce more of that kind of work, why wouldn’t I use it? Even so, there’s a rather inconvenient catch: producing more doesn’t mean anyone will need to read the original.

Why the Hell Am I Still Writing?

A desolate landscape of twisted metal and crumbling debris stretches toward a glowing sunset, the header image from Blogging at the End of the World. It returns here as Donny Kimball asks why he is still writing about Japan when Google and LLMs answer travelers' questions by citing his articles without sending anyone to the source.

Making it easier to write more articles raises an uncomfortable question that all of my AI cheerleading above conveniently sidesteps. If Google or the LLMs are going to answer a traveler’s question by pulling from my pieces without sending that person to the source, why the heck am I spending so much time producing them? Sure, the tools can help me turn one trip into several pieces, but efficiency is a pretty lousy answer to a question about purpose. I raised a version of this dilemma in Blogging at the End of the World, and the problem hasn’t magically gone away just because I now have a machine helping with the boring parts.

That’s where I keep coming back to the fact that the articles themselves were always the point. I never set out to turn the blog into a funnel for courses or memberships. Every time I take a trip, photograph somewhere, and put what I learned into words, I add another piece to a growing body of work about Japan that I actually know. These days, though, some of the people consuming that work aren’t people at all. They’re machines pulling from it, citing it, and recommending places I wrote about. Someone might click through, but they might just get the answer and move on. The writing still did its job, even if the traffic graph resembles an alligator’s jaws.

I suppose the real shift is that the blog is no longer the only way for someone to access what I’ve accumulated. Maybe they read an article after spotting a paid post. Maybe an AI sends them to one of my destinations without the two of us ever crossing paths. And maybe they want the whole editorial take in one shot rather than digging through my ramblings. The underlying material stays the same, but its role is changing as fast as the ways people consume information. Rather than insisting everyone become a regular reader, I’ve started thinking about what forms would make all of this useful to someone who just wants to plan a trip and then disappear.

Duplicating Myself with AI

A row of small red daruma dolls lines a stone ledge at Katsuo-ji in Minoh, Osaka, with the vermilion Benten hall and autumn foliage blurred behind them. Like this lineup of identical daruma, the Our Japan AI trip planner lets Donny Kimball duplicate himself into a whole row of Donnys, each one built on a decade of A Different Side of Japan articles and ready to fix the badly planned Japan itineraries that generic LLMs produce.

So, as many of you know, earlier this year I launched an AI-powered trip planner as the first step in resurrecting the Our Japan brand that I birthed back during the pandemic. For those unfamiliar with the name, Our Japan actually started as an online marketplace for connecting Japan’s struggling craftsmen with people trapped overseas. Unfortunately, that little experiment never really got off the ground, thanks to a combination of reluctant makers, the logistical nightmare of international shipping, and the borders reopening. Quietly, I made peace with the marketplace’s demise and pivoted to pro bono consulting powered first by FAM revenue, and now tips for the planner.

Originally, I had envisioned building a community where people could come together and talk about their adventures in Japan. The problem, though, is that travelers aren’t necessarily looking to join some online group where everyone holds hands and sings Kumbaya like you see with communities centered around learning a skill on Skool. Most people just want rapid-fire answers to their travel questions before they arrive, and then they go ghost once their vacation is over. They aren’t looking for a lifelong relationship with the guy who helped them plan their itinerary. As strange as it feels to send readers to an AI agent instead of my own articles, Google is increasingly doing much the same thing anyway.

Now, you might wonder why anyone would bother using my AI trip planner instead of just asking ChatGPT where to go in Japan. Well, the trouble with generic AI is that knowing a bunch of facts about a destination isn’t the same as understanding how to put together a sensible itinerary. Take Yamadera and Dewa Sanzan, for example. Both are spiritual allures in Yamagata, but that doesn’t mean you can easily combine them into a convenient day trip, especially when relying on public transportation. A #BasicBitch LLM might see the shared prefecture and happily recommend both without considering the actual logistics. What makes the difference is having some editorial judgment about what works on the ground.

