The Three Waves of Zero

In 1999, a CD cost $16.99. Fifty minutes of music encoded on a disc that had to be manufactured, shipped, stocked, and sold. By 2010, that same music was a file. By 2020, it was a stream — one of a hundred million tracks available for $9.99 a month.

That was not a format change. It was a phase transition in how civilization assigns value to a fundamental human capability: creativity.

Here is the pattern that most people have not yet generalized: human capabilities are being digitized and automated one by one, in sequence — and each domain follows the same deflationary arc.

First came the creative capabilities — the human capacity to produce images, moving images, and sound. Photography, film, music, design, writing. These were the first to be encoded into digital form, the first to become infinitely reproducible, and the first to see their per-unit value collapse. The music industry was simply the canary.

Now comes the next layer: the human capacity for ingenuity, problem-solving, and the creation of business tools and software solutions. For forty years, the dominant model has been to package that ingenuity into persistent software products — operating systems, SaaS platforms, B2B tools, dashboards, workflows — and sell access to them. That entire model is now entering the same phase transition.

And beyond software lies a third domain: the material capabilities — the production of food, goods, housing, physical infrastructure. The things that sustain life in the physical world.

Each domain follows the same arc. Each will transform the economics of everything built on top of it. And much of the industry is still underestimating the magnitude of what's coming.

Three eras, one pattern

There is a recurring structure to how civilization digitizes human capability. It moves through three phases, and each collapses the economic logic of the one before it.

The Ownership Era. Value lives in the artifact. You buy the CD, the software box, the license. Scarcity is enforced by physics or by law. Make a thing, sell the thing.

The Access Era. Value shifts from the object to the library. You don't buy songs; you subscribe to all of them. You don't buy software; you subscribe to a platform. The individual artifact loses standalone value. Breadth, curation, and convenience become the product.

The Generation Era. Value shifts again — from the library to the moment of creation. You don't access a pre-built tool; you describe what you need, and it materializes. The artifact isn't purchased or accessed. It's instantiated on demand — tailored to a single user, a single business, a single task — then used and potentially discarded.

A note on what "zero" means throughout this piece: not that all value disappears, but that the marginal value of the generic unit collapses, and durable value migrates to new layers.

Wave one: creative capabilities

Music, photography, film, design, writing — the human ability to create images, moving images, and sound. This was the first domain to be fully digitized, and the first to experience the collapse of per-unit value. Physical scarcity disappeared. Copies became free. Ownership weakened and access became the dominant commercial model. Streaming was not merely a new format — it was a new economic logic.

Now, with generative AI, creative production is entering the generation era. Music, images, and video can be produced on demand, personalized, and iterated in real time. The recording, the photograph, the design asset — these are no longer scarce. The entire creative stack, from conception to finished output, is being compressed into a prompt.

Wave two: problem-solving capabilities

For decades, software companies captured value by packaging the human capability for problem-solving and business logic into persistent tools. Find a recurring problem, build a product that solves it, sell the product at scale. CRMs, analytics platforms, project management suites, operating systems — all are crystallized human ingenuity, sold as artifacts.

Generative AI begins to unbundle that model. The specific tool may no longer need to exist in advance. The dashboard, the workflow, the analysis, the interface can be created at the moment the need arises.

The deeper shift is this: we are moving from the interface era to the outcome era. In the old paradigm, you interacted with software to get a result. In the emerging paradigm, you describe what you need and receive what you asked for — the tool, if it existed at all, was ephemeral. The user doesn't operate software; they receive an outcome. The entire layer of persistent, branded, packaged software between the human and the result begins to thin.

This is not a uniform process. Music is consumed; software is used to act. A business workflow must connect to live systems, handle permissions, produce reliable results, and operate within regulatory constraints. High-reliability systems, compliance infrastructure, and regulated industries will hold onto the access-era model longer. The first domains to shift are the ones closest to pure information work — analytics, reporting, internal tooling, content workflows.

But the direction is clear. Most of today's SaaS companies — the ones whose primary value proposition is generic workflow functionality — will not survive this transition. The companies that emerge on the other side will be unrecognizable, in the same way that buying a CD at Tower Records is unrecognizable from asking an AI to compose a soundtrack for your morning.

The interface keeps getting simpler

There is a parallel trend that accelerates everything: the access point itself keeps getting cheaper and simpler.

The first software required a desktop computer and technical literacy. Then laptops. Then smartphones. Each generation of hardware lowered the cost and complexity of accessing digital capability. The next step is already visible: voice-only and dialogue-only interfaces, where the interaction is a conversation rather than navigation of a screen.

This matters enormously. When the interface to generated software is a spoken sentence — available on a $50 phone — the barriers to entry collapse across geography, literacy, and economic status. The endpoint is not a better app. It is the disappearance of the app as a concept, replaced by a fluid layer of capability accessible through the simplest possible human interaction: speech.

