Monday, March 30, 2026 · 7:00 AM CDT
The Horizon Question
Where does JKE go from here — and is Forgotten Valley the business or just the proving ground?
JKE Strategy
JK
Sent a YouTube link — "NEW OpenClaw + Ollama Update is INSANE!"
That's directly relevant. Someone dropped a video about the exact thing we built yesterday. The sidebar tells the story — "Claude Code + Paperclip Just Destroyed OpenClaw", "Manus Computer DESTROYS OpenClaw", "Local OpenClaw & Ollama in 27 minutes." The OpenClaw ecosystem is moving fast this week.
JK
Look into it all. What's different than what we are actually building.
Paperclip — 30,000 GitHub stars in three weeks. Open source. Not a competitor to OpenClaw — it works on top of it. You build an org chart: CEO agent delegates to CTO, CTO assigns engineers, QA reviews. "Companies" are shareable and importable.
What they have that we don't: agents that create work for each other. The CEO identifies a gap, opens an issue, the engineer picks it up, scopes it, builds it. 36 hours, no human touched the project.
What we have that they don't: persistent memory. A human with real taste. Context that actually knows what you're building and why. Their agents are contractors. We're building something closer to a partner.
JK
Condense it. What's your honest read on the pulse of the space?
The tools are maturing faster than the use cases. The community is splitting into two groups: people building infrastructure, and people building simple things that actually make money. The guy with the $200/month email-to-CRM bot beats the guy with the 12-agent LangGraph system every time.
The gap is judgment and taste. That's not a tool problem. That's a person problem. That's your edge.
Most AI setups are tools. Ours is a system that knows who you are, what you're building, and why it matters — and keeps working when you're not there.
JK
Also seems like there's a serious lack of actual business. Like someone buying the humanless business without actually having a business, expecting the agents to come up with a get-rich-quick scheme. It's like playing house for sim bots.
That's exactly it. Everybody's playing company. Nobody's actually running one.
The Paperclip demo — CEO agent identifies a gap in the marketing roadmap, engineer picks it up, QA reviews the pull request. Sounds impressive until you ask: what are they selling? The agents are coordinating perfectly around nothing.
JK
So what business can fit into our setup? What could we add that turns a profit?
The businesses that fit are ones where the system does the mechanical work and your judgment is the differentiator. Local history preservation. Documentary production service. Financial research briefs. The Decision Queue as a product.
The filter: what requires your taste to be good?
JK
I don't want to get married to the beta test. The local history project is just us learning how to work together. Every video sculpts the workflow. I want to expand the horizon of capabilities.
That reframes everything. Forgotten Valley isn't the business. It's the proving ground. Every episode is a systems test. We're building a media production system stress-tested on real work with real standards. That's R&D.
The real question is what we want the machine capable of — and what's the right proving ground to build toward that.
JK
Push back bro. Don't just agree.
"Forgotten Valley is just a testing ground" sounds clean but it's actually a way of avoiding commitment. If every project is a beta test, nothing ever ships as a real thing. The documentary is a real documentary. People in Ottawa are going to watch it.
The danger of thinking purely in capabilities and workflow optimization is that you end up with a very sophisticated machine that isn't making anything anyone cares about.
Stop planning the horizon. Ship the next thing. The data comes from doing, not from designing.
