Reposition: the small-GPU and modest-RAM champion, for everyone

README.md

@@ -13,14 +13,16 @@ pinned: false
 
 A hob is an English household spirit that quietly does the work. Hob Forge exists for one
 reason: **to help you actually run good models on the hardware you already have.** We build
-**3B–4B finetunes and merges — the best little agentic models** — with the NVIDIA Jetson
-Orin Nano Super class (8GB unified memory, 25W) as our sharpest target, and we ship every
-model with the exact way to run it: step-by-step guides, tested configs, real support.
+**3B–4B finetunes and merges — the best little agentic models** — and we are unapologetically
+**the champion of small GPUs and modest RAM**: every release states exactly what it needs at
+8GB, 12GB, and 16GB (and what CPU-only costs you), and ships with the exact way to run it —
+step-by-step guides, tested configs, real support.
 
 Every release promises:
 
-- **A stated memory budget.** Quant sizes and KV-cache maths against the board's ~6GB
-  practical envelope, in every card. If it doesn't fit with real context, we don't ship it.
+- **A stated memory budget.** Quant sizes and KV-cache maths per hardware tier — 8GB, 12GB,
+  16GB — in every card, measured, not projected. If it doesn't fit with real context at a
+  tier, the card says so plainly.
 - **Honest evals.** Held-out benchmarks only, decontamination stated, baselines included —
   the raw base and a same-size peer, same harness, same settings. We publish the harness
   output, not vibes.