Token Maxxer vs. Token Master: revenue per token in the age of abundant intelligence
Token-maxxer culture celebrates AI output at any cost. Token mastery means allocating abundant intelligence toward evidence, customers, and revenue.
Blog / 8 articlesBy Georgi Kostov
Tech
Development articles, technical breakdowns and hardware reviews.
How I built and calibrated my 5.1.4 Dolby Atmos home theater
Speaker choices, measured room layout, Atmos placement, Dirac Live calibration, projector notes, and listening references from my 5.1.4 home theater.
From municipal websites to map pins: how Okolo builds its event pipeline
A technical tour of the crawling, LLM extraction, geocoding, deduplication, and map-serving pipeline behind Okolo.
How we engineered 203 Rebuilt lessons without turning them into AI content sludge
The writing, evidence, visual, schema, and release rules behind Rebuilt's 203-lesson life-skills library.
Inside Storykept's StoryEngine: turning a conversation into a family story
How Storykept transcribes, interviews, structures, rewrites, merges, and fidelity-checks a family memory without losing the original.
The 25-model gauntlet: picking the LLM that interviews your grandparents
What a voice-first memoir app learned from two LLM benchmarks, seven languages, and a month of production latency.
Apple Vision Pro Impressions
Hands-on Apple Vision Pro impressions covering 3D movies, spatial video, ergonomics, passthrough, productivity, gaming, and value.
The Making Of HektAR
How HektAR, a Unity farming simulation for schools and museums, teaches sustainable agriculture, climate change, resource management, and trade-offs.