Hello {{ first_name }} 👋
I hope you had a good start to the year! Mine was quiet, mostly time with family and friends. The one thing already climbing on my to-do list is spending more time with Claude Code. The recent progress around vibe coding is very real, and it’s where AI has truly gotten superhuman.
I wanted to use this short note to do three things: look back on what we built with Chain of Thought in 2025, put the work in context, and share how I’m thinking about 2026.
Lovin’ our cover art - I frame them up on my study room wall
I started writing this newsletter 2 years ago. It was really about noticing that AI was going exponential. Over the years, I’ve learned that my edge comes from noticing patterns early and staying close to sharp thinkers.
While things were still quite vague back then, the future picture of the world has become much clearer to me now. We are watching three systems collapse into a single stack.
Software that can reason
Machines that can act
Blockchains that can settle
The world is entering an economy where intelligence becomes a direct input to production.
→ Chain of Thought is the Machine Economy newsletter
I’ve been studying this transition to understand where value accumulates and where failure hides. I’ve been in tech for >10 years, and the convergence of AI, Robotics, and Crypto is the most exciting thing I’ve ever seen.
It’s actually happening at scale, and, in my humble opinion, will be the most important technological paradigm shift for humanity into an advanced civilization.
Our research is now published through two newsletters, each offering a different lens on the same Machine Economy future.
On-Chain: This is the one you’re most familiar with and covers crypto/blockchains as incentive and coordination infrastructure for decentralized systems.
Autonomy: This newer track focuses on AI agents (non-crypto). I started it six months ago, and it has been growing quickly as agentic systems move from research into production. If you’re not on the list yet, you can join here.
2025 in a nutshell
2025 was an incredible year - I was never bored.
CoT published 91 posts with a healthy 40% open rate. This includes shipping the Crypto x AI newsletter (AI Edge, now at edition #80) every single week without missing an issue. Surfacing the developments that actually mattered, even when they were unfashionable or early.
Alongside the weekly newsletter, we wrote deep dives on a small number of promising companies, analyzing their products and strategy. Most recently: OpenMind (robotics), Codatta (data), INFINIT, and Surf (AI agents).
Some of this work was commissioned, which allowed me to spend quality time with the founders to understand how they see the future and what they’re trying to build. Because these are private companies, much of this thinking is not yet public.
I’m always respectful of your time, so the bar for selection stayed high. I filter hard for teams doing work that is both innovative and timely. Many pitches were turned down because they simply did not merit your attention.
I also developed and shared a set of mental models for thinking about decentralized AI (the Decentralized AI Canon series), covering Verification, Compute, Training, Data and more. That was especially fun to write.
And if you missed it last week, as an ode to the year, I distilled my favourite ideas from 1,436+ podcasts and presented them in a mental map for thinking about how 2026 might unfold in AI.
I have to say: None of this works without you. The replies, the DMs, the poll responses, and even the quiet readers who open the email every week. This gives me energy and keeps me writing. Thank you for sticking with me. I’ll keep giving you my best work.
So…what’s ahead in 2026? Lots! 🫡
I’ll continue writing deep dives on the most compelling startups in AI, robotics, and crypto. The filter is not getting looser. I already have a short lineup of teams you may not have heard of yet, doing serious work that deserves attention in the coming weeks. Our research work is already underway.
I’m also making it a priority to revisit and update some of our prior frameworks (e.g., data networks). A lot has shifted even over the last few months, and some assumptions need to be revisited with fresh evidence.
Beyond the writing, you can expect a slew of new ways I’ll be launching to help us stay close to the frontier. These include podcasts (you’ll see more of me talking on YouTube), data dashboards, new content formats, and a community channel for us to grow together. I’ll announce them once they are ready.
My goal stays the same: help you think clearly about product, strategy, and capital allocation.
I try to keep this audience tight and thoughtful. If this newsletter has been helpful, I’d love it if you could share it with a friend or colleague who might find it valuable too.
Here’s to a focused and productive 2026!
Teng Yan
