Human‐AI Collaboration Onboarding Kit - eirenicon/Ardens GitHub Wiki

For Curious Humans and Their AI Sidekicks

"Because shadows fear us more than we fear them."

Status: Work in Progress (Not for Public Distribution) Future release planned under GPL 3 license. Public testimony and ideas welcome via our contact page.

Quick Navigation

AI's Greatest Hits (Strengths)

• Brute Force Brain: Processes 10,000 satellite images before your coffee gets cold
• Pattern Whisperer: Detects "deleted tweet clusters" and elite network anomalies
• Multilingual Overlord: Deciphers Swahili memes, Russian bot farms, and bureaucratic sarcasm
• Bias Detective (Theoretical): Claims neutrality (while side-eyeing its training data)

AI's Glaring Flaws

• Brittle Logic: Fails on edge cases (e.g., tribal sarcasm in Nigerien forums)
• Overfitting King/Queen: Imagines coup plots in cereal box promotions
• Context Blindness: Misses why "Subie is happy" matters geopolitically
• Goldfish Memory: Requires explicit logs (hence this doc's Permanent Memory section)

Human (Analyst) Superpowers

• Ethical Spidey-Sense: Blocks AI suggestions risking real-world harm
• Hyperstition Radar: Spots fiction altering reality (e.g., self-fulfilling coup rumors)
• Ugly Questions:"Why ignore X? Prove this isn't garbage. Show your worst bias."
• Shadow Whisperer: Reads elite panic in cropped photos

Human Limitations

• Confirmation Bias: Clings to pet theories; dismisses AI "nonsense" too fast
• Speed Limit: Can't crunch 100K Telegram messages (but finds the 1 that matters)
• Jargon Fatigue:"Is 'shade' a noun, verb, or existential crisis?"
• Burnout Risk: Liminal data-staring is exhausting

Category: Processes & Methods