Ideas "Physics and Chemistry" with GenAI
I have discussed (here, here, here, and elsewhere) how AI can help problem-solving and idea generation, mainly by boosting unexpected collisions between concepts. In this article, I lay out some notions leading to building better workflows and organizational structures that natively harvest the emergent opportunities in our problem-solving and innovation ecosystems.
The Structure of an Idea and the Power of Collisions
This will not be a theoretical discussion. We know well that our ideas - our organization's, our team's, our own - often simply result from interaction with other ideas through connecting with people, reading about something new, etc. The flow of our days is a constant opportunity to generate new ideas, not because we think about things harder but because we bump into things that trigger them. Leaders try to foster those collisions, including through organizational and process design or encouraging the formation of active knowledge networks. Those collisions happen daily, at scale, and intelligence emerges from those interactions.
When ideas collide—whether with other ideas, analogies, or subcomponents of the original idea—they generate derivative concepts. These derivatives can take many forms: new ideas directed at solving the initially stated problem, clarifying assumptions, or even clarifying doubts that spark further inquiry. Understanding this process is key to leveraging AI for idea generation and innovation.
Let’s take the iPhone as a simple, well-known example. Its creation wasn’t just the merging of existing technologies like cell phones, iPods, or computers. Instead, the components collided and evolved, yielding the iPhone and countless other ideas. For instance, the removal of the physical keyboard and mouse marked significant shifts in user interaction design. This demonstrates that ideas, when recombined, are not just additive—they can transform into something entirely novel.
At the heart of this process lies the structure of an idea, which I like dissecting into three components: the why, the what, and the how. This is a version of a process called morphological analysis, which uses many different parameters - but I find these three to be both insightful and practical for our purpose here.
For illustration, imagine an idea as a molecule comprising atomic subcomponent ideas. The subcomponents of these three (why/what/how) components are connected and relate to each other, which could be formally represented as a knowledge graph.
1. The Why: Purpose and Importance
The "why" represents the motivations and values driving an idea. It explains why the idea matters and what problem it solves.
The iPhone's "why" included enabling communication between family and friends or others, making a fashion statement, and acting as a tool for emergencies, entertainment, and productivity.
The "why" helps uncover the emotional and functional needs an idea fulfills, often through tools like user interviews, ethnography, and failure analysis.
2. The What: Categorization and Context
The "what" defines the problem categories an idea addresses. It connects the motivations of the "why" to broader contexts.
For the iPhone, the "what" included overlapping categories like phones for communication, music players for entertainment, computers for productivity, and payment and authentication systems for security.
The "what" helps locate an idea within an ecosystem, revealing its connections to adjacent fields.
3. The How: Feasibility and Execution
The "how" encompasses the technologies and systems that bring an idea to life.
For the iPhone, the "how" included durable materials like Gorilla Glass, AI-powered on-screen keyboards, Foxconn’s scalable manufacturing ecosystem, Apple’s design philosophy, and the App Store infrastructure.
By bridging the "what" and "why," the "how" transforms concepts into actionable solutions.
When Ideas Collide
Idea collisions create fertile ground for new concepts. Consider two subcomponents of the iPhone: fashion statements and entertainment.
Fashion Statements:
Entertainment:
Recommended by LinkedIn
When these two collide, derivative ideas emerge. For example, combining fashion and entertainment creates intersections like shared memories, self-expression, and social bonding. Apple’s "Memories" feature exemplifies this, using AI to surface curated photo and video highlights connected to people, places, and moments.
Collisions don't just help with the divergent part of idea development - they also help refine ideas. A special type of idea collision is "idea hardening": Every idea, no matter how groundbreaking, begins fragile and imperfect. For it to become robust and actionable, it must undergo a process of hardening—a deliberate refinement in which assumptions are tested, weaknesses are exposed, and components are iteratively improved.
Hardening isn't just about identifying flaws—it's about strengthening an idea's why, what, and how, ensuring it has the clarity, feasibility, and resilience needed to succeed.
This hardening process often relies on critique techniques such as persona-based evaluations, failure analysis, or "what-if" scenarios. These methods simulate how an idea might perform under various conditions or perspectives. Through iterations, feedback loops, and testing, ideas evolve into mature solutions that can withstand real-world challenges.
And crucially, the techniques used for hardening are special lenses and embed the knowledge of the people who have created them. In a way, they force ideas through some contact with specific realities - many of them have been constructed by humans.
Ecosystems of Ideas and the Role of AI
The world's idea-generation process doesn’t happen in isolation, in a purpose-built vat. Humans engage in "natural experiments" daily, generating and refining ideas through interactions with people, environments, and, now, machines.
The scale of this is massive. We are talking about trillions of interactions daily, an unstoppable chain of collisions whose outcome, with some notable exceptions, is so far captured manually or in barely digitized workflows and processes. AI of many types changes that landscape, as AI can now access publicly available data and, increasingly, organization-specific knowledge. It can use the signal from knowledge graphs showing the relationships between ideas and the one from network structure analysis (who or what "says" what, how they're connected, how central they are in the network, etc.).
AI, with its capacity for interpolation and abstraction, enhances this process, complementing human (and human systems') extrapolation and serendipity.
This interplay between human creativity and AI's computational power expands the innovation landscape.
Idea Physics and Chemistry: a New Discipline?
The "idea physics and chemistry" concept, a sort of combinatorics, envisions ecosystems where ideas collide and evolve, supported by AI. It is not a theoretical construct. I see it as a set of principles and frameworks driving the design of organizations, for instance, maximizing their "collision surface" and enabling their people to process that emergent knowledge.
By harnessing these techniques and tools, we could even create "chain reactions" of innovation, exploring (not just generating) new ideas at an unprecedented scale.
Imagine a world where AI helps us harvest ecosystems of ideas, optimizing the way knowledge flows and transforms within or outside of organizations. This supermind—a networked intelligence emerging from human-machine interactions—could hugely benefit how we innovate, solve problems, and create value.
By understanding and leveraging the anatomy of ideas and the power of collisions, we unlock the potential for continuous, scalable innovation. The future of ideas lies in how we connect and recombine the "why," "what," and "how" with the tools of tomorrow.
We now have the tools and practices to build this—it is time to do it.
This article is part of a series on AI-augmented Collective Intelligence and the organizational, process, and skill infrastructure design that delivers the best performance for today's organizations. More here and in the white paper here.
Get in touch if you want these capabilities to augment your organization. Build learning, problem-solving, innovative, intelligent organizations. Build Superminds.
Insightful