Weekly impact chains
Curated chains of what each AI release changes across orgs, products, and techniques.
The impact chain tracker
Map the ripple of every AI release
Weekly impact chains across orgs, products, and techniques.
US and CN ecosystems on one canvas. Free to read.
What you’ll find here
An atlas of weekly AI motion
Multi-layer map
Six layers from silicon to end user — see where a shift lands and where the ripple ends.
US and CN coverage
Both ecosystems on one canvas. Qwen and Claude sit on the same graph, not two blog posts.
Timeline and graph
Switch between the relational graph and the timeline of when each node moved.
Markdown-native chains
Each chain is a markdown file. Review diffs, suggest edits, and submit via the repo.
Free to read
The chains and visualization are public. Paid work happens in the consulting lane.
This week
5 chains across 20+ nodes
Each release mapped to its downstream effects on adjacent orgs, products, and techniques. Hand-curated, reviewed, versioned in markdown.
Layers L1 - L6
From silicon to end user
Six layers cover hardware, foundation models, middleware, applications, organizations, and end users. No layer swept under the rug.
US + CN
Both ecosystems, one canvas
Qwen, Claude, Gemini, DeepSeek — same graph, same coordinate system.
Open
Markdown + Git
Every chain is a PR-reviewable markdown file.
Testimonials coming soon
We’re in v1. Once teams have used Unmapped for a few weeks, their words go here.
FAQ
Questions we keep getting
What is an impact chain?
An impact chain is a small directed graph that links one AI release to the downstream products, techniques, and orgs it changes. Chains are hand-curated and reviewed.
How often are new chains published?
Weekly. Each Monday a fresh batch of chains is merged and the site redeploys automatically.
Do you cover Chinese AI companies?
Yes. Qwen, DeepSeek, Moonshot, Zhipu, and others sit on the same graph as US labs. We treat the two ecosystems as one system, not two blog posts.
Is the data free?
Yes. The chains, graph, and timeline are free to read. Consulting engagements are where we bill.
Can I suggest a chain?
Yes. Every chain is a markdown file in the repo. Open a pull request or email us.
What are the layers L1 - L6?
L1 silicon, L2 foundation models, L3 middleware and frameworks, L4 applications, L5 organizations, L6 end users. A chain usually touches two or three adjacent layers.
Why no login or paywall in v1?
Because signal quality wins subscribers faster than paywalls do. If and when a paid tier ships, existing readers will be grandfathered.
Start mapping the ripple.
Explore the full graph, or get the weekly digest when the next batch of chains lands.