2026年8月1日 技术热点总结

技术栈：TypeScript

核心介绍：Send a file between two devices using nothing but a screen and a camera. One page displays the file as an endless stream of animated QR codes; another device points its camera at it and reconstructs the file. No network path between the devices, no app, no pairing, no permissions beyond the camera. The payload travels as light. This is a minimal proof of concept extracted from a larger experiment that reached 128 KB/s phone-to-phone with denser frames,

项目数据：⭐ 1,938 Stars，🍴 222 Forks

项目简介：Multiplayer agent harness for work

技术栈：TypeScript

核心介绍：A multiplayer agent harness for work. In Slack and on the web. Most agents are designed like personal assistants. You can make one work for a whole company, but it quickly gets complex. QM is designed for startups. Employees each get their own isolated workspace and work independently without affecting each other, and they can also collaborate with the agent in channels, group messages, and projects. Each person and each room has its own scoped memory, files, keychain view, permissions,

关键特性：**Personal and shared…

项目简介：A realtime voice runtime that keeps Agents talking, working, and present. Real-time Voice Runtime for AI Agents

技术栈：JavaScript、agent、agentic-ai、voice-agent、voice-ai、voice-chat

核心介绍：中文 | English 真正的交流，不该在说完一句话后，就陷入漫长的等待。也不该因为 Agent 正在查资料、调用工具或处理任务，整场对话就此暂停。 交流应该是连续的，Agent 也应该始终在场。 所以，我们做了 qwen-audio-agent——让 Agent 持续交流、持续工作、持续在场的实时语音运行时。无论是聊天、思考，还是处理任务，Agent 都始终在这场对话里。它会倾听，会回应，也会在任务完成时自然地告诉你： 🚀 正式版发布，推出内置 Gateway 的 macOS 桌面版。 🌍 项目正式开源，后台 Agent 统一接入 ACP 架构。 对话不会因为后台任务而停下；任务完成后，结果会自然回到当前对话： 能直接回答的问题会立即回答；需要工具或持续处理时，任务会交给后台 Agent。 整个过程中，用户面对的始终是同一个助理。 查看详细架构 更完整的设计与模块说明见…

项目简介：FDE（前沿部署工程师）从零入门指南（基于范冰《增长黑客》原书框架）

核心介绍：范冰 著 · 免费公开全文，欢迎在线阅读与分享 2025 年夏天，我的朋友圈被同一个数字刷屏：95%。 麻省理工学院的一份报告说，过去三年，全球企业在生成式人工智能上烧了三四百亿美元，其中 95% 的项目没能产生任何能写进财务报表的价值。几乎同一时间，另一条新闻在往相反的方向狂奔：硅谷的招聘网站上，一个叫「前线部署工程师」（Forward Deployed Engineer，简称 FDE）的岗位，发布量九个月涨了八倍。OpenAI 在招，Anthropic 在招，YC 孵化器里一百多家创业公司都在招。 一边是企业人工智能项目 95% 的阵亡率，一边是一个岗位 800% 的抢手度。把这两条新闻摆在一起看，答案不难猜：模型已经不稀缺了，能把模型塞进客户真实业务里的人，才稀缺。 这本书就是那个研究过程的完整沉淀。它聊清楚三件事： 书里所有数据和案例，都在附录 C 里标明了出处。整理的过程本身就是学习，我尽量让每一条引用都经得起核查；如有疏漏，欢迎通过 Issue 指正。

项目数据：⭐ 545 Stars，🍴 66 Forks

项目简介：Wilonity Loader – cheat lib w/ spoofer, driver bypass, undetected injector for 20+ games (RUST, CS2, Valorant, Tarkov, Warzone, R6S, ArcRiders, Apex, Roblox, Meccha Chameleon). Kernel spoof, HWID cleaner, AC bypass (EAC/BE/VG). ESP, aimbot, WH, wallhack, triggerbot, radar, no recoil, silent aim, chams, skin changer, unlock all, no spread.

技术栈：arc-raiders、counter-strike-2、game、hack、meccha-chameleon-tools、minecraft

核心介绍：> 📋 *Full list and detailed module descriptions available on our website*

关键特性：Unified Launcher – one app for all your games；**Smart Module System*…

Existing token compression methods for omnimodal large language models typically rely on one modality to determine what to retain in the other. We show that this assumption often breaks down: for the same query, audio and video relevance often peaks at different moments. This cross-modal salience mismatch makes unidirectional guidance prone to discarding answer-critical cues under aggressive compression. We propose OmniScope, a training-free token compression framework that uses the query as a shared semantic anchor while estimating relevance separately for audio and video. OmniScope alloca…

On-policy self-distillation (OPSD) is a promising approach to improve reasoning language models, but it remains brittle in practice: making it work reliably often requires substantial engineering effort. We identify a structural source of this difficulty: vanilla OPSD is precisely the β=1 member of a broader policy-optimization family, where β weights the KL penalty anchoring the student to a reference policy. This equivalence turns β from an implicit value fixed at one into a controllable regularization parameter, yielding a more general formulation that trades off proximity to a reference…

Sparse mixture-of-experts (MoE) language models route each token to multiple experts, suggesting a geometric account of their benefit: co-selected experts should contribute distinct representation directions. Existing evidence often conflates route coherence, candidate quality, and candidate-by-context interaction. We distinguish these quantities using an Expert Subspace Separation Index (ESSI), matched-route residuals, and a prefix-controlled 2times2 factorial; frozen-route interventions and a controlled Top-k study assess functional value. Three paired contrasts organize the findings. Fir…

This work presents Fairness Pruning, a lightweight structural intervention method designed for the management and future mitigation of demographic bias in large language models (LLMs). As a foundational empirical validation of this method, this work focuses on causal bias localization. Using minimally contrastive prompt pairs and inference-time activation capture, the method identifies neurons that react differentially when processing demographic attributes in GLU architectures, evaluating the signal at the down_proj input. Empirical evaluation was conducted on models of up to 3 billion par…

Memory is central to long-horizon LLM agents, yet existing memory systems primarily preserve interaction content rather than modeling which agents can be trusted and under what conditions. This limitation is particularly important in multi-agent systems, where a central model may be unable to directly verify plausible or correlated peer responses. We introduce Σ-Mem, an online reliability memory that records historical competence evidence for individual peers and peer relationship evidence across the peer set. Both forms of evidence are maintained as real symmetric states and updated from p…

Multimodal large language models increasingly use sketches, annotations, tools, and intermediate images during reasoning, but it remains unclear whether they truly rely on these visual states. Existing benchmarks are limited both by task collections with narrow coverage or partially text-solvable samples and by evaluations that emphasize final answers without diagnosing how intermediate visual states are generated, rendered, and used. We introduce See2Think, a unified evaluation framework comprising See2ThinkBench and Visual Action-of-Thought (VAoT). See2ThinkBench contains 1,200 open-ended…