Notes on AI:
For the Young Ones.

I was in a board meeting in 2016 when my client, the then EVP of research & analytics, someone I revere said, Netflix was "just a distribution play." The studios, he argued, still held the real power — they owned the content. Netflix needed them. They didn't need Netflix.

Three years later, that same executive watched his company sign a survival deal with a streamer. His library, built over decades, was now someone else's algorithm.

I think about that room a lot when people tell me AI is like the printing press.

The printing press comparison is seductive because it's comforting. It says: we've been here before. Technology democratizes, things get messy, then we adapt, and progress continues. The arc bends forward. Go back to sleep.

But I've lived through enough disruption cycles to know that the most dangerous moment isn't when the threat is obvious. It's when it looks like a partnership.

That's exactly how Netflix looked in 2012. And that's exactly how AI looks right now.

Here's what I've watched happen in my own life over the past two years.

I used AI to draft a strategic memo that used to take my team three days. It took forty minutes. Then I asked it to analyze a competitor's pricing model. Then to synthesize six months of customer interviews into a framework I could present to investors. Each time, I told myself: I'm still driving. This is just a faster car.

But somewhere in there, something shifted. I stopped building the muscle. The friction that used to force me to think — really sit with a problem — started disappearing. I was producing more. But I'm not sure I was thinking more deeply.

That's not a productivity story. That's a dependency story.

And I'm a person who noticed it happening.

The printing press put knowledge into more hands. More people could read. More people could question. Power decentralized.

AI is doing something structurally different. It's not distributing intelligence — it's mediating it. Every time you ask a model how to think through a negotiation, structure an argument, or interpret a market signal, you're not just using a tool. You're routing your cognition through infrastructure owned by three or four companies.

The internet democratized publishing. But it also created Google, Facebook, and Amazon — platforms that now control what gets seen, bought, and believed. We got Wikipedia and algorithmic radicalization in the same package.

AI will do the same. It will give individuals extraordinary leverage and consolidate the underlying layer of intelligence into a handful of firms. We're already watching it happen. OpenAI, Anthropic, Google DeepMind, and Meta are spending hundreds of billions to build the cognitive rails everyone else will ride.

That's not the printing press. That's Standard Oil with a chatbot.

I came up in an era when "learn to code" was the answer to everything. Disrupted by globalization? Learn to code. Worried about your job? Learn to code. The promise was that knowledge work — analytical, technical, cognitive work — was the safe harbor.

We built entire middle classes on that promise.

Now I watch companies quietly running experiments: what happens when one senior person plus AI does the work of a twelve-person team? In some cases, they're finding out the answer is: pretty well, actually.

This isn't science fiction. I've seen it in my own portfolio companies. A two-person growth team outperforming what used to require eight people. A solo founder shipping product at a pace that would have required a full engineering squad eighteen months ago. The leverage is real, and it's accelerating.

The question no one wants to say out loud in a board meeting is this: if the middle layer of cognitive work gets compressed, what happens to the people who were standing in it?

I'm not a pessimist about this. I've made my career betting on technological change, not against it.

But I've also learned that optimism without structure is just denial with better branding.

The industries that survived platform disruption weren't the ones that ignored it or the ones that panicked. They were the ones that got structural early — built direct relationships with customers, owned their data, created switching costs, diversified their distribution. They played the long game on agency.

That's what I think the right response to AI looks like too. Not resistance. Not cheerleading. Structure.

Open-source models matter because concentration requires alternatives. Regulation matters because infrastructure without accountability becomes toll roads. Education reform matters because we're still training people for a world that's quietly being automated underneath them. And individual awareness matters — because the first step to not outsourcing your judgment is noticing that you're doing it.

Here's what I actually believe, after sitting with this longer than I'm comfortable admitting:

The most valuable thing in an AI-saturated world won't be the ability to use AI. Everyone will use AI. It'll be commoditized within a decade, like electricity or search.

The scarce thing will be judgment. The ability to ask the right question before the model answers it. To know when the output is wrong. To hold a framework in your head that the model can't synthesize because it requires lived context, domain experience, and intellectual honesty about uncertainty.

History, philosophy, economics, systems thinking, domain depth — the stuff that got deprioritized when coding became king — may be exactly what gets repriced upward.

Technology becomes abundant. Wisdom stays scarce.

The printing press gave us the Reformation and the pamphlet wars and eventually the Enlightenment. It also gave us centuries of religious conflict and centralized propaganda before those better outcomes arrived.

The story wasn't clean. It never is.

AI won't be clean either. It will give us extraordinary things — medicines, scientific breakthroughs, access to expertise that people couldn't afford before. I genuinely believe that.

But it will also concentrate power in ways that look invisible until they don't. It will create dependencies we don't notice until we've already lost the muscle. And it will raise a question that capitalism has never really had to answer before:

What do you build an economy around when intelligence itself stops being scarce?

That's not a technology question.

That's the question of the next fifty years.

And I don't think the printing press gives us much help answering it.