I've been thinking a lot lately about how much my investment philosophy has changed over the years.
I started investing pretty young. One of the first stocks I ever bought was Hawaiian Electric. I was still in high school and needed a custodial brokerage account that my mom had to sign off on.
At the time, what attracted me to investing was dividends. There was something that made sense to me about owning stable businesses, collecting the cash flow, reinvesting it and watching that income grow over time.
It felt predictable. If you owned companies that were deeply embedded into the economy and American lifestyle... utilities, telecommunications companies, consumer staples, things people used every day... you could make some reasonable assumptions about the future. Obviously nothing was guaranteed, but it felt like you could plan ahead.
How much income could this portfolio produce in 10 years? What about 20? If the dividend grew 5% per year, and I continued reinvesting it, what would that eventually become?
I liked that. I still do.
But over the last several years I've started looking at investing through a different lens. And I think 2020 was probably the turning point.
When COVID happened, it seemed like the world flipped upside down almost overnight. Businesses shut down. Supply chains stopped working. People stopped going to work.
The government stepped in. The Federal Reserve stepped in. M2, a broad measure of the money supply, expanded rapidly. There was suddenly an enormous amount of money moving through an economy where the actual productive capacity had not increased at the same rate. Initially, we got a huge bump in spending. Then inflation followed.
I'm not trying to make the argument that increasing M2 was the sole cause of everything that happened afterward... economics is obviously more complicated than that. But it made me start thinking differently.
What exactly is money? What am I saving? What happens to purchasing power when the supply of the thing I'm storing my work in can expand?
At the same time, I started buying Bitcoin. I wasn't all-in on Bitcoin. I was still a dividend investor. For a while, I was doing both.
But the more I watched what was happening, the more Bitcoin started making sense to me for reasons that went beyond simply hoping that the price would go up.
Hawaiian Electric was one of those investments that always stuck with me. Partly because it was one of my first stocks. Partly because I'm from Hawaii. And partly because it represented exactly what originally attracted me to dividend investing.
It's a utility. People need electricity. It had been around forever. You would think that if there was ever an example of a relatively predictable business, that would probably be one.
Then the Maui fires happened. HE suspended its dividend.
I'm not saying that was the wrong decision. Given everything happening around the company, preserving cash was understandable. But it changed something in the way I looked at investments. Something that I had thought of as permanent... wasn't.
I had owned the stock since high school. I had literally watched this company go from being one of the examples that attracted me to dividend investing, to showing me why even the things that appear to be the most stable can change.
I've seen similar things happen elsewhere. AT&T was another classic dividend stock. Then acquisitions, debt, restructuring and changes in the business eventually led to a significant reset of the dividend.
Again... this doesn't mean dividend investing doesn't work. There are companies that have increased their dividends for decades and may continue doing it for decades more. What I've started questioning is something a little different.
How good am I really at predicting which companies will still have durable economic moats 20 or 30 years from now?
And more recently, AI has made that question much harder for me to answer.
I spend a lot of time thinking about AI. Probably too much time.
And the deeper I get into actually implementing AI into a business, the more I realize how difficult it is to predict exactly what the economic impact will be.
Software is getting cheaper to build. Knowledge is getting cheaper to access. Tasks that historically required people with specific training can increasingly be performed, assisted or checked by software. Entire workflows can be rebuilt. Companies themselves may eventually be structured differently.
What happens to the labor market? What happens to a company's cost structure? What happens to businesses whose moat is largely access to specialized knowledge? What happens when a competitor can recreate portions of your software for a fraction of what it used to cost?
I don't know. And that's really the point.
AI can pick apart many economic moats. It can also create new ones. There are probably companies operating today that will use AI incredibly well and become substantially more profitable because of it. Others may get disrupted. Some businesses may suddenly find that what they thought was a competitive advantage really wasn't much of an advantage at all.
The hard part is figuring out which is which.
The obvious answer would be... If AI is going to change everything, invest in AI.
Semiconductors, data centers, power, networking, cloud infrastructure. The picks and shovels. And everyone and their mom seems to be looking for the picks-and-shovels play right now.
I'm not saying those are bad investments. Some of them may end up being incredible investments. But I'm having a hard time convincing myself that I understand where the long-term economics of AI actually settle.
We have frontier models that are becoming unbelievably capable. At the same time, it isn't completely clear to me yet how the frontier model companies ultimately turn those capabilities into durable profits relative to the enormous amounts of capital required to train and operate them.
And then there's another possibility that I think is interesting. Maybe the future isn't simply one massive model doing everything. Maybe increasingly specialized models and tools start making their way through the economy.
