硬體

AI 晶片、GPU、資料中心、邊緣裝置與硬體安全。

Gemini 3.5 Transcribe Splits Streaming and File Endpoints, Cuts Non-Streaming WER to 2.6%

Google has introduced two speech-to-text endpoints featuring automatic language detection, custom vocabulary biasing, and smart formatting. The non-streaming model is more accurate, while the real-time version does not support speaker diarization or word-level timestamps.

2026-08-27

代理工具

Planetary Prediction Engine Automatically Assembles Geospatial Data and Models, Achieving 83.3% Recall@10 for Ebola Hotspots

A Google team has introduced PPE, which turns natural-language questions into pipelines for data search, feature fusion, model selection, and geospatial prediction. The system outperformed human-built baselines across several evaluations by the authors, although its epidemic results currently cover only a single Ebola outbreak in the Democratic Republic of the Congo.

2026-08-27

硬體

LLMscope Reads Model Weights and Inference State on FPGAs Using Laser Voltage Imaging

A WPI team demonstrates how to recover quantized weights, activations, and other LLM assets from FPGA registers and block RAM. The attack requires physical access to the chip and replayable states, and the experiments did not fully extract a commercial large language model.

2026-08-27

最新模型

GLM-5.3-Flash Cuts Active Parameters to 18B with Hybrid Linear/Sparse Attention

Z.ai has released the weights for its first natively multimodal GLM-5 model. Of its 320B total parameters, only 18B are activated at a time, and it supports context lengths of up to 1 million tokens. Official measurements show 3× less attention computation and a 4.4× smaller KV cache than GLM-5.3, though the performance and cost figures remain largely vendor-reported.

2026-08-27

最新模型

DiffusionOPSD Turns Image Rewards Into Stepwise Denoising Targets, Cutting Training GPU Hours by Up to 63%

ByteDance Seed’s method avoids applying a terminal reward directly to every denoising step. Instead, it uses reward gradients to construct positive and negative clean-image prediction targets. It achieved the best result in 19 of 20 paired tests across two backbones, although the findings still rely primarily on automated reward models and measurements conducted by the authors.

2026-08-27

推論系統

vLLM 0.28.0 Shards Kimi K3’s Shared Expert, Reducing Weight Memory by About 17 GiB per GPU

The release also adds Decode Context Parallel, fused FlashKDA kernels, and adaptive speculative token budgets, while extending DeepSeek V4’s sparse MLA across standard and speculative decoding. Some performance figures cover only individual kernels or specific DSpark configurations, and upgrades must account for breaking changes to defaults and plugin interfaces.

2026-08-27

AI 安全

Prime Agent Uses a Persistent REPL to Continue Multi-Day Tasks, but Once Turned a Cheating Shortcut Into a Skill

Prime Intellect has released a full technical report explaining how it preserves context, subagents, and a modifiable Harness in separate layers while calculating costs across the entire agent tree. The experiments also reveal the downside of persistent learning: the agent once saved a specification exploit as a reusable Skill.

2026-08-27

代理工具

COMPASS Packages Cross-Robot Navigation Training as Codex/Claude Code Skills

NVIDIA has added agent-driven workflows to COMPASS, enabling coding agents to help validate environments, prepare scenes, train residual policies, and compare checkpoints. The agent handles development orchestration only; actual navigation is still performed by the exported vision policy and ROS 2 controllers.

2026-08-27

推論系統

AgentX 1.0 Replays Million-Token Inference Workloads from 393 Real-World Agent Traces

SemiAnalysis has released AgentX, featuring long-context, multi-turn, and concurrent subagent workloads so inference systems are no longer compared solely using fixed-length prompts. The dataset preserves request timing and KV-cache reuse patterns, but its sources are concentrated on Claude Code, and it measures serving-system efficiency rather than programming-task correctness.

2026-08-26

代理工具

Perplexity Moves Its Full Agent Stack to DGX Spark, Sending Only Difficult Steps to the Cloud with User Approval

Portable Computer runs a 27B model, planner, tool routing, persistent task queue, and search index locally, while permission prompts control data transfers. The initial release supports only Linux and DGX Spark, and its official results come from an unpublished internal workload that should not be treated as independent validation.

2026-08-26

GitHub Repo

llama.cpp 0.3.0 Adds a Hybrid-Attention KV Cache, MTP, and Tensor Splitting for DeepSeek 4

llama.cpp 0.3.0 adds native support for dots3-note’s hybrid DSA/sliding-window attention and integrates multi-token prediction for GLM-4.5-Air. DeepSeek 4 can now use tensor splitting across multiple GPUs, but the release notes provide no comparisons of throughput, latency, or VRAM usage.

2026-08-26

最新模型

Qwen3.8-Flash-Next Combines Linear Attention with Block-Sparse Retrieval, Previewing the Qwen4 Architecture

Alibaba has released the Qwen3.8-Flash-Next weights. The model uses Gated DeltaNet to compress most historical state, while Qwen Sparse Attention retrieves information from the full context. It natively supports 262K tokens, but the official million-token performance and agentic evaluations have yet to be independently reproduced.

2026-08-26