2026年11月6日 技术热点总结

项目简介：Google_AI_Pro_2Year_ZeroCost_Auto

技术栈：Python

核心介绍：======================================================================== Google AI Pro (5 TB) 零成本两年 (US$0.00 -> 24个月) 全自动控制台 ======================================================================== 【一、运行环境要求】 1. Windows 10 / 11 系统 2. 已安装 Google Chrome 浏览器（默认路径即可） 3. 已安装 Python 3.10 或更高版本（安装时勾选 Add Python to PATH） 4. 网络环境：本机开启美区（US）代理，或在控制台左侧填写代理模板 【二、快速启动步骤】 1. 首次使用请双击运行：【1_首次运行_安装依赖并启动.bat】 2. 之后日常使用直接双击：【2_启动控制台.bat】 3. 浏览器会自动打开控制台页面：http://127.0.0.1:8765 【三、自定义填写支付卡 & 跑号说明】 1. 填写支付卡（必填）：

项目数据：⭐ 2,311 Stars，🍴 945 Forks

项目简介：VERY experimental Nintendo Switch 2 emulator

技术栈：C++

核心介绍：Lichii is an experimental, open-source Nintendo Switch and Switch 2 emulator written in C++. It builds on Eglantine, Eden, yuzu and Sudachi, preserving their contributions while developing this fork’s game compatibility and renderer fixes. This is an independent fork, not an official release of those upstream projects. Lichii is not affiliated with Nintendo.

项目数据：⭐ 2,220 Stars，🍴 344 Forks

项目简介：Entire MacOS running in your terminal.

技术栈：Swift、cli、gui、macos、shell、swift、terminal、tui

核心介绍：A separate macOS desktop for computer-use agents, viewed and controlled from a terminal. terminal-mac is an early release candidate for Apple silicon. Apple Virtualization runs a real macOS guest. A bundled helper inside that guest captures its screen and posts input events. A terminal viewer streams that screen using Kitty or iTerm2 graphics. Agent commands use the same helper without moving the physical Mac’s pointer or changing its clipboard.

项目数据：⭐ 1,947 Stars，🍴 119 F…

项目简介：知了 Cicada · 任何会议和直播的实时字幕、同声传译、AI 会议纪要，Mac、Windows、Linux、Android 免费开源。Live subtitles, interpreting and meeting notes for any call or stream, on Mac, Windows, Linux and Android.

技术栈：Swift、android、linux、live-captions、macos、meeting-notes、menubar-app、simultaneous-interpretation、speech-recognition

核心介绍：知了 · Cicada 任何会议、任何直播，没有语言障碍。 听得懂，说得出，记得住。Mac、Windows、Linux 和 Android 上的实时纪要、双语文字记录、字幕与同声传译。 简体中文 · English · 日本語 · 한국어 无论是用 Zoom、飞书、腾讯会议开会，还是听 X Spaces 和 YouTube 直播，知了都能实时给出双语字幕、边听边写结构化纪要。遇到需要发言的时候，按住说话键（Mac 上是右 ⌥，Windows 上是右 Ctrl）开口，对方就能在会里直接听到翻译好的语音。知了直接读取系统声音，不用往会里塞机器人…

项目简介：Autonomous pentest legion — AI commander + worker swarm + layered fact-intent graph. 3rd Tencent Intelligent Attack-Defense online champion (634 teams · 8h unattended · 1B tokens beat 4.5B). The Nüwa Test: give AIs Create/Language/History, let a civilization grow.

技术栈：Python、ai-agents、autonomous-agents、claude-code、ctf、llm、multi-agent、pentest

核心介绍：> Champion of the 3rd Tencent Cloud Intelligent Attack-Defense Challenge online qualifier > (634 teams · 80.18/100 · 8 hours fully unattended) · First Prize, GeekCon 2026 AI Challenge > Same 85 flags as the runner-up, 25 …

Sparse attention cuts the cost of long-context inference by making the attention matrix sparse: each query attends only a subset of the sequence, chosen by a fixed pattern or, in recent production LLMs, learned per query. In DeepSeek Sparse Attention (DSA), the attention of DeepSeek-V3.2 and the GLM-5 family, every layer is preceded by a lightning indexer that scores the whole cache and keeps the k=2048 highest-scoring tokens for the layer to attend. At long context this indexer scan, not the sparse attention it feeds, is where the decode time goes. But adjacent layers select nearly the sam…

A frontier AI model is trained in an AI data center, on tens of thousands of accelerators joined by a dedicated network, and every training step ends with all of them exchanging data at once. Between racks, that network is optical, and each of its four layers, from lasers to switches, runs a fixed or after-the-fact policy. Lasers sharing an enclosure, cooling, and power can fail together. At a fixed spare ratio, correlated failure provably caps a pool's availability however large it grows. In this paper, we present a comprehensive predictive control plane framework, called Halley, that read…

Streaming digital humans produce detailed video while continuing speech as new text arrives. High-resolution denoising is costly, and text updates can cause repeated or interrupted speech. We present TaoMate-HD, an autoregressive model for high-resolution audio-video streaming. For each short segment, or chunk, two denoising steps establish coarse video structure at low resolution. A final step refines the video to high resolution, with audio generated throughout the same trajectory. Each stage passes its audiovisual predictions to the same stage of the next chunk as history. Across text up…

Generative AI is billed per token, yet tokens are not equally expensive. Each token enters the KV-cache and raises the cost of those that follow, so energy grows non-linearly along a sequence whose tokens look identical. No analytical model states what a token costs, and the most accurate SOTA model is a fit, not a law: it absorbs architecture into fitted coefficients, requiring per-LLM fitting. We derive closed-form expressions of the computational and memory-access complexity of LLM inference in terms of architectural hyperparameters, covering MHA and GQA, and compare a family of analytic…

Historical archives are invaluable for multidisciplinary research but remain inaccessible globally due to severe language barriers. While manual translation is prohibitively expensive and time-consuming, simply applying existing machine translation models is inadequate due to unique linguistic nuances and the temporal bias inherent in scarce human-labeled corpora. To address these challenges, we propose HERIT, a systematic domain-adaptation pipeline that leverages retrieval-augmented generation (RAG) to generate high-quality pseudo-labeled data from abundant Hanja-Korean corpora. This appro…

Classification accuracy and true-positive and false-positive rates alone do not determine whether restarting an AI agent improves success. Using an exact alarm-to-action identity for an independent clean restart with zero monitoring overhead, we show analytically that a monitor with fixed 90% accuracy can swing from a five-percentage-point loss to a 35-point gain in success rate, depending solely on the replacement attempt's residual-budget success probability. We introduce two techniques for evaluating monitors and progress signals in decision terms. First, a decision-sufficiency criterion…