The Napster Moment for AI

The revolution went open source. The gatekeepers don't stand a chance.

A few weeks ago, a developer named Matt Shumer published an article called “Something Big is Happening.” When I read it, it had around 36,000 views.

It hit 55 million within days.

The piece laid out what a lot of us in tech have been feeling in our bones: AI capability has hit an inflection point. Massive job displacement is coming fast. And the people paying attention now will have a serious head start. Definitely worth the read.

I’d been thinking the same thing for months. I ran into an old college friend at Billy’s memorial—surrounded by grief and nostalgia and the weight of time—and we ended up talking about AI for an hour. Not because we wanted to. Because we couldn’t stop. Something big is happening, and neither of us could look away.

Shumer just wrote it down eloquently and struck the zeitgeist.

The Lobster That Broke GitHub

Around the same time, a weekend project called Clawdbot exploded onto the scene. An Austrian developer had built a simple tool connecting AI models to messaging apps, and giving it wide authority to perform unsupervised action on it’s host machine. It racked up 9,000 GitHub stars in its first 24 hours.

Then Anthropic sent a trademark notice—”Clawd” was too close to “Claude”—so it became MoltBot. Then OpenClaw. The name changed twice in 72 hours, and nobody cared. They just kept starring.

157,000 stars in 60 days. For comparison, Kubernetes took about three years to hit 100,000. OpenClaw was 18 times faster.

Why? Because it gave everyday people a taste of what AI agents can actually do. Not chatbot stuff. Real automation. Direct access to your tools, your data, your workflows, all with a human friendly communications interface.

It’s essentially a harness around your LLM that drives autonomous operation, creates Input output via tools like Telegram and email, and allows all manner of other tool connectors. There’s more to it that I don’t want to diminish, but it marked a fundamentally different way to use and interact with AI that your typical chat interface.

People wanted that power badly enough to overlook some genuinely terrifying security implications. There was a spike in Mac Mini sales as Custafarians (Yes, ClawdBots created their own “religion,” evidently) rushed to stores to provide hardware and compute to their AI assistants.

I work in enterprise identity security. I can tell you: large enterprises won’t touch OpenClaw. They’re still working out security implications for basic LLM use cases while their legal teams draft acceptable use policies.

But millions of individuals and small teams? They dove in headfirst. Jason Calcanis has his team building OpenClaw agents they call “Replicants“ to replace their own busywork. His plan: let each employee level up as the agents absorb their routine tasks.

(Note: After a quick podcast media tour including a few hours on Lex Friedman, OpenAI acqui-hired Peter Steinberger, the altruistic creator of OpenClaw for an undisclosed sum. The official statements suggest OpenClaw itself will remain open source. We’ll see. This is how fast the space is moving.)

I’ve seen this before.

$20 for Twelve Songs

When I was a kid, I saved up $20 for a CD. That’s about $43 in today’s money—for twelve songs, most of which I didn’t pick.

Then Napster showed up and blew the whole thing apart.

For those who need the refresher: Napster was a free app that let you share music files directly between personal computers. By February 2001, it had 80 million users. Anyone with an internet connection could find and download virtually any song ever recorded. Instantly. For free.

The music industry sued it into oblivion. But the genie was out of the bottle.

Napster didn’t just disrupt an industry. It demolished a bloated and entrenched cartel of middlemen and gatekeepers who profited from their control of the entire pipeline—from which artists got chosen to be “made,” to promotion channels, to distribution, down to the physical media formats.

That system wasn’t working for listeners. It wasn’t really working for most artists either, though a select few obviously benefited tremendously. It primarily served the “industry”—the agents, producers, and distributors. The horror stories from that machine are endless.

The mp3 revolution cut out layers of middlemen. Now we have Spotify, where we can listen to almost anything for about $20 a month. Coincidentally, the same price as that CD. Any artist can upload and compete for an audience. It’s not all sunshine and roses—artists are still generally underpaid—but the point is what the technology instigated.

As Stewart Brand said: information wants to be free.

OpenClaw is Big AI’s Napster moment. Once people see the power that’s becoming available, they want it. The security risks are real and troubling. But people are going to build around it anyway.

I’m curious how the industry digests all this. They have to recon with it.

