Every research note the shop has shipped — on-device models, grounded decoding, mesh trust, and the reviews that tried to tear them down. Tap to open.
// a solo engineer's paper trail, measured not vibed
A fine-tune is a small delta on a shared base. Measure that delta for Feedseed and compress it (SVD × quantization × Matryoshka nesting) as hard as quality allows.
The architecture the whole ladder hangs on — nested sub-models where facts come from retrieval and skills come from the weights, so the model can never hallucinate a name it wasn't handed.
Trimming a 262k vocab to a 48k FC-core cuts a third of the params and the per-token bytes with it. The governing constraint behind every speed claim in the stack.
Re-measured every ONNX dtype through onnxruntime. A tiny model has too little weight redundancy for 4-bit — RTN noise collapses borderline tool-selection.
Why a 4× smaller model can hit half the bandwidth efficiency of a bigger one — per-token fixed overhead, KV traffic, and the runtime path, ranked by the formula.
The model emits the tool-call structure; the name slot is filled from a ranked retrieval enum with phonetic + fuzzy matching. The browser twin of MoM's grammar enum.
100% valid tool-call JSON with a real tool and a real retrieved name, enforced at decode time in the browser — grammar fixes structure, the model handles which tool.
Design for pulling waypoint names off physical trail signage so the grounded name set matches what a hiker is actually looking at.
The cascade that decides which brain answers, and the guarantee that the emergency path never depends on any of them.
A signed, hash-linked Merkle log (Git + Certificate Transparency, not a blockchain). The same leaves that prove a build is authentic reveal exactly which inflated chunks are wrong, so a node re-pulls only those and proves each before patching.
A per-device self-signed CA + IP-scoped leaf (and iOS .mobileconfig) satisfies the secure-context wall offline, so the app — model included — installs from the phone next to you with no signal.
Independent teardowns of the claims. The flagged one: an early "keystone" number was retrieval-top-1 with no model on the causal path. Re-measured 2026-07-27 with the production brain (Gemma E2B) actually in the loop — model-selection over retrieval's top-8 scores 90% (36/40), far above chance and its own no-rail 15%, but slightly under retrieval's 100% rank-1. So names come from retrieval, the model from intent — the honest version, on record.