Nate B Jones is selling a product, and I think the product is pointed at something real.
A portable context server, MCP-shaped, pitched as the missing infrastructure for AI working intelligence. The framing is sharp and I think the framing is right. There’s a layer between you and the model that nobody’s standardized, and whoever standardizes it owns something valuable.
Here’s the thing. I already have it. Not because I bought it. Because I built it over five months of disciplined instrumentation, one decision at a time, without ever calling it infrastructure.
My personal AI system is named Bishop. And structurally, Bishop is already most of what that product promises.
i.The four layers
The pitch breaks working intelligence into context layers the model needs to actually be useful to a specific person over time. Roughly: who you are and what you value, what you’ve decided and why, what you’re working on right now, and where every artifact came from.
I went down the list against what Bishop already has.
Three of four, strong, with the fourth identified as the gap. That’s not a coincidence. That’s what five months of building the thing the product is selling looks like from the inside.
On building it yourself A product can sell youthe schema. It can't sellyou the five months.
ii.Bought versus earned
The contrarian point isn’t that Nate’s wrong. He’s right that the layer matters. The contrarian point is that you don’t have to wait for someone to ship it.
The infrastructure that makes an AI genuinely useful to you over time is mostly discipline, not software. None of it required a vendor. It required showing up and writing down the decision and the reason, every time, for five months, instead of letting the context evaporate at the end of each session.
A product can sell you the schema. It can’t sell you the five months of actually populating it with your own ground truth. That part you earn.
iii.What the gap tells me
The useful output of this exercise wasn’t “I don’t need the product.” It was finding the one layer I’m thin on.
Artifact provenance is the next thing I build. Not because a video told me to. Because mapping my own system against an external framework surfaced exactly where my system is underbuilt. The external pitch became a diagnostic for my own architecture.
That’s the move I’d recommend to anyone building this for themselves. Don’t buy the framework. Use it as a checklist against what you’ve already got. The gaps it finds are real even if the product it’s attached to isn’t the answer.
iv.The frontier is an invitation
There’s a version of watching the AI-infrastructure market that’s pure FOMO. Someone shipped a thing, I’m behind, I’d better buy in. That version makes you a customer of your own future.
There’s another version where the market is a map of what’s worth building, and you check it against what you’ve already earned. That version makes you an owner. Five months in, three of four layers strong, the fourth one named and next. I’ll take owner.
Drafted with Bishop, my AI partner.
Words picked, edited, and approved by me.
Model provenance: Claude Code (Claude Opus and Sonnet)