Ray Dalio published a piece today called "Principled Thinking and AI Need to Go Together." When I read it, I did not learn anything new. I recognized it.
I worked at Bridgewater Associates. The culture there does something to how you think. It forces you to name the criteria behind every decision, to separate instinct from logic, to demand that reasoning be visible. That discipline stayed with me long after I left. It shaped how I evaluated opportunities, how I advised companies, how I thought about what was actually driving outcomes versus what people assumed was driving them. Over the past year, building a relationship intelligence platform, I have started to see just how deeply that way of thinking runs through everything I design.
What Principled Thinking Actually Means
Dalio defines principled thinking as "the examination and systemization of one's decision-making criteria rather than just thinking to make decisions." Most people will read that as organized thought. It is sharper than that.
That practice rewires how you operate. Once you start naming your decision-making criteria explicitly, you notice how often you were making decisions on instinct dressed up as logic. The instinct might be right. But until you articulate the criteria underneath it, you cannot test it, refine it, or teach it to anyone else. You cannot improve what you cannot see.
That is the distinction most people miss. The gap is not between good judgment and bad judgment. It is between judgment you can name and judgment you cannot.
Understanding, Not Data Mining
Dalio makes a point in his piece that most AI builders are ignoring. He says principled criteria "are not best derived by looking at what would have worked in the past and assuming that it will work in the future, i.e., data mining, or simply asking an AI what to do. They are based on logical understandings converted into decision-making systems."
The default approach in AI is pattern recognition. Feed in enough data. Find correlations. Optimize for what worked before. The problem is that correlation-driven systems break silently. They cannot tell you why they made a decision. They cannot be debated. They cannot adapt when the environment shifts because they never understood the environment in the first place.
Dalio describes the alternative: each criterion has a reason attached. "If this happens, do this, because XYZ." The because matters more than the rule. When a principle stops working, you trace back to the causal logic and find where the world changed. You do not throw out the system. You refine your understanding. At Bridgewater, that discipline was cultural. You lived inside it every day.
That habit shaped how I think about building intelligence systems. When I sat down to design Revolv, I kept coming back to the same question: are the criteria underneath this system based on a logical understanding of how relationships actually work, or are we just mining patterns and hoping they hold? The Bridgewater instinct was always to demand the former.
Dalio tests his principles across time and geography. As far back as possible. Every country. Every type of environment. If a principle only works in certain conditions, it is not a principle. It is a local optimization. I apply the same test to the principles underneath Revolv. Trust builds through consistent follow-through. Context compounds over time. Relationships weaken without reinforcement. People remember being remembered. These hold across industries, roles, and eras. They were true before phones. They will be true after whatever comes next.
The Partner, Not the Oracle
The most counterintuitive thing Dalio says is that even the most advanced AI is not good enough to follow blindly. He describes the ideal as a partnership: the system makes moves based on systematic criteria while you make moves based on the criteria in your head, and you compare the two. Most AI products are built on the opposite assumption. Hand off thinking. Accept the answer. Move on. Dalio's framework rejects that entirely.
Another line stayed with me even more: "The system always speaks to you, explaining its logic so you can understand each other and align your thinking." Most AI products are silent. They process in the background and surface a result. If you ask why, you get a confidence score or nothing at all. At Bridgewater, the culture demanded that reasoning be visible. You were expected to challenge logic, not defer to it. That friction was the point. The alignment happened through the disagreement.
You are not outsourcing your judgment. You are augmenting your memory.
I internalized that years before I built anything. Revolv is built around both principles: the human stays in the loop as a partner, never as a passenger, and the system explains its reasoning so you can challenge it, calibrate against it, and improve alongside it. A relationship intelligence system that cannot explain itself cannot earn trust. And without trust, you either follow blindly or ignore entirely. Neither produces intelligence.
The Domain That Was Missing
Dalio writes from investing. He has spent fifty years encoding cause-effect relationships about markets into systematic criteria that partner with human judgment. The framework he describes is proven in that domain.
But there is a domain where the same framework applies and nobody has built it yet: relationships.
People already have good instincts about their relationships. They know who matters. They know when something feels off. They know when a connection is going cold. What they lack is the systematic memory to act on those instincts at scale. Their judgment is sound. Their bandwidth is not.
The insight Dalio surfaced decades ago, that principled thinking bridges human intelligence and computerized intelligence, turns out to be exactly what professional relationships need. Encode cause-effect relationships about people, trust, timing, and context into systematic criteria. Build a system that partners with memory the way Dalio's systems partner with judgment. Make the logic visible so the human can debate with it and improve alongside it.
That is what I am building with Revolv. Not a CRM. Not a networking tool. A relationship intelligence platform that applies principled thinking to the domain of human connection, preserving the context that makes professional relationships compound over time rather than decay.
The Line That Connects
The same instinct kept showing up across my career: the hardest problems are the ones where human judgment is irreplaceable but human memory is insufficient. That was true in every company I advised, every product I built, every team I watched struggle with decisions that depended on context nobody had written down. I wrote about why systems thinking is the skill that survives when everything else gets automated. This is what I meant.
Bridgewater taught me to take that instinct seriously. To build systems around it instead of working around its absence. The criteria should be explicit. The logic should be visible. The human should stay in the loop. The principles should be durable enough to survive whatever technology comes next.
That is what I am building toward. And reading Dalio's piece today reminded me where it started.
AI did not make relationships simpler. It made the cost of forgetting impossible to hide. — Jon Chu
