Dean-aligned Llama(): 5 kwargs only; hf_hub_download at module scope; poll 25s

Tier 1 safety: n_ctx 8192, TypeError guard on perf kwargs

Tier 1 perf: page-cache warm + n_ctx 6144 + q8_0 KV + mlock + poll 10s

CPU fallback when ZeroGPU quota exhausted; GPU/CPU chip in topbar

Drop Flux 2; 4 starting chars; pin Ada=engineer Bram=diplomat

Named-place movement: atomic move_to, drop coord reasoning, 10/gather, per-archetype reward, default llama-3b, smart-mock updated

Recording prep: JS mutex, mixed models (Ada/Bram=Nemotron, rest=SmolLM2), drain=3, init=3, duration=90s, idle-drift, mock latency profile

Sync latest local snapshot

Mock: tighten SoC-detection so it doesn't swallow normal decisions

Mock: templated SoC per archetype style for local testing

Smart mock: tune social pull mid-range

Smart mock: dial down social pull, walk away after posting

Smart mock: social pull to town hall + bump post probability

Smart mock: context-aware tool emissions + shell-session pattern

Townlet: initial private upload with Zero-GPU fixes