-
Sun, Nov 1, 2026 9
-
H5 energy run (laptop), probe +32% at 256 tokens, audit update, paper energy section
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Gate G: bootstrap statistical audit (table in paper), manifest generator
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Paper: abstract/intro/limitations reflect measured verdicts (H1, H4, E)
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Gate F script (Qwen3 two-node, cold local vs odd/popular placement), nocache flag, LOG (probe depths, host-key fix)
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Figures (sim, capacity sweep, two-node), paper results v0.1 with two-node evidence, LOG
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
SSD tier: parallel demand reads + expert-contiguous TITAN layout (cold 16-slot 8-10 -> 36 tok/s); sim v2 results; probe fix; LOG/DECISIONS
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Qwen3-30B parity on M3U96a (bit-identical, 121.7 vs 121.5 tok/s), make_tables.py, cold-NVMe sweep script, LOG
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Paper skeleton v0.1 (protocol + preliminary), README, models config, stacked-layout registry (Qwen3-30B), generalized tests
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-
Gate A/B: scaffold, mandate, inventory tooling, prior art (59 refs), registry/store/slot-table runtime, numerical tests bit-identical to mlx-lm
…
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
-