2026年11月2日 技术热点总结

项目简介：244 pixel art effects, every one a small piece of code. Plus a Claude Code skill to make your own.

技术栈：JavaScript、agent-skills、animation、canvas、claude-code-plugin、gamedev、pixel-art、sprite-sheet、vfx

核心介绍：explosions, magic, weather, UI, pickups, critters and more. Play them live, recolor and mirror them, drop them into a browser game, or export sprite sheets for any engine. Plus a Claude Code skill that makes new ones the way a picky pixel artist would: draft, render, look, refine. Every effect that involves a character is acted out: he winds up, lets go, freezes on con…

项目简介：VERY experimental Nintendo Switch 2 emulator

技术栈：C++

核心介绍：Eglantine is a free and opensource (FOSS) Switch 2 emulator, derived from Eden, Yuzu and Sudachi – started by developer Camille LaVey. It’s written in C++ with portability in mind, with builds for Windows, Linux, macOS, Android, FreeBSD and more. Compatibility | Development | Building | Download | Support | License The emulator is capable of running most commercial games at full speed, provided you meet the necessary hardware requirements.

项目数据：⭐ 1,982 Stars，🍴 238 Forks

项目简介：See how your app actually fits together. An interactive, plain-English map of your codebase, drawn by Claude or Codex.

技术栈：JavaScript、ai-coding、architecture-diagram、claude、claude-code、claude-code-plugin、codebase-visualization、codex、codex-plugin

核心介绍：Claude or Codex reads your codebase and draws an interactive map of it. Every part, file and web request gets a plain-English name, and every connection between them is shown. ▶ Watch the walkthrough: installing Floorplan, then a tour of Brewline, a demo coffee-subscription app built across four repos. Try the live demo → Exp…

项目简介：Procedurally generated Delft-blue floral page margins for LuaLaTeX, drawn in Lua

技术栈：Lua、generative-art、latex、latex-package、lua、lualatex、postscript、procedural-generation、svg

核心介绍：The Reddit post asked how to make flowers like these in LaTeX. I don’t know how the original was made. This project is one possible way, and the way I did it uses hand-authored primitives with randomized parameters, painted in Lua and placed by a procedural layout whose “fill the unused space” part borrows a 1998 idea from the ornament literature. It gets a similar look, but I explicitly tried to…

Quarterly earnings forecasting requires an agent to connect financial evidence to the income-statement line items that produce a company's reported earnings. We introduce EPSBench, a benchmark for studying this ability across 149 firm-quarters from 86 US companies. Agents read a fixed collection of filings, earnings-call transcripts, and news, then estimate company-specific income-statement drivers with cited evidence. A shared accounting calculator converts these estimates into GAAP diluted earnings per share (EPS), providing a common basis for comparing forecast assumptions. We evaluate s…

Self-distillation offers a way to improve reasoning models when a suitable external teacher is unavailable. However, common approaches rely on privileged information to construct a stronger self-teacher, limiting their use when only unlabeled prompts are available. In this paper, we introduce Weakening-Induced On-Policy Self-Distillation (WISD), which instead constructs a weaker counterpart of the original model and uses the contrast between the two as a training signal without requiring privileged information. We explore two ways to construct the weaker model: low-bit quantization and magn…

Autoregressive (AR) image generators are powerful in principle but inflexible in practice: training with a fixed raster-scan order hardcodes a single factorization of the image distribution into the model, leaving it unable to inpaint, outpaint, or edit at inference time. Existing solutions, including the modern discrete-diffusion family, train from scratch on a mixture of token orderings, paying roughly 3x the compute of standard raster-scan training; discrete diffusion further sacrifices compatibility with the AR inference ecosystem (KV caching, vLLM) by construction. We argue that this c…

Language agents increasingly act through external tools and interactive environments, making the interface between the model and its executable capabilities part of the decision process. Model Context Protocol (MCP) and command-line interfaces (CLI) can expose the same semantic operations through markedly different model-facing representations. We study whether these representational differences change agent behavior when executable capabilities are matched. We introduce Matched Interface Comparison and Analysis (MICA), which renders MCP and CLI from a shared semantic action contract and ro…

Developers download quantized large language models (LLMs) from public registries to run them locally. None of the registries we examined documented functional testing of their files. We tested all 305 eligible code-capable GGUF tags (the llama.cpp format) across 15 model lines in the official Ollama library, plus 22 popular Hugging Face community files. A 15-task code screen identified suspects, which we retested on all 164 HumanEval+ tasks. Seven official tags (2.3%) solved at most six of them. All were low-bit quantizations of models with at most four billion parameters. We rebuilt five …

Foundation models are typically trained in stages: pretraining, midtraining, supervised fine-tuning (SFT), and reinforcement learning (RL), with changes in both data and learning algorithms. We show that this algorithmic separation is not fundamental. By deriving suitable RL objectives as maximum-likelihood fitting to model-generated target distributions, we place all of these stages under a single likelihood-based training algorithm, unified training. This perspective turns the foundation-model training lifecycle into a problem of choosing which data to train on, in what proportions, and w…