As much as I’ve enjoyed writing here over the past three years, that writing hasn’t been very focused. I’ve covered a broad array of topics. While there are a handful of recurring themes—in particular, politics, economics, and broad tech and crypto topics—I never made any specific attempt to focus on any particular areas. I just followed my heart and wrote about the things that I felt like writing about: the things I was most curious about or that were the most on my mind.
I’ve thought for a long time about narrowing the focus here a bit, and I think I may try this as an experiment this year. To be honest, it’s difficult because, in general, I’m bad at focus. I have many disparate interests and I don’t like to work on, or write about, the same thing day in and day out. But some degree of focus still makes sense and it’s still worth a shot. Success requires dedication and focus, and I feel like it would benefit my career and my professional life, too, to focus a bit more.
Here are three things I’d like to focus on next year: both here, in Three Things, and also in my life and career more generally.
Thing #1: AI Governance 👨⚖️
I’m as interested as the next person in the latest AI trends, but worry not because Three Things isn’t going to become yet another general AI blog. There’s enough of those already, there’s enough people covering general AI trends and the market overall, and I’m not the person to write such a thing.
It makes more sense to focus on an area where I have more experience and a deeper interest. For me, this is the intersection of AI and governance, and in particular, the role that modern AI tools have to play in governance. This intersection is extremely promising, it’s underexplored to date, and it’s a niche where I can potentially add value.
Governance has been an on-again-off-again area of focus and passion for me over the past few years. I was very active in Ethereum governance during 2018-2019, and subsequently in a number of DAOs and other open source, cryptoeconomic ecosystems. But I became quite frustrated and disenfranchised with the state of decentralized governance and with the “amateur hour” attempts at governance that I saw in a number of projects and ecosystems. I took a hiatus from governance work for a few years to focus on other, more technical, projects, including Spacemesh and Athena.
Recent trends in AI have rekindled my interest in the topic, and cast it in a new light. I’ve seen firsthand how ineffective and downright malicious human actors can be, and often are, in governance. I’ve seen the role that corruption, selfishness, ignorance, and small-mindedness play in scuppering even the most ambitious governance schemes. At the same time, I haven’t given up on fundamental concepts like democracy, self-sovereignty, and open, decentralized governance. I believe we should continue to experiment with these, and that we should put new technologies and new ideas at our disposal.
AI is the most obvious such new tool. It gives us an entirely new set of tools and a new set of paradigms with which to approach governance in a fresh light. Most people who are excited about AI today are using it for things like “agentic” Twitter bots, or baking it into applications that we’re already familiar with. But the most interesting use cases for AI, the non-skeuomorphic use cases, i.e., the AI native use cases, haven’t even emerged yet. I feel strongly that governance is one of these use cases.
The potential applications of AI in governance run the gamut from minimalistic interventions up to the Big Kahuna. At one end of the spectrum, the most obvious role that AI has to play in governance is as an advisor or research assistant of sorts. At the very least, in the most minimalist intervention imaginable, AI should make it easier for everyday people to stay informed about the issues and to contribute constructively to governance. This could go a long way to addressing some of the biggest issues faced in decentralized governance: ignorance of the issues, apathy, and a lack of participation.
In the middle of the spectrum, still very much within the bounds of what’s possible in the immediate term, would be something like an AI delegate or lawmaker with at least some formal powers, “sitting” among a panel of human delegates. It’s eminently possible to begin experimenting with both of these first two example use cases in the immediate term.
At the opposite end of the spectrum is a much more expansionist, maximalist view on the role that AI has to play in governance. The Big Kahuna, sci-fi inspired view would see us trying to build a powerful AI tool to which we can, over time, delegate more and more responsibility. The utopian view here is that the AI can make decisions in a fairer, more credibly neutral, more transparent fashion than a human actor, and can in theory be incorruptible and not motivated by the baser desires that motivate humans such as fame, wealth, and power. (I won’t bother describing the dystopian angle, since it’s fairly obvious and has also been done to death by sci-fi including The Matrix.)
I hope to not only explore the theory, and not only write about the topic here, but indeed to run some real world experiments with AI governance in 2025, especially with the “assistant” and “delegate” use cases described above.
Thing #2: Better Software 💾
One major theme here for the past few years, and indeed a theme of my career, has been a focus on making software more humanistic. I believe that software is far more broken than most people realize. As we rely on software more and more in our personal and professional lives, the ways in which it’s broken pose a bigger and bigger threat.
