Generative AI:
The good vibes and the bad
What it is, how it works, and what it's already changing — explained without assuming you took linear algebra.
The deck: ~30 min, 24 interactive slides
Walk through the math underneath today's AI, a 5-second working "transformer," the actual architecture of a modern assistant, what's changing in coding and knowledge work — and the honest costs (energy, wealth concentration, jobs, ethics). Every chart and widget is editable; try your own numbers.
What you'll see
The math, plainly
Linear algebra without the algebra. A 3×3 mixer you can drag. A toy transformer that actually spells a word.
What's inside a modern AI
Hover over each piece of a real architecture diagram — safety filters, tool use, retrieval, the model core, and more.
AI & code
How "vibe-coding" works, why models pass software-engineering benchmarks now, and the honest caveats.
The new plumbing
MCP servers and "Skills" — the duct tape connecting models to the rest of the software you use.
Costs & trade-offs
Datacenter energy, hardware shortages, wealth concentration, and the entry-level jobs being hollowed out.
Where it's going
Three plausible 5-year scenarios — without hype and without doom.