The Our Japan trip planner is not some revolutionary open-source model that I’ve cooked up in my kitchen but rather a layer of instructions and my accumulated writing that helps an existing AI make recommendations informed by my perspective on the country. Instead of immediately rattling off the usual suspects, it asks about your interests, previous travels, and practical constraints before suggesting places that might actually suit you. It’s not infallible, of course, and you’ll still want to double-check things like rural bus schedules but, at least so far, it’s been great at helping people connect with places they wouldn’t otherwise find.

A Decade of Diatribes, Distilled

Of course, not everyone wants to outsource their vacation planning to an LLM. Some people would rather sit down with a curated resource than spend an afternoon going back and forth with a glorified chatbot. And, seeing as I’ve finally come to accept that I’m more of a publisher than an influencer, I figured it was about time to do something a little more… traditional. So, in addition to resurrecting Our Japan with the AI trip planner, I’ve also brought the brand back to life with an actual guidebook. While I never wanted to really write one of these (as the blog was the whole freaking point), the agentic shift finally pushed me to publish one.

Available for free over at our-japan.store, the guidebook takes more than a decade of my travels and distills them into 345 pages built around 60 places I call trip anchors. These are destinations with enough to see and do in the surrounding area to justify building a significant chunk of an itinerary around them. The selections grew out of my ridiculously long, 28,000-word essay on where to go in Japan, which was itself an attempt to make sense of everything I’ve documented over the years. The blog is still the foundation, but nobody planning a vacation should have to parse a whole damn decade of my diatribes just to figure out where the heck to go.

Better still, the guidebook gives all of those aforementioned people discovering my writing through paid social a way to see the bigger picture from the jump. Someone might stumble across my Shuten Doji article because it appeared in their feed, learn about Mt. Oe, Amanohashidate, and Ine, and then wonder what else I’ve been writing about all these years. Unlike before, I can now point them to one curated resource that introduces the destinations, recommendations, and philosophy behind the entire A Different Side of Japan publication.

The Work Is the Constant

Donny Kimball, dusted with snow in a hooded jacket, stands among the snow-laden cedars at Togakushi Shrine in Nagano. The travel writer behind A Different Side of Japan still heads out in all weather to document the country, while reader support for the Our Japan AI trip planner and guidebook funds his pro bono consulting for Japanese craftsmen.

Back during the pandemic, Our Japan began as an attempt to connect Japan’s struggling craftsmen with people who might appreciate their work, but the marketplace never got off the ground. What stuck was my desire to help them, and that became the pro bono consulting that I now do. For a while, the FAM trips funded the journeys out to the countryside to meet with them, earn their trust, and get them set up to sell online through their own storefronts. When I walked away from government gigs, I wondered how deep I could cut into my own pocket, but with reader support for the AI trip planner and now the guidebook, I haven’t had to slow down.

While the blog has always been an unmonetized monster of a time sink, these two offers have been a much-needed way for all of you to contribute to the cause of keeping Japanese crafts around a little longer without me needing to beg for Buy Me a Coffee support. While I make my living as a freelance digital marketer and fractional CMO, reader support has enabled me to offer my time to people who could never afford my professional rates. And now, I can finally give back to those of you who generously chip in with something of my own, though both the Our Japan AI toolset and guidebook are available for free to anyone who wants them.

For now, these two things are also my best guess at what publishing can look like for the next year or two. Turning the blog into a forcing function for my work with AI has certainly been one hell of an education, but even I can’t see much past that horizon. With what looks like a fast takeoff ahead, perhaps the eventual interface won’t be a guidebook or even my current agent. Maybe people will simply talk to a system that draws directly from everything I’ve written. At that point, the articles might be more of a substrate than something anyone reads from start to finish. That’s conjecture, of course, as the world is changing faster than ever.

So that’s where I am at 41, after a year spent drifting and trying to work out what any of this is for. Oddly, I feel really in a good place right now. The agentic skills that I’ve learned through making AI mastery my reason for continuing have paid dividends time and time again in my digital marketing work. As we literally march towards the singularity, I feel really well equipped to face the future, both as someone who finally gets he’s a publisher and as someone who has the unique set of skillsets needed to help Japanese businesses as we waltz into the agentic age.

Until next time travelers…


Subscribe to My Newsletter

Donny Kimball
Donny Kimball

I'm a travel writer and freelance digital marketer who blogs about the sides of Japan that you can't find in the mainstream media.

Articles: 439