Скачать или смотреть From Nazi Psychology to AI Auditing: Inside the System I Built

Along the way we get into how software systems actually communicate (APIs and the Model Context Protocol), why the same engineering pattern that gave us USB and HTTP is now connecting AI to external tools, and why the word "agent" is doing more work than the technology behind it currently justifies.

This is the third video in a series. The first traced a line of Python down to electrons. The second connected Mad Libs to AI memory. This one starts at the top and takes an AI system apart to find traditional engineering at every layer.

TIMESTAMPS:
0:00 - A test designed to study authoritarians, now running on AI
1:15 - What "agent" actually means (and doesn't)
3:30 - The orchestration layer: async programming and concurrency
6:30 - How systems talk to each other: APIs explained
9:30 - MCP and the pattern: USB, HTTP, LSP, and now AI tooling
12:00 - The data and the math: parsing, scoring, statistics
15:00 - The reframe: what's actually inside these systems
16:00 - Close

THE ETHICS ENGINE:
Paper: https://arxiv.org/abs/2510.11742
Code: https://github.com/RinDig/AuditEngine
App: https://ethicsengine.eduba.io

PREVIOUS VIDEOS IN THIS SERIES:
How One Line of Python Triggers 12,000 Lines of Code
   • How One Line of Python Triggers 12,000 Lin...  

How a 1953 Word Game Explains AI Memory
   • How a 1953 Word Game Explains AI Memory  

TOPICS COVERED:
AI agents, software architecture, APIs, Model Context Protocol, MCP, psychometric assessment, authoritarianism research, RWA scale, async programming, orchestration, JSON, HTTP, USB, protocol standardization, Language Server Protocol, Python, SciPy, NumPy, statistical analysis, Cronbach's alpha, test-retest reliability, AI auditing, ethics, Frankfurt School, prompt engineering, context engineering, software engineering

CONNECT:
Website: https://eduba.io
LinkedIn: /in/jakevanclief

#softwareengineering #programming #ai #python #api #computerscience #coding #tech #machinelearning #psychology #aiethics #mcp #artificialintelligence #opensource #datascience