What Is This?
In the early 19th century, Charles Babbage designed a mechanical computer — the Difference Engine — that was conceptually correct and would have worked. It was never built. Not because Babbage was wrong, but because the manufacturing precision required to produce its parts didn't exist yet. The idea was right. The moment wasn't.
In 1900, six years before the Wright Brothers flew at Kitty Hawk, Nikola Tesla demonstrated a radio-controlled boat on the Hudson River. The crowd assumed it was controlled by a trained monkey hidden inside. The adjacent possible — people's ability to imagine remotely-controlled machines — hadn't opened yet.
In 1999, Pets.com raised $82.5 million, shipped dog food via FedEx, and went bankrupt in 268 days. The idea wasn't wrong. The infrastructure wasn't ready: broadband wasn't widespread, online payments weren't trusted, last-mile logistics weren't optimised. A decade later, Chewy built the same business for $3.35 billion.
This is the adjacent possible — the idea, developed by complexity theorist Stuart Kauffman and popularised by science writer Steven Johnson, that innovation doesn't happen in a vacuum. It happens at the edge of what's currently possible. At any moment, a specific set of ideas, technologies, and combinations can be reached from where we are. Not all ideas — specifically the adjacent ones. Each step that gets taken opens new adjacencies that weren't reachable before.^1
The concept began in evolutionary biology: how does chemical evolution work? At the origin of life, only certain chemical reactions were possible from whatever molecules existed. Each reaction that occurred produced new molecules, which opened new possible reactions. Life didn't explore all of chemistry at once — it explored what was adjacent, and adjacency expanded with each step taken. Innovation, Kauffman and Johnson argue, follows the same logic.
Why Does It Matter?
- It explains why the same idea emerges independently from multiple people at the same time. Darwin and Alfred Russel Wallace both developed natural selection simultaneously, independently. Bell and Elisha Gray filed patents for the telephone on the same day. Newton and Leibniz invented calculus simultaneously. This happens because when an idea enters the adjacent possible — when enough prerequisite ideas and tools exist for it to be reached — multiple people exploring adjacent territory will find it near-simultaneously. It's not coincidence. It's the structure of the adjacent possible being traversed by multiple explorers.^2
- It reframes the question every builder should be asking. The standard innovation question: "What's a good idea?" The adjacent possible question: "What has just become newly possible that wasn't possible two years ago?" The second question is more tractable and more reliable. Good ideas that are outside the adjacent possible fail — not because the idea is bad, but because the prerequisites don't exist yet. The most consistently successful innovators are those who accurately map what's just become reachable.
- It explains why so many "obvious" billion-dollar businesses only became possible in a specific year. Uber required simultaneous availability of: smartphones with GPS (post-2007), mobile payment rails (PayPal mature, Stripe launching), Google Maps API (open to developers 2005), and a critical mass of smartphone-carrying urban users (reached ~2009-2010). Any one of those arriving five years later would have prevented Uber from existing in 2009. The business idea was obvious — people have always wanted cheap rides on demand. The adjacent possible wasn't open until all four prerequisites arrived simultaneously. ^3
- It explains why some founders are serial successes while others keep failing. The most successful founders aren't necessarily more creative — they're better at sensing the adjacent possible. Paul Graham, Peter Thiel, and Marc Andreessen all describe some version of the same skill: the ability to sense what's about to become possible before the mainstream notices. This is adjacent possible awareness — a combination of broad knowledge, pattern recognition, and comfort with uncertainty about timing.
- It changes how to think about building now. The question isn't just "is this a good idea?" It's "is this idea in the adjacent possible right now?" Specifically: what existing capabilities does this rely on, do those capabilities exist and work reliably, what's the critical constraint, and is that constraint dissolving? If yes to all — the moment may have arrived. If a critical constraint is still locked — the Babbage problem.
