Zeroth-Principles Thinking
How does one come up with a truly original idea? One that reshapes society for centuries.
Learning about the discovery of the number zero, and its philosophical representation — nothingness — changed my orientation with transformational thinking.
My first thought after reading Zero: A Biography of a Dangerous Idea was that I wanted to build a sturdy mental model for exploration of analogously difficult-to-see concepts. I wanted to find capital-Z Zeros.
It was intuitively what I had been trying to do for years after selling my company Braintree Venmo but I just didn’t have words or concepts to make concrete what I was consciously experiencing. The question I was trying to answer once I sold the company and could explore anew: What one thing can I do to help humanity thrive most? What is missing from the zeitgeist but is essential?
Silicon Valley was full of “first principles” companies and ideas from engineers and designers. But every “first principles”-derived possibility I evaluated felt like it fell short of what was expansively possible and what humanity needed right now for its future existence. Thinking outside the box can lead to productive insights, but that wasn’t enough; I wanted to call the box itself into question.
Now, I think about Zeros everyday as I run Kernel and OS Fund, two companies trying to usher in the future of what it means to be human. Each is simultaneously enabled by first-principles thinking and fueled by Zero discovery time-locked to meaningful horizons in all of our futures — 2025, 2050, and 2500.
Zero discovery is difficult to express with the words and concepts at our disposal today, but I know the emotional profile of finding a Zero. In the professional analysis of one of AlphaGo’s games against the best human Go player in the world, one observer remarked that the moves were like watching the AI play “Go from an alternate dimension.” That’s the profile. That’s what it feels like to witness Zero discovery.
Since hearing that, I keep coming back to this “from an alternate dimension” idea. To the trained eye, AlphaGo seemed to play with new ideas or strategies that were either entirely new or ancient enough to have been abandoned by modern professionals. The AI did not take our human assumptions as a given and played with new conceptual primitives lurking inside the structure of the game that not even experts could see. The moves already existed. A human could have played them, if they knew to. They were always possible. But something about the human brain made us never see them, even after we iterated for thousands of years of Go culture and strategy. Why not?
AlphaGo found zeroes in the game of Go, but Go is a finite, complete-information game. My primary interest: How do we find the best moves — those ”from an alternate dimension” — for the future of intelligent existence?
No previous generation has ever had the opportunity that we do right now: to look out over our expected lifetime and see the real possibility of evolving as/into entirely novel forms of conscious experience. An existence so transformed that our current selves, and today’s realities, will appear historically primitive and uninteresting.
What might these be? Zeros by definition cannot be defined and are hard to imagine. Some that come to my mind: What if we weren’t motivated by social status, others approval or wealth accumulation? What if we felt no tribal proclivities? What if we primarily cared about harmonious play, of a type we don’t even have words or concepts for today, with all things around us? (What [enter your most unintuitive, counterintuitive and unlikely assertions] are you thinking of that I can’t see?) After all, our modern world would appear all those things, and equally foreign, to our ancestors.
These ruminations even a few decades ago could have easily been dismissed as practically impossible or as a waste of time for anyone actually wanting to accomplish something. (As we build our Autonomous Selves, our mental energies will be freed up to explore these frontiers.)
Hunting for Zeros is relevant for everything currently top of mind, from how we address climate change, how we adapt to technological disruption, how we build AI, how we govern ourselves, and how we evolve and motivate our reasons to keep going, to keep existing.
Recently, I’ve condensed these ideas and have started practicing “Zeroth-principles” thinking, which aims to uncover transformative new elements of thought. It is a twin with “first-principles” thinking. Zeroth-principles thinking is about building blocks, or the structure of all things, whereas first-principles thinking is about system laws, or how things interact. First-principles thinking sets goals in known terms and then pursues them, inventing and learning new things as necessary, but it rarely uncovers brand new conceptual primitives of the kind which the world needs.
When people say, “From first principles…” to start their argument, what they mean is that they wish to assume as few things as possible from within a given frame. Innovation in space exploration is a good example of the good that can come from first principles thinking alone. Our agenda in space today is valuable, difficult, and we don’t have all the answers for what’s out there in the universe, but we do have a good sense of the questions and the building blocks. Few new building blocks have been discovered for space travel in a long time — we’re still working out the implications and laws of the same basic concepts from the 1970s and, arguably, all the way back from Newton.
