What Happens to Content When We Can All Create It?
Why Less Is More
For the past several weeks, I've seen plenty of articles and tweets raving about how easy it is to create long-form content with AI. One well-known investor on Twitter mentioned how he was able to write a lengthy book (almost 100,000 words) in literally one day with the help of Claude.
It may be an egregious example. But it goes to show how people are using AI to drastically increase their written output. I've even heard stories about authors trying to create literally one book (or multiple books) per day with AI, upload the books to Amazon, see which books stick, allocate more marketing spend toward those books, and then create more books like those successful books. It's startup-style iteration and speed with the written word.
I'm probably like you in that I don't care to consume this type of content. The more I consume recent digital content, the easier it is to tell that it was written by AI. While it's getting better, there is certain syntax and certain grammatical choices that raise red flags. It immediately diminishes the author's credibility.
And I'm not alone. I read there is this new term going around called "AI;DR." Like TL;DR, the idea is that any inkling of AI-generated content will lead readers to move on to another piece of content.
So why are these individuals continuing to create books made of AI slop and upload them to Amazon?
Most likely, it's just a cash grab. Our economy and society are in the middle of a massive transition. There is a unique window of opportunity to capitalize on increasingly powerful models and some individuals are trying to do that.
At the same time, I still think the long-term winners will be original content creators who lean into the human element of their work. Relying on AI-written prose, even as the models get better, is essentially a Pyrrhic victory.
Yes, you can "get more content out there" and potentially build a larger audience. However, that audience is likely brittle. The number may be impressive, but the underlying engagement likely isn't.
It's yet another example of one of my main theses from these posts. All of us need to think about our edges as humans. Then, with some AI-enabled help, we need to ruthlessly exploit those edges. If you're a strong writer, for instance, lean as much as you can into your unique voice and experiences. I'm not against using AI in some part of the process, but it should never overshadow the edge that you have. The same principle applies to any other skill or expertise that you have.
The world is changing super quickly. I only started writing these posts several months ago and we've seen truly substantial advancements in AI technology. It'll probably get even more impressive.
The bottom line, though, is to see this as a tool to supplement your real strengths. It isn't to replace those skills entirely.
Prompt of the Week
I'm not a macro investor. I prefer to look at individual companies instead of speculating where the global economy is going in the next 12-24 months. That said, I think that it would be a fool's errand to entirely avoid macro trends. Because of this, I thought to ask ChatGPT about historical parallels and how (if possible) that could be translated into actionable investment ideas. Here is the prompt I used:
"Pretend that you are an expert financial historian who translates historical events into actionable investing insights. What are some extremely understudied, yet highly applicable analogs to this current period in history? And considering those periods, what are some sectors or companies that could represent attractive long or short ideas? I'm not looking for obvious ideas, as they are likely priced in already. The more esoteric analog, the better."
Until next week,
Adam