Last Cast Letter #32: The Three D's

Hi All - Happy Friday and Happy Halloween. It’s the last day of the month, which means it’s time for the Last Cast Letter.

I, like many of you, have been thinking a lot about AI lately.

More specifically, I've been thinking about what an AI-driven world looks like in two years, ten years, and twenty years.

And even more specifically, I've been thinking about bets I can make now that will put me and my family in a position where, at the very least, we are secure during each of those time periods, if not thriving.

Because a Tweet this week caught my eye, amplifying a trend that I've already been following:

That's over 170,000 individual lives impacted by layoffs. These people don't live in a silo. Many have significant others and families that rely on this income.

Outside of straight earnings, that's 170,000 people who won't have a "purpose" until they find a new job, which is arguably even more detrimental to the fabric of their families and society at large.

Now, most of the blame for these job losses is being pinned on AI. Here's a header and subheader from a Wall Street Journal article:

“Tens of Thousands of White-Collar Jobs Are Disappearing as AI Starts to Bite. Layoffs at companies ranging from Amazon to Target are sending young and experienced workers alike into an unwelcoming market.”

With that said, some, including billionaire investor and podcast host Chamath Palihapitiya, think, "It’s convenient to blame AI, and this diagnosis may eventually turn out to be right, but the current wave of job losses are not because of AI. It is companies unwinding ZIRP and DEI hiring excesses that left them bloated and inefficient."

I don't fully disagree with Chamath. In fact, I Tweeted this last week, which got a little love:

But, based on what I see from running my own business, if I had to choose just one culprit, I think AI-driven layoffs and a lack of hiring due to AI feels more correct.

It's something I've been watching the whole year. In July, I wrote about "The Disappearing First Rung" and what feels like a brutal job market for entry-level hires.

I think the recent layoffs are part and parcel with the sentiment of this piece and the trends in the junior hiring market.

So, if this is now, and AI is only going to get "better", I revert to my original question: where are we going to be in two years, 10 years, and 20 years, especially as it relates to the job market?

Are we really going to have mass layoffs?

McKinsey estimates that "between 400 million and 800 million individuals could be displaced by automation and need to find new jobs by 2030". And that was published in 2017!

And if we do have mass layoffs, who benefits? Where does the value accrue? How can I position myself to be on THAT side of the fence, where the value is?

That's the question I'm trying to answer, and a few thoughts come to mind.

Before I outline where I think value, and by extension wealth, will accrue, I want to state for the record that I do not consider myself to be a "doomer".

By nature, I am a "glass half full" guy. But I'm also a realist and someone who tries to be proactive rather than reactive.

And so, with that said, here is an initial list of where I think value, wealth, and power will pool as AI accelerates. For now, I'm referring to them as the three-D's:

1. Data

2. Distribution

3. Deeds

I'm not going to spend a ton of time on 1 and 2 because this is a real estate-focused newsletter. With that said, both data and distribution could be viewed through the lens of real estate as well (i.e., datacenters). Let's touch on them briefly, starting with Data, but focus more time on Deeds for the purpose of this letter:

The Data Thesis

If intelligence is being commoditized, proprietary data becomes the real differentiator. AI models are only as valuable as the data they can access or fine-tune on.

This is why companies like OpenAI have entered into licensing agreements with major publishers (AP, Axel Springer, Dotdash Meredith, Reddit, News Corp, etc) and business agreements with companies ranging from Shutterstock to Salesforce. This is why Palantir has gone parabolic this year (up 160%+ YTD).

Individual model users, based on their queries, are providing data daily to LLMs like ChatGPT. Slowly but surely, we're all giving companies like OpenAI a window into our thoughts, fears, ailments, and more.

At the very least, this will be used for advertising purposes. Take a look at the Shopify and Etsy partnerships that OpenAI struck recently. To summarize, if you control or own unique, first-party data, I think you will be in a good spot.

The Distribution Thesis

Data is foundational, and it will be used to train AI models that let anyone write a song (Suno), produce a movie (Sora), or create a simple website (Lovable).

But if the cost of content, code, and design is going to zero, and anyone can launch a business during their lunch, how do you stand out in this flooded marketplace? How do you get your content or products in front of people who will pay for them?

This is where I think distribution will be crucial and why value will accrue to those who own the pipes, or the eyeballs. This is why celebrities launch products and services. They have built-in distribution and marketing through massive followings. Not all of these businesses ultimately succeed, but they certainly have a head start.

