Selected Work · AI Practice

Finance × AI
Knowledge Network

A local system that turns financial and AI sources into a searchable, reusable network. One source leads to its key concepts; each concept opens into related mechanisms, context, and my own judgment.

Role
System design + AI collaboration
Stack
Obsidian · Web Clipper · AI
Status
Active · Continuously growing

Financial work produces an endless stream of sources, terms, and opinions. The real problem is not scarcity; it is that concepts, underlying mechanisms, and personal judgment remain disconnected. I wanted a network that compounds in value—not another folder of things I once saved.

From one source to connected understanding

The goal is not to save a webpage somewhere else. A source is extracted, connected, and reviewed until it becomes a reusable knowledge asset.

01 · SOURCE INPUT
NEW SOURCE · TO PROCESS
If rate-cut expectations are rising, why can long-term yields still move higher?
02 · EXTRACT & CONNECT
TreasuriesTerm premiumInflationFed policy
03 · KNOWLEDGE ASSET
01Structured concept pagessearchable
02Backward links across nodesconnected
03A reviewed personal judgmentreusable
📥
STEP 1
Capture
Save a useful article with its title, link, and body in one click
🔍
STEP 2
Extract concepts
AI identifies domain terms such as gold, Treasury bonds, or real rates
🔗
STEP 3
Open a concept page
Each term gets a concise explanation and links to related concepts
🕸️
STEP 4
Expand the network
Follow the links as adjacent mechanisms and context come into view
🧠
STEP 5
Build understanding
Turn a transient source into context that remains useful later

The system compounds: every source adds another set of connections. Sources provide input, AI assists with structure, and I remain responsible for verification and judgment.

Open one term. Reveal the system around it.

This is not a decorative diagram. Change the entry point, then open any node to see how a term connects to mechanisms, judgment, and a useful output.

Interactive knowledge map
NOW EXPLORINGWhy gold rises
Direct driver Mechanism Extended link
Open a node to light up its immediate relationships

What is already in place

VAULT SNAPSHOTSynced at build time · 2026.08.23
01 · CONCEPTS
111
Structured concept pages
02 · DOMAINS
5
Growing knowledge domains
03 · SOURCES
21
Sources and market logs
04 · JUDGMENTS
11
Personal viewpoints and judgments
How the 111 concepts are distributedActual pages by domain
FIN
Finance
64 concepts
AI
AI
31 concepts
COG
Cognition
9 concepts
TECH
Technology
5 concepts
OPS
Content Operations
2 concepts

More than a notebook

The difference is not “saving more.” It is whether an old source can help produce an answer when a new question appears a month later.

BEFORE · BOOKMARK FOLDER
Saved,
but not connected
Thirty days later: I remember seeing it, but still have to search and rebuild the explanation.
AFTER · SECOND BRAIN
Every input adds
another relationship
01
Source: rate-cut expectations rise
02
Concepts: Treasuries · term premium · inflation
03
Mechanism: short policy rate ≠ long-end pricing
04
Judgment: cuts do not guarantee a long-bond rally
Thirty days later: a new question hits reviewed nodes instead of restarting from search.
CAPTUREOne-click browser capture
CONNECTBidirectional concept links
RETRIEVEDashboard and topic search
OWNLocal, durable plain text

One source → one concept → a set of mechanisms → a better judgment.
A knowledge system is valuable when it makes the next decision faster and more grounded—not when it merely stores more.

More selected work