Welcome to
Hey MHK
A digital third place for thinking out loud about AI, data, and the infrastructure underneath.
This is where I cultivate ideas. Some polished enough to share widely, others still growing. Not work, not home, but somewhere in between.
The Highlights
Finished work ready for the world. Essays, series, and ideas that have matured enough to stand on their own.
↓ Scroll down750+ Connected Notes
The living knowledge graph. Ideas at every stage of development, linked together and growing in public.
→ Visit the gardenBrowse All
AI: From the Engine Room
13 Articles-
01
Mechanisms →I've Heard This Engine BeforeIntroduction to the series
-
02
Mechanisms →The Attention Mechanism ExplainedWhat attention actually does
-
03
Mechanisms →Training EconomicsWhat training costs and what it buys
-
04
Mechanisms →The Efficiency IllusionWhat efficiency gains actually deliver
-
05
Mechanisms →Context Windows and MemoryWhat models actually remember
-
06
Data →The Multiplier EffectHow reference data transforms proprietary data
-
07
Data →The Tacit Knowledge BottleneckExpert knowledge resists capture
-
08
Data →Why Vectors Aren't EnoughThe case for structured knowledge
-
09
Data →UI as the Ultimate GuardrailDesigning interfaces for systems that can mislead
-
10
Governance →Hallucination is a Feature, Not a BugUnderstanding why AI makes things up
-
11
Governance →Working With the MachinePrompting as interface design
-
12
Governance →AI Readiness Is a Governance QuestionWhy the work before AI is the work that matters
-
13
Governance →What the Engine Room Taught MeSynthesis and looking forward
Who's Here
I'm Mike Kuehne. I work in applied AI: architecture, strategy, and the infrastructure that makes it real.
Before the current AI wave, I was building attention-based neural networks for NLP research. Before that, I built a knowledge graph from public domain data to understand how products relate to each other.
The through-line: I'm interested in how knowledge hides in relationships, and how to surface it.