My mother-in-law was at our house a few weeks ago, and at some point in the visit she told me she'd been reading Dollar Llama. She said she liked the writing. I told her I appreciated it, and then I told her the truth, which is that I use AI to help me write.
I didn't say it defensively. I said it because the compliment was real and I wanted to be honest about how the work actually gets made. I kept thinking about it after she left. I write this blog because most people don't have access to a family office, and publishing it here is one way to narrow that gap. AI is part of how I do that. It seemed worth explaining why I think that's fine, and why I think it's part of a much longer story.
The pattern
Every time humans have gotten a new way to share ideas, the same thing has played out. The new technology reshapes who can communicate. It displaces specific jobs. It amplifies how far a single voice can travel. And it takes a generation or two for people to figure out what to do with it. Writing with AI is the newest entry in that pattern. The pattern itself is worth looking at.
Want to skip the history? Jump ahead to how I use Claude.
Before writing
For most of human history, we had no writing at all. Knowledge moved through stories, songs, and images. This was not primitive. It was an information technology with its own professionals: bards, elders, storytellers. Some of it held up astonishingly well. Aboriginal Australian communities preserved stories describing coastal geography from around 7,000 years ago, before sea levels rose and the old shorelines disappeared. Homeric epics traveled for centuries by memory before anyone wrote them down. Cave paintings at Lascaux and Chauvet, 17,000 and 30,000 years old, appear to be more than decoration. They were records.
The limit was that knowledge could only travel as far as a human voice, and it only lasted as long as the person carrying it.
Writing
Written language showed up in Sumer around 3200 BCE, and for the first few centuries it was mostly used for accounting. It spread from there into law, contracts, astronomy, and history. For millennia, less than one percent of the population could read or write. The scribe became a profession.
Writing did not just preserve knowledge. It made new kinds of knowledge possible. Hammurabi's code meant laws no longer had to be remembered by a priest. Trade could happen at scale because obligations could be written down. But the same technology also enabled the earliest bureaucratic surveillance and the first detailed records of enslaved people. The tool was neutral. The people using it were not.
Copying by hand
Before the printing press, the only way to reproduce a book was to copy it by hand. A medieval Bible took a single scribe somewhere between fifteen months and several years to produce. It required 150 to 250 animal skins for parchment. A quality illuminated copy could cost about what a small farm cost. Monasteries ran scriptoria full of copyists. Later, universities in Paris and Bologna supported secular stationers who did the same work for tuition money.
Think about what that meant. Books existed, but a single book was a luxury good. Most people would live and die without owning one. The harm here was not in what the technology did. It was in who got to use it.
The printing press
Then Gutenberg arrived around 1450, and within decades the economics of writing completely changed. In the century after the press spread across Europe, book output went from a few million total manuscripts accumulated over all of human history to roughly 200 million printed books. Prices fell by about two-thirds. Luther's 95 Theses spread across Germany in weeks rather than generations. The scribe as a mass profession collapsed, though scribes held on in legal and luxury work for a long time.
The same press that gave us the Reformation and the scientific revolution also printed witch-hunt manuals, anti-Semitic pamphlets, and the pamphlet literature that helped drive roughly 150 years of religious warfare. Gutenberg did not change human nature. He gave us more of it, faster.
Radio, television, and the internet
For four centuries after the press, the printed page was the dominant way to reach a mass audience. Then, in about a hundred years, three more waves arrived.
Radio made it possible for a single voice to reach millions in real time. By 1940, roughly 83 percent of American households had a radio. FDR used it for the fireside chats. Goebbels used it for Nazi propaganda. In 1994, a radio station in Rwanda was used to help coordinate a genocide. Same technology, different hands.
Television added images. The 1960 Kennedy-Nixon debate is the textbook case. Radio listeners thought Nixon won. Television viewers thought Kennedy did. Vietnam became the first war Americans watched from their living rooms, and public opinion moved with the footage. Along the way, narrative control concentrated in a handful of networks, and civic engagement started a long decline that researchers are still studying.
The internet cut the cost of publishing to nearly zero. Anyone could reach anyone. It hollowed out newspaper classifieds, travel agencies, encyclopedia salespeople, record stores, and video rental, along with many other jobs. It also opened the door to misinformation at scale and algorithmic radicalization.
The pattern is consistent. Each leap amplified whatever humans brought to it, for good and for ill. Each one displaced specific jobs while letting more people participate in public conversation than before. And each one took a generation to find its footing.
How I use Claude
Now the new tool. I use an AI assistant named Claude to help me write Dollar Llama. The ideas are mine. The frameworks come from sixteen years of sitting in family office meetings and watching what works for wealthy families and what does not. The voice is mine. The judgment about what is true and useful is mine, and it has to be, because if I got that wrong the blog would not be worth reading. Claude does not decide what is true here. I do, and I check the facts and data in every post before it goes up. What Claude does is help me turn outlines into prose faster and more cleanly than I could alone.
The closest comparison is probably photography and painting. When cameras arrived in the 19th century, people worried that painting was finished. Instead, painters were freed from the job of being accurate recorders, and painting moved into territory cameras could not reach. Photography became its own craft with its own masters. AI writing sits somewhere in that neighborhood. The craft does not disappear. It shifts.
Which brings me back to my mother-in-law's compliment. It was real. What she was responding to, the way the ideas connect and the voice they are delivered in, are mine.
One more thing worth saying plainly. I am a practitioner, not a writer. My job is managing money for families. For most of history, someone in my position had two choices. Stay a practitioner and keep what you learn inside the firm, or step away from the work to write about it from the outside. AI collapses that tradeoff. It lets practitioners share what they know without having to stop practicing. That is new, and I think it matters.
Why this matters for Dollar Llama
Dollar Llama exists because most people don't have access to a family office. The whole point is to take the thinking that usually lives inside that world and put it within reach of anyone with an email address. AI is another lever on that same idea. It lets one person produce the quality and quantity of content that used to require a team. If I had to write every post from a blank page without any help, this blog would barely exist. With Claude, it does. The only version of this blog that actually reaches anyone is the one that gets written.
Every big shift in how humans share ideas has eventually made us more capable. None of them have been smooth for the people living through the transition. This one will be the same. In the meantime, I am going to keep writing, with help, honestly disclosed, in my own voice. If you are reading this, that is already working.