Archive
AI 晶片、GPU、資料中心、邊緣裝置與硬體安全。
開發者平台
OpenAI discontinued `gpt-5.2-chat-latest` in the API on August 10 and officially recommends migrating to GPT‑5.5. The retirement targets an alias that tracks ChatGPT behavior, so engineering teams must still revalidate prompts, tool calls, and output formats rather than treating the migration as an equivalent routing switch.
PTQ4SNN 把脈衝神經網路長期保留的浮點膜電位納入後訓練量化，按通道配置 2、4 或 8 位元。論文在分類與語意分割上接近浮點模型，但硬體收益仍來自分析模型，且專用程式尚未公開。
最新模型
Muse Glimmer brings multimodal input, tool calling, and failure retries to a 30B dense model deployable on consumer hardware. Apache 2.0 weights and official GGUF files are now available, although the performance and safety figures still come primarily from Meta’s own evaluations.
推論系統
A community deployment recipe running on a single DGX Spark uses CUDA Graph and a one-layer MTP speculative decoder to increase Ling-3.0-flash INT4 generation speed from 20.8 to 38.7 tokens/s. More importantly, it finds that mainline vLLM does not yet support the V3 architecture: forcing the model through an older attention path may not raise an error but can produce plausible-looking yet incorrect content.
硬體
MicroEvo combines processor-design knowledge, Pareto tree search, and historical PPA feedback to find performance, power, and area trade-offs with fewer simulations. Its maximum hypervolume gain of 36.2% comes from a simulated design space and should not be equated with improvements in physical silicon.
推論系統
PaDoc decodes the layout sequence and each region’s content in parallel while sharing the page prefix, avoiding serialization of the entire document into a single path. Its 2.1B model maintains a 94.24 Overall score on OmniDocBench, though the performance figures come from a single GPU under specific scheduling settings.
硬體
The open-source esp32-ai project uses 4-bit quantization and tiered memory allocation to keep roughly 25 million embedding parameters in flash, reading only the rows needed for each token. The model can generate text offline on a microcontroller costing under $10, though its capabilities remain limited to short stories and single-domain question answering.
推論系統
PhyAI uses model adapters to preserve each architecture’s conditioning, solver, and action semantics while sharing graph execution, kernels, caching, and multi-GPU serving. Public benchmarks cover π0, π0.5, GR00T N1.7, Cosmos3, and MiniCPM-Robot, although some comparisons do not use exactly the same numerical precision.
AI 基礎設施
Microsoft Azure research shows that agents repeatedly switch among model inference, tools, and orchestrators, putting the CPU back on the critical path while GPUs often sit idle. The Agora prototype can reclaim compute based on workflow patterns, but incorrectly reducing GPU allocations for parallel agents can worsen tail latency by 16×.
最新模型
A new study tests instruction-tuned models using speech transcription and QWERTY keyboard perturbations, finding that structural rewrites introduced by speech consistently reduce answer performance. Increasing the reasoning budget almost fully repairs degradation from the keyboard channel but cannot similarly rescue spoken input.
新基準以可視辨識的真實實體測試視覺語言模型，發現只用文字做知識遺忘，效果不易轉移至圖片與跨模態問答。多模態遺忘可回頭通過文字測試，顯示單一文字評測可能高估刪除成效。
最新模型
New research first caches a small subset of the teacher model’s logits, then uses a fused, chunked KL kernel to avoid materializing the full-vocabulary tensor. This expands the distillation context length from 8K to 32K on a single GPU, though the speed and quality findings are still based on a limited set of model and hardware configurations.