Personal Reference Guide

Understanding AI Infrastructure:
A Map of Inquiry

Eight domains of inquiry, organized so that each conversation thread can stay focused without losing sight of the larger picture. Items marked ✓ Covered have been addressed in earlier sessions.

How to use this guide
Use this as a checklist before starting a new conversation thread. Pick a domain, identify where you left off, and begin there. Ask for a summary at any point to capture what's been established before the thread grows too long.
Domain Key
Physical / Infrastructure
Political / Regulatory
Technical / How It Works
Societal / Human Impact
Economic / Financial
01 Data Centers: Physical Reality Physical
Where are data centers being built — Pacific Northwest, California, nationally? ✓ Covered
What are data centers actually used for — AI, cloud, surveillance, entertainment? ✓ Covered
Water cooling: the problem and the alternatives ✓ Covered
Power consumption: how much electricity does the AI buildout require?
Land use: how large are these facilities, and what do they displace?
Who owns what — Microsoft, Google, Amazon, Meta footprints?
Undersea and modular data centers — experimental or viable?
Waste heat reuse: capturing and redirecting thermal output
02 Energy & Environment Physical
What energy sources power data centers — coal, nuclear, hydro, renewables?
How does AI power demand interact with regional power grids?
Carbon footprint of training vs. running AI models
Nuclear revival: are data centers driving new reactor investment?
Hydropower dependency in the Pacific Northwest — vulnerabilities?
What happens to power grids when AI demand spikes?
Water table and watershed impacts of large-scale cooling operations
03 How AI Actually Works Technical
What is a large language model and how is it trained?
Training vs. inference — what's the difference in resource terms?
What are GPUs and why does AI require so many of them?
What does it mean for a model to "hallucinate"?
Open-source vs. proprietary AI — what's at stake?
AI agents: what are they and how do they differ from chatbots?
How do AI companies improve their models over time?
04 Surveillance, Privacy & Data Societal
What personal data is collected and stored — and by whom?
How is behavioral data used for advertising and profiling?
Government access to data center contents — legal frameworks
Facial recognition and AI-enabled surveillance infrastructure
What protections exist — GDPR, CCPA, federal gaps?
China's AI surveillance model vs. Western approaches
Biometric data: what's being collected without clear consent?
05 Politics, Policy & Regulation Political
Which government entities govern data center approval — local to federal? ✓ Covered
What questions should voters ask candidates at each level of government? ✓ Covered
Federal AI legislation — what exists, what's proposed?
State-level AI laws — California, Washington, Oregon leading?
Tax incentives for data center construction — who benefits?
The EU AI Act — does it affect U.S. companies?
AI and antitrust — are a few companies too dominant?
Water rights law as applied to data centers
06 Labor, Economy & Displacement Economic
Which jobs are most at risk from AI automation — near term?
Which jobs and industries are being created by AI?
Data center construction: local employment vs. long-term local jobs
Who profits from the AI buildout — concentration of wealth?
Small businesses and AI: tool or threat?
How much does it cost to train a frontier AI model?
AI and healthcare costs — productivity vs. access
07 AI in Daily Life Societal
AI in healthcare: diagnostics, drug discovery, patient records
AI and education: tutoring, cheating, institutional response
AI in creative fields: art, music, writing — tool or replacement?
AI and accessibility — benefits for people with disabilities
AI and misinformation — deepfakes, synthetic media
Social media algorithms as a form of AI — how they shape behavior
AI companions and mental health — benefits and concerns
08 The Bigger Picture: Risk & Future Societal
What is AGI (Artificial General Intelligence) and how close is it?
AI safety — what are researchers actually worried about?
Concentration of AI power: a few companies vs. global access
AI and democracy — election interference, propaganda at scale
International AI race — U.S. vs. China, strategic implications
What does a world with abundant AI look like in 10–20 years?
Who decides how AI is governed — corporations, governments, or publics?
On Thread Strategy
Physical and environmental topics (Sections 01–02) work well together in one thread. Political and regulatory topics (05) can be woven in when they directly arise from a physical question, or kept separate for focused policy discussions. Sections 03 and 08 are good starting points for any conversation about the technology itself. At any point, ask for a summary to capture what's been established before starting a new session.