And to that end, I continue to think people are going to want to follow people.

I totally understand that AI already allows us to create avatars which look extreamly realistic. But people like takes and tone, and a lot of times these two things are delivered when they're off the cuff.

This is why I think there will be an "authenticity premium" for human-generated content, which will help build distribution. As information floods the system, trust, curation, and attention become premium.

The Deeds Thesis

Okay, now we get to the "real estate" section of this newsletter. If you've made it this far, thanks for hanging in there with me!

Here are my thoughts: in an increasingly digital world, I think physical scarcity reasserts itself.

He who owns the deed owns the demand because when intelligence, labor, and creativity become infinite, the only thing left with real value is what can’t be copied, coded, or conjured.

Value will accrue to those who own finite physical assets: land and real estate being one of them. Energy and infrastructure are others. I was going to throw rare earths in there, but realized that they can go in the land bucket.

If there's an obvious throughline, it's what I already mentioned above: those who own data need to acquire or work with deed holders to store/create it, i.e. data centers. This has been well covered; I don't think I need to elaborate here.

I also don't think I need to explain that not all real estate is the same. There are various asset classes, regions, rules, and regulations that will greatly impact which sub-sectors in this asset class outperform.

There are even nuances to these sub-sectors.

Take student housing, for example. That sub-sector as a whole is not a silver bullet when it comes to steady returns. In fact, I think some student-housing backed locations are "cooked" as Gen Z would say. I'm working on this paper for a later date.

Broad strokes, however, I do think that if you own physical assets, you will put yourself at an advantage compared to those who don't.

And I want to take this a step further, into the theoretical, because real estate as an asset class still doesn't feel as frenzied and "hot" as it did during the ZIRP / COVID days.

I still struggle to understand how some deals are getting done. Where I live, I chalk it up to stashing foreign (and/or illicit currency) and/or 1031 exchanges.

But what if you aren't locked into a fund with a 7-10-year time horizon? What if you can act like China and take the long view?

Are these deals actually not as bad as I think? Is half-way decent real estate undervalued? Not because of its yield today, but because of its future strategic necessity.

Rates are high, debt is painful, and the narrative says “wait until the Fed cuts.” But if AI really hollows out the wage economy, then maybe the market is mispricing what “hard assets” mean in that new equilibrium.

But then I play devil’s advocate with myself.

If millions of people lose their jobs, who’s paying the rent? That’s the paradox. The same dynamic that makes assets valuable could also shrink the pool of solvent tenants.

If AI kills the middle class, you can’t assume a growing stream of renters. The system would need a new payer. And the only one large enough is the government.

If a large share of the population can’t afford housing, expect a massive expansion of state-sponsored rent support: vouchers, UBI, tax credits, maybe even direct public ownership. I'm not sure this necessarily destroys landlords; it just changes who’s writing the check.

But governments are monopsonies (good word, right?). When they become your primary tenant, they eventually become your price-setter. Rent ceilings, renewal limits, “good-cause” eviction laws, all of it justified as “housing stability.”

So yes, maybe you still get paid, but it’s more predictable, less profitable, and more political. The yield survives, but the upside dies.

En Fin

The most seductive narratives are usually the ones that are the easiest to exclaim, but difficult to explain. They’re hyperbolic one way or the other. “AI will make everyone rich” or “AI will make everyone unemployed.” Both could be partly true, and both could miss the second-order effects that matter most.

Real estate sits at the intersection of all those effects:

• It’s physical in a digitalizing world.

• It’s politically exposed in a populist age.

• It’s needed but hated.

That mix makes it uniquely complex and possibly uniquely mispriced.

Maybe the real answer isn’t that real estate will win, but that it will survive.

And survival, in an era where work itself may no longer confer security (hence the layoffs I started this letter with) might be the most valuable thing of all.

I don’t have a clear call on where cap rates settle or how fast automation eats the job market. But I do know this: owning income-producing, real-world assets feels like the only genuine hedge against a future where labor’s value approaches zero.

If AI rewires the economy, you want to own something it can’t simulate. Something it can’t replicate or make more of. Maybe that’s farmland. Maybe it’s data centers. Maybe it’s just a small building with real people living in it.

It’s easy to think about AI as code. But code still needs somewhere to run. And as long as that’s true, there’s a chance the oldest asset class in history still has a future.