From Eternal September to the Cambrian Explosion: Why AI's Evolution Might Be More Complex Than We Think
History doesn’t repeat itself, but it often rhymes
I've been thinking about how different technological eras have their own unique patterns of transformation. Eternal September – that 1993 moment when AOL brought millions of regular people onto the internet – was about sudden, massive democratization that completely changed who got to participate in digital spaces. Overlooked in this era were so many elements that haunt us today, primarily the loss of individual agency. This is something that needs retroactive work.
I've been exploring why AI hasn't had its equivalent moment yet, despite ChatGPT's record-breaking adoption. The promised democratization feels stuck, with most people still intimidated by prompt engineering and treating AI more like a novelty than a transformative tool.
Diving into insights about context engineering and abstraction layers reveals that we might be experiencing something entirely different. Each technological revolution has its own character, and our current moment is starting to look less like Eternal September and more like the Cambrian Explosion, the concept that spawned this Substack.
Think about it – the Cambrian Explosion wasn't one dramatic event where suddenly complex life appeared. It was this incredibly rich period of experimentation where all sorts of weird, wonderful, and ultimately influential life forms emerged simultaneously. Some survived and became the basis for everything that followed. Others were evolutionary dead ends. But the whole period was this explosion of diversity and innovation that fundamentally changed what life on Earth could become.
That's starting to feel more like what we're seeing with AI right now. We've got prompt engineering, context engineering, agentic systems, multimodal models, and probably a dozen other approaches I haven't even heard of yet. Some of these will probably turn out to be evolutionary dead ends – interesting experiments that don't quite work at scale. Some will merge, and evolve into something entirely new. But others are going to become the foundation for whatever comes next. And that’s just the ‘software’ side of things; hardware is going through it’s own explosion, as the fundamentals of integrated circuit development are being rewritten.
The thing about the Cambrian Explosion is that if you were living through it, you probably wouldn't have recognized it as this pivotal moment in the history of life. You'd just see a lot of weird creatures trying different approaches to survival. The significance only becomes clear in retrospect.
We are in AI's Cambrian period right now, as every layer of the technology stack - seemingly - is unsettled and exposed to massive evolutionary change. All this experimentation with different interfaces, different ways of managing context, different approaches to making AI more accessible – maybe that's not a sign that we haven't figured it out yet. Maybe it's a sign that we're in the middle of figuring out something much more complex and interesting than a simple democratization moment.
Sure, the path forward looks clouded and uncertain. We're still dealing with hallucinations, prompt ambiguity, and the fact that most people find AI more frustrating than empowering. But underneath all that noise, the fundamental building blocks are coming together. Context engineering is maturing. Natural language interfaces are getting more sophisticated. The infrastructure for truly human-compatible AI is being built piece by piece. It is also the time to correct a failing from the Eternal September era, and guarantee that individual agency is factored in the core DNA of every evolved layer. (More about this in coming posts, but
does an excellent job of explaining this opportunity with the case for MyTerms. Full Disclosure: Along with Doc, I am a board member of Customer Commons.)Progress doesn't always look like progress when you're in the middle of it. Sometimes it looks like chaos, experimentation, and a lot of false starts. But that might be exactly what revolutionary change is supposed to look like.
The explosion is happening. We're just still figuring out what survives.