It’s not immediately obvious what to do about this situation. Many people smarter and more experienced than I am have written and spoken compellingly about the problems, such as lack of privacy, lack of control over our data and identity, and deplatforming, and I’ve written about them before as well so I won’t spend much time on that list here.
The key point is that, so far, workable solutions are few and far between. As broken as it is, software today does work after a fashion, but it mostly works for big companies, not for individual users. Nevertheless we have a degree of path dependency on the way software is designed and built today, and there’s no simple, easy path out. Client-server architecture, cloud computing, CDNs, and many other aspects of the way modern software is built and operated also contribute to these problems, and it’s going to be very hard to change them.
Nevertheless we’ve come a long way in the years since I began thinking about this problem. We have powerful tools like open source software, open data standards and protocols, and P2P networks. Now we also have blockchains: money, like bitcoin, but also networks that allow us to build unstoppable applications other than money using tools like smart contracts. We have other types of decentralized networks and frameworks for building decentralized, P2P applications such as Urbit and Holepunch, and decentralized social networks like Nostr and Farcaster. These tools are exceptionally powerful but, even after all these years, blockchains are still expensive and slow, and it’s too difficult to develop blockchain and P2P applications.
I spend a lot of time thinking about how to escape from this situation, which is a difficult coordination problem. Some of the next steps are obvious. We need blockchains that are faster, cheaper, and easier to use. We need easier onboarding and onramp/offramp experiences. We need much better interfaces—it’s time to escape the broken wallet paradigm.
But we need to go even further. Zooming out, we need to rethink the relationship among the user, the application, and the data. We need to get back to the user agent mindset, where software works on behalf of the user, not on behalf of the company that published and operates it.
Here, too, AI offers powerful new potential solutions. Eventually, we’ll have AI agents that actually act as agents, on behalf of the user. We’re not there yet, not even close, but it’s a promising avenue of exploration. We need AI agents that are open source, that can be trusted with our data and, to some extent, with our assets, and we need agents that are aligned with our interests. Each of these is a hard problem.
This is another pursuit that I’ll focus on, both here and in my career, in the next year. It’s one of the highest leverage things we can do, and one of the best uses of my time.
Thing #3: Wellness 🧘♂️
If your goal is to accomplish things, make a difference, and make the world a better place, the right place to start is with improving yourself. It took me a long time to understand that wellbeing needs to proceed impact. I spent years not taking care of myself—not sleeping well, not eating well, not being physically active, not focusing at all on mindfulness—and I didn’t realize the negative impact. In particular, I didn’t understand how much it reduced my productivity and effectiveness.
Wellness has grown in importance for me every year. It’s not “one and done”: it’s something that you must constantly strive for to make progress or even maintain your current level of wellness. While you eventually get used to it and even start to enjoy things like eating healthy, prioritizing sleep, and being active, it’s never easy. The older I get, the more I care about and learn about wellness, and the more I have to share about the topic.
I hope to share more about it here. Since I began prioritizing wellness a few years ago, wellness has also become a recurring theme here, including topics such as mindfulness, fitness, improving diet, and the importance of sleep. But there’s so much more to say and there’s so much more to do.
I’m not interested in writing about wellness just for the sake of wellness. It serves a greater purpose: namely, enabling the potential impact of the previous two things. None of these three things exists in isolation, and in fact all of them are connected. Improving wellness, especially things like sleeping better and being more mindful, allows me to focus more on work and to get more done while I’m working.
There are many aspects to wellness and all of them are important, but mindfulness deserves highlighting. Despite its importance and despite years of study and practice I still can’t say that I understand it well. I’ve shared many of the things I’ve learned here, but I’m constantly learning more about mindfulness through this study and practice, and I want to write more here about these ongoing learnings.
My hope is that, by exploring the topic more here, I can encourage more people to invest time and effort in understanding and experimenting with meditation and mindfulness as a way of unlocking happiness, peace, and productivity. There’s simply no other topic I could possibly speak to that has this same, large, universal potential. At the very least, even if readers here aren’t ready, willing, or able to commit to something as intense as a ten day silent meditation retreat, then they will at least be aware that such a thing exists, and of its life-changing potential. And even in small doses mindfulness has enormous benefits.
Expect more content about wellness this year.