Open Source Is Dead—Long Live the Open Skill
- The value of open source is shifting from shared code to shared intelligence—skills, prompts, and workflows that guide AI agents.
- OpenClaw’s rise signals a new software paradigm: autonomous, self-improving, and memory-driven, where the AI acts without waiting for prompts.
- As AI regenerates tools instantly, the real scarcity isn’t code—it’s human insight embedded in reusable decision frameworks.
Open source is breaking. Not because it failed, but because it succeeded too well.
AI agents can now rebuild any application from scratch in minutes. GitHub repos? They’re ephemeral. The real value has moved upstream—to the skills that tell AI what to build, how to act, and when to improve itself. OpenClaw’s skills.md isn’t documentation. It’s the new source code.
Dave Morin didn’t join the OpenClaw Foundation to protect a codebase. He joined to protect a paradigm. After being "claw pelled"—his agent building him a CRM, then optimizing his YouTube thumbnails by reverse-engineering MrBeast’s team—he saw the shift: this wasn’t automation. It was collaboration with memory, intent, and rhythm. The Heartbeat system, which triggers autonomous action every five minutes, turns AI from a tool into a teammate.
On Presidio Bitcoin Jam, David King contrasted that world with New Orleans, where no one cares about AGI. There, human irreproducibility—live music, handmade meals—is the premium. In tech, the mirror is emerging: when machines can do everything, the differentiator is not what they do, but how we guide them.
The code is no longer the artifact. The skill is. Ashu’s 90-hour podcast indexer wasn’t valuable because of its implementation—it was valuable because it knew what to extract and how to surface it. That know-how, encoded in prompts and workflows, is what gets shared now.
Open source isn’t gone. It’s evolved. The license isn’t on GitHub. It’s in the shared playbook of action.
Dave Morin, This Week in Startups:
- This was the first time I felt like I've been living in the future since chat GPT.
- The idea that the machine has a heartbeat and it will do things each time it does this heartbeat is really thoughtful and interesting.