devotd.
Browse the Asia/Seoul (KST) issue calendar. Retrospective entries show their assigned coverage day; actual publication times remain separate.
Cloudflare’s October 12 technical post states that a failing Logpush Transformer does not send untransformed raw logs instead. The transformation boundary holds, but if failures continue, records in the affected batches are eventually dropped and cannot be recovered.
A Project Maya contributor reported 26.0 tok/s decode on one RTX 3090 and 53.0 tok/s on two. The two-card setup also changed the layer split, MTP drafting and prefill chunk size, so it cannot be read as a scaling factor from GPU count alone.
The DNS root began signing with KSK-2024 at 14:00 UTC on October 11. Validating resolvers, down to recovery images and standby systems, need to trust key tag 38696, and ICANN’s guidance is not to assume automatic updates succeeded.
llama.cpp merged runtime support for Prism Bonsai 2 27B on October 11. The model card and backend docs checked on October 13 still require the Prism fork, so file packing, runtime revision and backend combination need separate checks.
PXA 3.1 adds an Overdrive preset for older NVIDIA cards such as the Tesla P100 and V100, but its speed work assumes PXA quant packs and the PXQN library bundled in the package. Keeping a learned expert map outside a container also needs a named volume.
Basalt forks Strata into an engine for one target, Qwen3.8-Flash-Next on Linux and NVIDIA Blackwell (sm_120). The developer’s reported 665 and 354 tok/s come from a particular model pack, workload and tuned configuration; built-in defaults and context splitting under concurrency shape the result just as much.
Bun 1.4.3 adds bun check, a type checker ported from typescript-go, and a --check flag that gates running, testing and bundling. File-level --check follows only static imports, so covering workers or preload files still needs a separate whole-project check.
strata-mlx is an independent MLX engine that reads Strata's model files. A user's measurements on a real 32GB MacBook Air showed that cutting the expert budget from 20.5GB to 16GB nearly doubled generation speed and sharply reduced cumulative swap-out.
uv 0.13.0 changed the default Python download, when no version is requested or pinned, from 3.14 to 3.15. Handling of --require-hashes and editable entries in constraints files included with -c also became stricter, so installs that used to succeed can fail.
Ai2 replaced a priority-based scheduler with GPU-time budgets, hierarchical fair-share and a minimum-runtime contract. The reported 98% figures are cluster occupancy and delivery of owed time; whether useful computation survives preemption is a separate question.