Tests on Qwen 3.6 27B show how an experimental BeeLlama KVarN preview shifts KV-cache quality down a memory tier: q5 quality at q4 size, q6 at q5, and q8 at q6.
Tests on Qwen 3.6 27B show why TurboQuant is overrated but saved by TCQ, q5 deserves more attention, and symmetric q8 might be a waste of VRAM.
How I built 3 massive AI mods for Paradox grand strategy games (Stellaris, Victoria 3, Imperator: Rome) in a scripting language that doesn't even have arrays.
Advanced AI game rule, road building system, fort placement optimization, and a comprehensive overview of all AI improvements.
Mercenary recruitment, governor policy rework, research efficiency-based building decisions, and economic policy improvements using the new modifier: syntax.
Added law management, tribal reform, diplomatic stances, state investment, character loyalty, and economic policy improvements.
Reworked AI decision making for buildings and inventions, introduced the Invention Relative Chance system, and improved city founding logic.
Anbeeld