2026年9月20日 技术热点总结

项目简介：The Photoshop alternative for Mac

技术栈：Swift

核心介绍：Adobe Photoshop costs too much and tools like GIMP don’t feel familiar enough for me to stay in flow. That’s why I built Compositor. The goal was to create a full-featured image editor that is completely free and open source. I use Photoshop for compositing and post-processing, so Compositor is built around that workflow – with the tools needed to create a pixel-perfect final image.

关键特性：Layers and folders, with blend modes and opacity；Layer masks: paint, fill, invert, blur and feather them; link or unlink them to tran…

技术栈：Python

核心介绍：Laya evaluates typed questions (choice, score, noul) over any state (text, email, ticket or JSON document) in a single forward pass — 33 ms for one question, 7.2 ms/question batched, measured on a T4. No text generation, so nothing to parse and nothing to hallucinate. Three checkpoints, and a Router that picks between them per request:

项目数据：⭐ 1,246 Stars，🍴 109 Forks

项目简介：A nano replica of Jev: parallel decisions, dynamic candidates, and an end-to-end training pipeline.

技术栈：Python

核心介绍：Model · Dataset Download the maze video (MP4) · 27 seconds · 1440 × 1120 · 30 fps Play Snake · Explore the 50×50 maze · Recorded sources and replay checks The standalone ChatGPT Sites demo presents Jev, NanoJev, and Untuned Qwen in three light panels. Playback advances by the same environment step across panels; completed runs freeze at their actual final state. Probability bars show the last decision that produced the displayed state. Shared code planni…

项目简介：🍊☁️ Open-source* apps that replace a SaaS product, running in your own Cloudflare account

技术栈：JavaScript、awesome、awesome-list、cloudflare、cloudflare-workers、self-hosted

核心介绍：> Open-source apps that replace a SaaS product, running in your own Cloudflare account. A Worker, a D1 database, an R2 bucket and a wrangler deploy now cover what used to need a VPS and a monthly invoice. This list tracks the software taking advantage of that: a booking page instead of Calendly, a feedback board instead of Canny, an inbox instead of Google Workspace. Every entry is checked rather t…

项目简介：A macOS menu-bar app that dresses your windows in knitted borders.

技术栈：C、desktop-customization、knitting、macos、macos-app、menubar、menubar-app、objective-c、window-management

核心介绍：A little Mac app I made to give my windows sweaters. 🧶 Knitted borders, colours inspired by your favourite apps, and a cosier desktop. You can just send this repo to your coding agent and ask it to install Window Sweaters for you: > Install Window Sweaters on my Mac: https://github.com/saragordic/window-sweaters Or install it yourself with Homebrew: brew trust saragordic/tap brew install –cask sarag…

项目简介：WeChat local data CLI (query/decrypt/export) — fresh start from wx-cli

技术栈：Rust

核心介绍：会话 · 聊天记录 · 搜索 · 联系人 · 群成员 · 群昵称 · 收藏 · 统计 · 导出 通过 skills CLI 一键安装到 Claude Code、Cursor、Codex 等 agent： npx skills add botiverse/wx-cli npx skills add botiverse/wx-cli -g 安装后 agent 会自动读取 SKILL.md，了解如何安装和调用 wx-cli。 源码与发布仓库：botiverse/wx-cli。

项目数据：⭐ 593 Stars，🍴 1,465 Forks

LLM assistants are widely used for daily social advice, yet evaluating their social reasoning in such consultation settings remains challenging since (i) it requires setups where the assistant learns about social situations from subjective user narratives, and (ii) social properties, such as others' intentions, typically lack verifiable ground truth. To address these challenges, we introduce Fuse, a multi-agent simulation framework for studying user-mediated social reasoning. In Fuse, a target agent with a hidden motive interacts with other agents including one representing the user, who th…

Information retrieval is increasingly important as LLM agents tackle complex tasks involving diverse information needs. Because retrieval relies on an index that represents each document through index keys, retrieval quality depends heavily on how effectively these keys expose the knowledge contained in each document. However, effective index representations vary across retrieval environments, making it difficult for any fixed optimization strategy to perform consistently. Yet evolving an index to its retrieval environment remains largely human-driven, requiring humans to diagnose retrieval…

Test-time scaling can improve large language model reasoning by generating and combining multiple candidate responses. In sampling-based methods, the inference budget is often described by the number of generated candidates, N. However, N tells us how many candidates are generated, not how they are executed. The same candidate budget can be produced in one batched generation call or split across several sequential calls with smaller batch sizes. We first study the effect of increasing N on reasoning accuracy using Phi-3-mini and Qwen2.5-1.5B on 500 GSM8K prompts. As expected, increasing N f…

Question answering has advanced rapidly with large language models, but predominantly for high-resource languages, in both text and spoken settings. Spoken question answering (SQA) benchmark for Telugu remains unexplored, and the reliability of automatic evaluation in this setting remains unquantified. We introduce VākQA, a Telugu SQA benchmark of 2,001 factoid question-answer pairs across six domains, with 2.53 hours of speech audio, bilingual transcriptions, and human-verified reference answers. We first validate evaluation methods against human judgements: Gemini-as-a-judge best approxim…

Agent trajectories record what an agent does and what happens next. Yet standard supervised fine-tuning (SFT) applies loss only to agent-authored action tokens, using environment observations as context but not as prediction targets. We ask whether this convention provides the best initialization for subsequent reinforcement learning. We introduce ActObs, which also supervises the observation tokens already present in each trajectory. Although deployed agents never generate observations, learning to predict them encourages the policy to model action consequences without adding data, paramet…

We present Srijika, a system for producing installable OpenType fonts for nine Brahmic scripts: Devanagari, Tamil, Bengali, Telugu, Kannada, Malayalam, Gujarati, Gurmukhi, and Odia. Rather than generating fonts from scratch, Srijika restyles glyph outlines from shaping-complete template fonts. It preserves the template's cmap and GSUB closure and its GPOS data under a documented metric policy, making every output a complete font by construction. This addresses a central challenge of Indic font generation: hundreds to thousands of conjuncts, half forms, and matra variants must remain mutuall…