Key People & Players
Stuart Kauffman — Complexity theorist at the Santa Fe Institute whose work on self-organisation and evolution produced the adjacent possible concept. His formal work is in theoretical biology and complexity science; the adjacent possible is one of his most generative ideas for non-biologists. His book At Home in the Universe (1995) develops the framework.^4
Steven Johnson — Science and technology writer who brought the adjacent possible to mainstream awareness in Where Good Ideas Come From (2010). He extended Kauffman's biological concept into a full theory of innovation — how environments (coffeehouse culture, coral reefs, cities) accelerate adjacent possible exploration by bringing diverse knowledge into contact. He now runs the Adjacent Possible Substack.^5
Brian Arthur (Santa Fe Institute) — Economist whose work on "increasing returns" and path dependence is the economic face of the same phenomenon. Technologies build on each other; early advantages compound; which technology becomes standard depends partly on which adjacent possible gets explored first. His book The Nature of Technology (2009) is the best companion read.
W. Brian Arthur + Kauffman — Together their work established that technological evolution and biological evolution follow structurally similar laws: both explore adjacent possibilities, both exhibit path dependence, both produce explosive combinatorial diversity once a threshold of complexity is reached.
Matt Ridley — His book How Innovation Works (2020) is full of adjacent possible case studies: how innovations consistently emerged from the accumulation of prerequisites rather than lone genius moments. The most historically grounded treatment.^6
The Current State
The adjacent possible is less a formal theory and more a powerful mental model — one that has been adopted by the technology and venture capital communities as a framework for timing bets.
The most clearly adjacent things right now (2026):
The adjacent possible expands every time a new enabling technology matures. The current wave of openings:
- AI-native applications — LLMs at human-level performance across many domains only became available 2022-2024. Applications that require this capability as a substrate (but weren't buildable before it) are currently in the adjacent possible. The window is perhaps 2-5 years before the space is mapped.
- Agentic automation — AI that can execute multi-step tasks autonomously (MCP, A2A protocols, Claude Cowork) only became viable 2024-2025. Workflows that require reliable autonomous AI execution are entering the adjacent possible now.
- Programmable biology — CRISPR, bioelectric manipulation, and synthetic biology tools have matured to the point that applications in medicine and agriculture that required them are now adjacent. Therapeutic applications of Levin's bioelectric work, for instance, are in the adjacent possible in a way they weren't five years ago.
The "multiple invention" signal:
One of the most reliable indicators that an idea is in the adjacent possible: multiple people are building the same thing simultaneously without coordination. If you notice two or three startups building essentially identical products within the same six months, the adjacent possible for that product has opened. The question is who will execute best, not whether the timing is right.
The premature innovation trap:
The opposite risk: a genuinely good idea built before its prerequisites exist. Symptoms: the product works technically but adoption is slow, users understand what it does but don't use it daily, the product requires users to change too many behaviours simultaneously. This is not a marketing problem. It's an adjacent possible problem — a critical dependency isn't mature yet. The honest response: either wait for the prerequisite to arrive (and re-enter when it does) or build the prerequisite itself.
Applied to personal development:
The adjacent possible applies to individuals as much as products. Your skills, knowledge, and networks are the prerequisites that determine what career moves are adjacent for you. Deliberately expanding what you know expands your personal adjacent possible — which is why broad reading, varied experiences, and cross-domain knowledge (the polymath argument) compound so dramatically over time. Each new thing you understand makes more things adjacent.
Best Resources to Learn More
- Where Good Ideas Come From by Steven Johnson — The most accessible treatment. Each chapter traces a pattern of innovation; the adjacent possible chapter is the best single explanation of the concept.^7
- At Home in the Universe by Stuart Kauffman — The foundational text, in the biology-of-complexity tradition. Dense but the source.^8
- How Innovation Works by Matt Ridley — The historical case studies. Best at showing the adjacent possible at work across centuries of technological change.^9
- The Nature of Technology by W. Brian Arthur — The economic and structural analysis. Complements Johnson's cultural analysis with a more mechanistic account of how technologies build on each other.^10
- Steven Johnson's Adjacent Possible Substack — Where Johnson continues exploring the idea with contemporary examples.^11