The End of the Middlemen

This is a bold claim, but I’ll say it: the era of VC-backed SaaS as we know it is over.

The entire tech stack is being re-envisioned. All the capital is going into AI infrastructure right now. And once that infrastructure matures, bloated user experiences will no longer be tolerated.

APIs and MCPs give AI agents direct access to create, read, update, and delete your data—the four basic operations that every piece of software performs. These agents can generate bespoke interfaces based on what a user or organization actually needs, rather than forcing everyone through the same one-size-fits-all dashboard.

I’m not saying this happens overnight. Enterprise software procurement cycles alone could outlast some startups. But the revolution is upon us. Nothing will look the same in three years.

The Missing Piece

Credit Integral Life for the visuals. https://integrallife.com/agency-and-communion/

Here’s where my head is at in all this right now.

The philosopher Ken Wilber describes two fundamental drives in any system: Agency—the drive for individual autonomy and action—and Communion—the drive for connection and collaboration. Every healthy system needs both. They exist in dynamic polarity. Too much agency without communion leads to isolation. Too much communion without agency dissolves identity.

If you want to travel fast, go alone.
If you want to go far, go together.
~The Internet

Right now, AI tools are almost entirely focused on agency. All the major platforms are built around a single person interacting with tools to increase their individual output.

But how do we work with these tools as a team? What does the communal landscape of AI empowered work look like?

Here’s an example from my world: my marketing VP creates an incredible Claude Code skill for LinkedIn posts. Totally on brand, understands our value proposition, researches existing conversations, guides toward effective content. How do they share it with the team so everybody can use it? Sure, they can pass the file around and hope everyone installs it. But what if someone improves it? How does that update get shared back? How to we all benefit from the expertise and work of one of our team?

The coming wave isn’t just more powerful agents. It’s figuring out how we work together using shared tools and skills—and continually evolve and improve as a team. “How we work” has always been a passion of mine as my title suggests: Director of People & Operations. And most of what I see in AI right now is completely individual focused.

OperatorOne

Which brings me to what I’ve been building.

OperatorOne is a prototype—an AI-agent-first backend stack for small businesses. It addresses some of what I’ve been talking about.

I started by working to address some of the core security concerns that OpenClaw made famous, AND working in ideas for how to make working with agents a team sport.

The team collaboration piece? It’s a first pass, but it’s built in from the start. Users can create and maintain Agents(I call them “Operators” and skills. The supervisor calls the sub-Operators as needed.

I pull in Authentik for SSO, Grafana for logs, Paperless NGX for docs and a few other open-source goodies, all deployed in Docker containers. (I integrated Paperless this morning, as an example of how this vibe-coding thing goes. It’s Fast)

The stack includes a supervisor agent to route inquiries, a system status agent to health-check the infrastructure, and a workflow agent that can directly create automations in a self-hosted n8n instance. It connects to your existing tools via secure API, supports the LLM of your choice, and runs on open-source identity and functionality tools.

Think of it as a small business AI starter kit with the base level security handled—something that can be customized and scaled by people who don’t have tech geeks like me to hack around on it. A low-cost accelerator for small businesses who want the power of an AI-augmented workforce.

This IS NOT production software. It’s a prototype. A proof of concept. It’s open source. It’s early. And it’s an application of my small thesis to a very big question.

If you want to start working with it, let me know and I can advise.

I include a Claude.md file so you get get a head start on customizing this using your own Claude code instance. Fork it, hack on it. Make it work for your business. Or get in touch with me and I can lend a hand.

---

Nobody knows how much devastation this revolution will leave in its wake. Every major technological shift creates winners and losers, and this one is moving faster than anything we’ve seen.

This invention of the tractor put 90% of farm hands out of work, but it took 100 years. It also tripled total farm output. This AI revolution will have similar dramatic impact, but it is happening at light speed.

If Napster taught us anything, it’s that once the technology exists, the old guard can’t hold the line. The middlemen get cut. The power shifts. And what emerges on the other side—messy, imperfect, still evolving—is usually better for the people who actually use it.

The revolution is here. I’d rather be building than watching.

~capshaw

If you need a little sonic inspiration while paddling in on this wave, here’s my vibe-coding playlist.