Sherlock Holmes is a classic first-principled thinker, achieving success as a detective with his own brand of careful logic: “When you’ve eliminated the impossible, whatever remains, however improbable, must be true.” In this, Holmes — trapped in a book, after all — can presume a closed universe of objects where, if X and Y are the only things that exist, not-X is the same as Y.
The deductions of Dirk Gently, however, the “holistic detective” created by Douglas Adams as a contrast to Sherlock Holmes, have a different approach: “I don’t like to eliminate the impossible”. This is Zeroth-Principles thinking. Seeing through the blind spots.
In a murder investigation, Holmes wouldn’t entertain the possibility, or even the hypothesis, that time travel could play a role. Dirk Gently, though, wouldn’t rule it out. In part, because how does he know what is and isn’t possible yet? The two fictional detection styles illustrate a fundamental distinction in logic.
What would it mean to say, “From zero principles…”?
Zeroth-order Technology
The future of human existence is not deducible from first-principles thinking alone. We need Zero Explorers. If we can’t fully model where we are going or what we can aspire to, it’s challenging to create practical plans which have clear goals and which galvanize large-scale cooperation. I think that is a feature, not a bug.
Our co-evolutionary future with AI will introduce a record-breaking number of Zero-like building blocks, which will in turn level up our aspirations. When brain interfaces allow for the real-time pairing of one’s mind with AI, we may experience a Cambrian Explosion-like emergence of Zeros. Our subconscious is to our thinking as deep learning is to AI which, when harnessed, will be capable of producing thoughts that seem like they are “from another dimension.”
By definition, you can’t “get to Zero” or introduce a novel element into thinking directly, so you have to try and stress first-principles efforts to the extremes until there are no marginal gains left. How can humans, the system architects of intelligence, increase the speed of Zero discoveries?
First, some more examples from tech and science:
- Andrej Karpathy’s Software 2.0 vision is zeroth-order software engineering that uses “abstract, human unfriendly language” to achieve an objective via a deep neural network (DNN). This in contrast to Software 1.0 where a human writes explicit code to achieve an objective. Perhaps Karpathy’s paradigm could be labeled Software 0.0?
- Daniel Dennett’s concept of “Joosting” (Jumping Out of The System) aims at discovering Zeros.
- Einstein uncovered a Zero when he hypothesized that the speed of light might be constant for all observers, which uncovered a new conceptual primitive. Literally a “move from another dimension”.
Or, consider the historical example of European geometry before and after the introduction of the zero (the original, literal number/concept). Everything Europeans thought they understood about math had to be rebuilt from the ground up to include the concept of zero once it was introduced. It’s not that Euclid’s Elements from the Greek era suddenly became obsolete or worthless. The power of zero was that it allowed existing mathematics to be rebuilt for a more powerful understanding of geometry that included the concept of zero. Which led to Cartesian geometry — including the idea of the origin of the coordinate system on the plane at (0, 0) — and, suddenly, geometry and algebra could be connected conceptually via the decimal system.
Before Europeans learned about zero, Descartes could not have come up with Cartesian geometry. All anyone had ever tried was to do better and better Euclidean geometry.
We need more such insights. In a world where we are continuously expanding our spheres of understanding, each Zeroth-principle insight can potentially unlock a set of more expansive spheres. This is bigger than just an exponential effect, of the kind popularized by Ray Kurzweil, the “up to the right” / hockey stick / knee-of-the-curve style stuff. Zeros are game changers. The graph is not just exponential — the units change. The graph reorganizes its axes. Dimensions are added to accommodate ideas “from another dimension.” That’s what Zeros get you.
You don’t know what big goals are worth pursuing until you’ve uncovered the new building blocks and understood their properties.
Here’s the big question, the Zero Explorer mantra: What concepts are hiding in plain sight but can’t be seen by anyone? How do we, with our limited human cognition, start thinking “from an alternate dimension”?
Instead of aspiring to be number one, be a zero.
