What We’re Doing (and Not Doing) with AI

TL;DR:
AI is transforming everything, but not in the way most people think. The real threat isn’t conscious machines—it’s the flood of low-quality, low-effort content. At MarketLife, we use AI as a tool, not a creator. I still write all long-form content myself. We use AI for things like editing, short-form posts, and code assistance—but never to replace human insight. This post lays out our current policy and values, and why integrity still matters in an AI-saturated world. (And this summary was AI-generated!)

I thought it might be a good idea to outline when and how we are using AI in our work.

None of us really know how to think about AI. It’s clear that the world is going to change. In some ways, this will be an even bigger shift than the Internet. Many jobs will be eliminated, and the landscape is going to change significantly. In a very real way, the world is going to fracture along the lines of those who use AI well, and everyone else.

As a composer, I’ve been concerned about algorithmic music generation since my first encounter with it in the 1990s. And it has, of course, gotten exponentially more powerful and nuanced over time. There’s always the argument that computer-generated music is soulless or otherwise deficient, but an objective evaluation of the output should certainly raise some troubling questions for creators.

We’re swimming in these questions:

Will AGI evolve? Is consciousness an emergent property of complexity? Will AI wipe humanity out? Will AI replace jobs? How will people support themselves, if so? Will we see the end of human produced art and writing? (Ummm… in no particular order…)

Some of these answers are clear. Some are completely unanswerable at this time. But I will say this: AI has already leveled the field in some important ways. The AI you interact with today is the worst AI you will ever touch—it’s only going to get better.

But the barbarians are probably not at the gate. (Well, the barbarians are always pounding at some gate. The question is which gate. I don’t think they are going to come in through this one!)

This has precedent. A pretty smart guy once railed against the invention of a little technology called writing. He worried that it would make men stupid—that someone would think they understood something just because they read it, that the written word was dead and could not respond to inquiry, that no one would remember anything because it was all written down. (Oh, and that people would forget how to listen to what the rocks and trees said.)

Plato’s concerns (in Phaedrus, around 370 BCE) were just a wee bit overstated, but I think we know where he was coming from.

Content is getting cheaper. And far more plentiful. And shittier. This is not a good trend.

In light of this, I thought it was worth sharing how we are using AI. This is a rapidly evolving landscape and what I write here may not be true tomorrow. But the overall commitment will carry through.

• We will never produce solely AI-generated content. It’s unthinkable to me that I will ever generate audio or video AI content.
• I write my own long-form content. I spend real time shaping it, word by word. If it’s here, on my substack, or in a published format, I wrote it.
• When I do use AI for writing, it is as an assistant: for tone check, to find outright errors as a copyeditor, and to simulate the experience of specific kinds of readers. It is currently less than fully reliable in any of these areas, but it is helpful.
• We may use AI to shape some short-form messages, like tweets or summaries.
• We may use AI to craft some emails. Why? Because it is a better copywriter for that type of content than I am.  I view this use as essentially hiring a copywriter. My writing is probably best in long-form, and I can sometimes tend to write overly complex sentences that require multiple parentheticals (which obviously has a time and place (and can even be ideal, in some cases (but which also obviously works against the clarity required for concise email messaging))). If you’re wondering, yes, I talk like this too, and I sympathize with the people who have to listen to me frequently!
• We use AI to assist with some graphics. In other cases, we have and will continue to hire designers and to support human artists.
• We are using AI to assist with coding behind the scenes and to check formulas for math errors. This is also a mixed bag, and if anyone is vibe coding trading systems I would love to take the other side of your trades!

This is subject to change, of course. No doubt, it will change and must change. If it changes significantly, I’ll update you.

But I wanted to put this stake in the ground. For transparency and also in the interest of integrity.