We are already experimenting with some of this in our own business. A smaller model or a JEV-type tool that performs one very specific job may be able to do that job more cheaply, consistently and reliably than asking the most intelligent model available to solve everything.
Maybe that's where a lot of the useful economics eventually develop. Specific models, specific workflows, specific agents, specific jobs.
Or maybe I'm completely wrong. We're still early. That's what makes it difficult as an investor.
I am extremely bullish on what AI will do to the world. I'm much less certain that I know exactly where all of the investment returns will accrue. Those aren't necessarily the same thing.
This is the question I've been wrestling with.
If technology is moving this quickly, economic moats are getting harder to identify, we don't really know what a post-AI labor economy looks like, and money itself can be expanded... where do you store capital?
Increasingly, my answer has been Bitcoin. Not because I think I've solved investing. And definitely not because I think Bitcoin is risk free.
Bitcoin can drop 50%. Regulation can change. Adoption can slow. There are custody risks, political risks, technological risks and probably risks that I haven't even thought about yet.
But there is one part of the thesis that I find incredibly attractive: scarcity.
There will only ever be 21 million Bitcoin. There isn't a CEO who can decide to issue another 20 million.
There isn't a board that can decide the payout ratio needs to change. There isn't an acquisition that suddenly puts the balance sheet under pressure. There isn't a natural disaster that causes Bitcoin to suspend its dividend.
The rules are different. And I think that's what I've slowly become more interested in.
Not certainty around the price... certainty around the scarcity.
Since I started buying Bitcoin in 2020, I've been watching this progression happen. Initially, it still felt relatively niche. Retail investors, Bitcoin people on Twitter, people talking about self custody and monetary policy that most normal investors probably didn't care much about.
Then Wall Street started coming in. Then the ETFs. Then more regulatory infrastructure. Then governments and sovereigns started having conversations about Bitcoin.
And now we're watching another interesting financial experiment develop with Bitcoin treasury companies. Companies are raising capital through common stock, convertibles and increasingly preferred securities, then using that capital to accumulate more Bitcoin. Some of these structures are essentially attempting to use traditional capital markets to accretively increase Bitcoin exposure per share.
Will all of these strategies work? Probably not. There will almost certainly be good versions and bad versions. But the progression itself is fascinating to watch.
Bitcoin has gone from something that was largely dismissed as an internet currency to something retail investors owned, Wall Street packaged, institutions allocate to and governments are thinking about. Now companies are building increasingly sophisticated capital structures around it.
That doesn't guarantee that Bitcoin succeeds. But it does continue to make me pay attention.
The more I've thought about AI and Bitcoin together, the more interesting this idea has become to me.
AI is creating abundance. More intelligence, more software, more content, more code, more analysis. Eventually, probably substantially more output from the same amount of human labor.
The marginal cost of producing many things may continue moving toward zero. And if that's the direction we're heading... what becomes more valuable? Maybe scarcity.
There are plenty of scarce things. Land, time, attention, trust, certain commodities, relationships. There are things that technology cannot simply generate infinitely more of.
Bitcoin is interesting because we have managed to create something digitally native that behaves this way. In a digital world where almost everything can be copied... Bitcoin can't simply be copied into 42 million Bitcoin.
You could create another cryptocurrency. You could copy the code. But you can't recreate Bitcoin's exact network, history, ownership distribution, infrastructure and accumulated adoption from scratch. At least that's how I'm thinking about it today.
I don't think I've solved investing. I don't think I've got this all figured out. In fact, part of the reason I'm writing this is to help myself wrap my head around these ideas.
That's kind of what this blog has always been for me anyway. Putting thoughts somewhere permanent so I can look back later and see how my thinking has changed. Maybe ten years from now I'll read this and realize I completely missed something. That's okay.
I'm still looking for companies. I'm still looking for opportunities. I'm still trying to understand where AI creates durable economic value. I'm still trying to figure out what the next economic moats look like. I'm still searching for alpha.
But right now, I'm finding it harder to forecast the permanence of many businesses while finding the scarcity of Bitcoin easier to understand. That's an important distinction. It doesn't mean Bitcoin is predictable. The price certainly isn't. It means one of the underlying properties I'm buying is predictable: the supply.
Maybe the next great investment opportunity becomes obvious as AI matures. Maybe entirely new categories of companies emerge. Maybe existing companies use AI to build even stronger moats than they had before. I'm sure my thinking will continue to change.
For now though... in a world that seems to be getting better and better at creating abundance, I've found myself increasingly searching for the things that remain scarce.
And today, that search keeps bringing me back to Bitcoin.