What retains value when the tool is free

When recorded music lost its per-unit value, three things gained value: live experience, curation, and artist identity. The artifact died; the ecosystem around it thrived. The same migration is beginning in software.

Proprietary data and context. The AI can generate the dashboard, but it cannot invent the data that makes it meaningful. Organizations with rich, structured, domain-specific datasets become more valuable, not less.

Judgment — knowing what to ask. The tool is free, but the insight that determines which question matters is domain expertise. In the generation era, taste and judgment become scarce assets.

Trust and verification infrastructure. When anyone can generate anything, the question shifts from "can we build it?" to "can we trust it?" Provenance, attribution, auditability — these become critical infrastructure.

This last point is where I spend my days. My current venture, Musical AI, is building attribution infrastructure for AI-generated music: the system that traces which human-created works an AI model drew from, ensuring original creators are properly credited and compensated. The future of creativity is not human versus machine. It is human and machine, properly attributed — co-creating, with value shared and exchanged fairly. That requires infrastructure that doesn't exist yet. We sit at the intersection of wave one and wave two, building a trust layer that exists precisely because creative generation has become trivially easy, and the question of origin, rights, and fair compensation has become the hard, valuable problem. It's one piece of a larger puzzle, but it's taught me that the most durable value in any generation era lives in the infrastructure of trust.

Even the AI layer commoditizes

The major AI companies are locked in a fight for dominance that mirrors the early streaming wars — and they know the model layer itself is heading toward commodity status. Today's frontier model is next year's baseline. Margins compress. Open-source alternatives proliferate.

The model layer is becoming infrastructure — closer to electricity or bandwidth than to a durable product. AI companies are subject to the very deflationary logic they are imposing on everyone else. Value doesn't vanish in these transitions. It migrates.

The 80% question

Most AI discourse centers on the part of the world that already has these tools. But roughly 80% of humanity does not yet have meaningful access to frontier AI — or to reliable connectivity, affordable compute, culturally adapted interfaces, and the educational scaffolding to make any of it useful.

The deflationary dynamics in developed markets create a crisis for existing software companies. Globally, they reveal an enormous construction opportunity — arguably the largest addressable market in history. Billions of people need localized infrastructure, multilingual AI, regulatory frameworks, and pathways to participation. The builders who bring AI-native capability to the next four billion people are not doing charity. They are building essential infrastructure for the next phase of civilization.

What this means if you're building now

If you are founding or running a software company today, the question to sit with is: where does your value actually live?

If your moat is generic workflow logic — a better interface for doing something that an AI agent will soon do on command — that moat is evaporating. If your moat is proprietary context, deep system integration, verified outputs in high-stakes domains, or irreplaceable data — it may be strengthening.

In the near term, expect internal dashboards and reporting to become generated rather than bought, light SaaS features to become indistinguishable from agent capabilities, and incumbents to defend themselves through data, distribution, and compliance rather than through the product itself.

The companies that thrive will not be the ones that build the best tools. They will be the ones that own the layers tools cannot replace.

The deepest obstacle — and the third wave

If the first wave digitizes creative output and the second digitizes cognitive tooling, the third will eventually reach the material world — manufacturing, agriculture, housing, food, physical goods. Slower, because atoms resist copying. But the convergence of AI-driven design, advanced robotics, and new fabrication methods points toward the same destination: material abundance.

And here we arrive at the biggest obstacle in the entire arc — one that is not technological at all.

Every institution, every economic model, every power structure humanity has built is organized around the management of scarcity. Scarcity of food, scarcity of tools, scarcity of shelter, scarcity of capability. Our markets price based on scarcity. Our politics distribute based on scarcity. Our psychology hoards based on scarcity.

When the material conditions begin to shift toward abundance — when creative output is generated on demand, when software is instantiated at the moment of need, when material production becomes radically cheaper — the deepest challenge is not building the technology. It is the civilizational paradigm shift required to actually live in abundance. To reorganize economic models, governance, education, and our own psychology around the reality that the fundamental constraints are changing.

This is not a problem any single company or technology solves. It is the defining transition of the coming decades.

Every previous wave of technology has eventually forced a deeper question: once a constraint is removed, what becomes worth optimizing for? When creative production is abundant, when software is on-demand, when material scarcity begins to ease — the genuinely scarce resource may be something that resists digitization entirely. The quality of human attention, care, judgment, and presence. The well-being of eight billion people, not just the fraction already inside the system.

In every generation era, the winners are not the ones who produce more artifacts. They are the ones who build the layer that scarcity moves into — and who help all of humanity make the leap from a world organized around scarcity to one that can finally be organized around something better.

Nico — CTO & Co-Founder, Musical AI. Building attribution infrastructure at the intersection of human creativity and artificial intelligence.