Original Content

You should think of each new piece of human-generated Personality Content as an investment in the ultimate compounding vehicle: your Personal Language Model.

And when is your Personal Language Model complete? When you are finished living. As long as you are alive, you are accumulating new experiences and perspectives and learnings. But to properly leverage them, it is your responsibility to build the skill of articulating them in the first place. Because, as I say often in our writing programs: “You can’t automate what you can’t articulate.”

Those who fail to commit to this process, and do not prioritize taking the time to write “how they became themselves” in the form of Personality Content will be at a severe disadvantage in the digital world. They will have no Digital Brain! They will have failed to transform their human-lived experiences into a Personal Language Model—one of the only data sets AI companies and Large Language Models do not have and cannot easily create. As a result, and as AI continues to get smarter and faster and more capable at producing higher-grade Commodity Content across a wider number of industries and topics, these Digital Brain-less humans will extract the least amount of value from these new technologies. They will not have built the skill (maybe one of the only remaining and necessary skills) of articulating and crystallizing “how they became themselves.” As a result, and as AI technology adoption accelerates, the everyday Digital Brain-less person will continue to create both “Human Slop” and “AI Slop,” causing them to sound like everyone else—a mime to the machine, the cognitive equivalent of a teenager lip-syncing to someone else’s song on TikTok.

Which is why, as we embark into this next era of Writing With AI (not “writing without AI” in technocrat rebellion, or “letting AI write everything for me” in a pursuit of some cyberpunk delusionaltopia), I am on a mission to educate the world on the right way to write with this new technology. This requires helping people understand what is entirely acceptable to outsource (Commodity Content —which we have been outsourcing for centuries with varying degrees of leverage), and what is unacceptable to outsource (Personality Content), not because of some moral conflict, but because Personality Content is an unmade data set.

When people slander the industry of ghostwriting, for example, this is often what they believe is happening: a writer has been hired to “hallucinate” someone else’s Personality Content. As one of the highest-paid ghostwriters in the world, I can tell you the vast majority of the time this is not what’s happening. What’s happening is a smart, accomplished person simply does not have the time, or the skill, to sit down and do “the writing” (the manual labor of organizing thoughts into groupings of words) themselves. It’s too low leverage. So, they hire a ghostwriter (someone like me) to do that work for them. I am not “hallucinating” their Personality Content. I am not making up details about their life that are not real, or choosing Commodity Ideas they have not chosen themselves. And I am certainly not creating truly Original Content on their behalf. I am talking to them (usually on a recorded phone call or Zoom meeting), asking questions, and getting their Personality in their words, spoken aloud. I then take the words they’ve said, cemented into a transcript, and begin whittling it down the same way a sculptor would a block of marble. I remove tangents. I remove unnecessary words. I organize. And I add or subtract any necessary pieces of Commodity Content—which were not the client’s to begin with, and also not “mine.” These are industry-specific words and groupings of words, open-source and owned by no one.

Whether “I” am the one doing this work, or a robot, is somewhat irrelevant. Human beings provide labor leverage, and Large Language Models provide technology leverage. It’s not the form of leverage that matters. It’s the quality of the input. And in order to create a high-quality Digital Brain, it is still the person’s responsibility (the client, in this example) to articulate “how they became themselves,” first. Whether that articulation happens by way of a monologue recorded alone, by way of conversation, through writing yourself, or writing with the help of an editor, doesn’t matter. It also doesn’t matter if it’s a human or a robot who helps you organize the content thereafter. Quite literally the only thing that matters is that YOU provided the input. If you do, you are likely to stand by the quality of the output. (And if you don’t, then don’t be surprised when it doesn’t “sound like you!”)

This is the right way to write with AI. And the beginning of what I believe will be a multi-generational transition in collectively identifying what is worth doing in the first place. AI is not going to steal all the writing jobs (technology leverage). And the ghostwriting industry isn’t going anywhere (labor leverage), because even the most powerful technologies have yet to find a solution to human inaction. Leverage, in and of itself, does not remove the bottleneck of you getting in your own way. You must take the first step, and you must provide the input—and then, and only then, can labor leverage or technology leverage help scale “you.”

If language is an open-source operating system for communication, then most combinations of words and groupings of words are Commodity Content.

If Commodity Content was never yours to begin with, then spending your finite time here on earth manually writing and re-writing the same Commodity Ideas in slightly different variations out of pride for “doing it yourself” is not a symbol of purity, but a Scarlet Letter of stupidity. You are making the conscious choice to write the wrong things.

Personality Content is the unmade data set. It is also one of the only forms of differentiation, association, and “ownership” in the age of infinite text-based Commodity Content. Time spent writing and re-writing Commodity Content, past a baseline level of proficiency, is low leverage and unnecessary. Time spent writing and re-writing Personality Content is higher leverage and additive to the long-term value of your Digital Brain.

But when all conventional wisdom has been said, and “how you became yourself” has been sufficiently articulated in your library of content (your Personal Language Model), what is left? What do you say when there are diminishing returns in repeating yourself, and a machine can repeat the same stories, insights, opinions, and lessons from your Digital Brain, infinitely, without needing your continued involvement?

All that’s left is the creation of Original Content: thinking thoughts that haven’t been thought before, and making connections that haven’t been made before, to say things that haven’t been said before. And I would posit nothing would have a bigger impact on the collective intelligence of our society than having exponentially more people spending exponentially more time “thinking different thoughts,” opposed to endlessly repeating and manually writing the same Commodity Ideas over and over again in the hope of establishing those words as “theirs.”

You do not create something Original simply because your brain thought the thoughts, or your fingers clacked the keys. You also do not create something Original simply because your creation consumed time or effort. Original means additive. It’s new. It’s net-positive. It’s moving the world forward in some big or small way, identifiable to you.

• Say something different.
• Say the same thing in a meaningfully different way.

An example of the former would be Pythagoras saying, in 500 BC, “The earth is round. Not flat.” These are six Original words. Six words that fundamentally changed the way human beings perceived the world. Before Pythagoras, no one had “thought that thought.” As a result, we now (centuries later) still associate those six words with Pythagoras. This is the epitome of “ownership.”

An example of the latter would be Aristotle, a century and a half later in 350 BC, essentially saying “the same thing” but in a meaningfully different way. He proved Pythagoras’s round-earth theory through empirical and observational evidence—pointing to lunar eclipses, horizon lines, and changing constellations.

This is the art of writing.

It’s not to manually type words. It’s also not to repeat Commodity Ideas (“conventional wisdom”) and call them your own. It’s to think deeply about a small number of topics and use writing as a mechanism for understanding. But as you understand each Commodity Idea, you should not remain attached to carrying the burden of repeating those chosen Commodities manually yourself. That’s what the robots are for. And you should feel empowered knowing you are continuing to focus on the next-highest-leverage task at hand: making new connections, forming new hypotheses, and thinking different thoughts, in pursuit of pushing your area of interest further than you found it, creating something Original.

Large Language Models are miraculous machines. But they are not going to replace the need for human beings to “think different thoughts.” And if writing is the most effective mechanism in human history for crystallizing one’s thinking, then AI is not going to replace writing as a skill anytime soon. Besides, in order to use AI, you have to know how to prompt. And in order to prompt, you have to know how to… what? You have to know how to write.

You can’t automate what you can’t articulate.

My hope is these Large Language Models will reduce the amount of time we spend as a society writing and re-writing Commodity Content, and give our Personality Content leverage, so we can spend more time thinking and writing what hasn’t been said yet.