Featured
Live mentions
Real-time mentions from X plus the latest from Reddit and Hacker News. Updated every 10 minutes
Live mentions
Real-time mentions from X plus the latest from Reddit and Hacker News. Updated every 10 minutes
2k77
A Programming Paradigm for Spatiotemporal Composability
分享12篇跟DeepSeek Harness相关的高质量学习资料，包括官方文档、Harness教程、插件文档、相关论文及解读等： 1、DeepSeek Harness 官方仓库，关于源码、README、示例、架构、版本变化和社区入口的权威来源 2、DeepSeek Harness官方Web UI教程，介绍安装、启动、模型配置、工作区设置和基本使用流程 3、DeepSeek…
143.4k14.4k
分享12篇跟DeepSeek Harness相关的高质量学习资料，包括官方文档、Harness教程、插件文档、相关论文及解读等： 1、DeepSeek Harness 官方仓库，关于源码、README、示例、架构、版本变化和社区入口的权威来源 2、DeepSeek Harness官方Web UI教程，介绍安装、启动、模型配置、工作区设置和基本使用流程 3、DeepSeek…
157.8k45.4k
Java 面试 & 后端通用面试指南，覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
分享12篇跟DeepSeek Harness相关的高质量学习资料，包括官方文档、Harness教程、插件文档、相关论文及解读等： 1、DeepSeek Harness 官方仓库，关于源码、README、示例、架构、版本变化和社区入口的权威来源 2、DeepSeek Harness官方Web UI教程，介绍安装、启动、模型配置、工作区设置和基本使用流程 3、DeepSeek…
Telegram bridge for Grok/xAI-powered bots. 1:1 + group chat, session memory, allowlist. npm @talocode/botlane | PyPI talocode-botlane | MCP server + HTTP API
We built BotLane because agents need interfaces people already use. Telegram is already where builders, teams and communities spend time. Your agent should be able to meet you there without forcing…
9.4k780
Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Codex, Gemini and more), by model, project, and task. npx codeburn
Your AI coding bill shows the total. It doesn’t show where the spend went. CodeBurn is a free, local-first tool for tracking token usage and cost across 40 AI coding tools and agents, including…
37460
What if your phone had no apps
what if your phone had no apps. appless is an open source experiment where the llm builds the screen you ask for on the spot. type what you need, it renders the ui. good for personal tools and one…
A curated list of plugins, skills, MCP servers, patch/profile layers, orchestrators & UIs for DeepSeek Harness (DSH). Visualization · PPT · Coding · Agents · Loops (auto-research) and more. #dsh
Discover curated plugins & tools for DeepSeek Harness. 🌟 See link below 👇
Where Regulations Meet AI Reasoning
Devlog, Clause & Effect #18: Cleared the last 8 blockers in my dependency gate. Then found the reports I'd been clearing them from are truncated. GuardDog's rules carry a max_hits cap. transformers…
I built an open-source tool to generate refactoring datasets — looking for feedback
A configuration-driven interactive personal space template for creators and independent builders.
这是我的个人网站。 它不只是在线简历，我把经历、项目、内容和兴趣，组织成了一座可以探索的数字空间。 做完以后，我又把私人信息和素材全部拆掉，整理成了一个开源模板。你可以替换一份配置，改成自己的个人空间。 网站： GitHub： 欢迎体验，也欢迎拿去改。反馈和 Star 都很珍贵。
2k127
Community recipes for serving LLMs on RTX 3090/4090/5090 CUDA gpus. Multi-engine (vLLM, llama.cpp, ik_llama) and model-agnostic. Currently shipping Qwen3.6-27B Qwen3.6 35B Gemma 4 26B Gemma 4 31B configs for 1× and 2× cards.
In our quality testing low and xhigh gave similar results (apart from time spent), medium was just slightly behind.
3.1k258
Cursor plugin specification and official plugins
Cursor just open-sourced the official repository behind their AI agent plugins. It's called `cursor/plugins`. Most people use Cursor like a basic coding assistant: "Write this function for me." That…
Unofficial Binance Android version notes: product, copy, APIs. No APKs.
新搞了一个项目，以后币安APP有更新，就做对应的新增内容分析
64.1k13.3k
Open-source AI job search: scan job portals, evaluate listings with a structured A-F rubric into a 1.0-5.0 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)
a friend of mine has been using this to cut through job-search noise, thought I'd share not an "AI got him the job" tool. an AI that tells you which jobs are worth your time it's called career-ops,…
FREE-MAD Orchestrator: Consensus-Free Multi-Agent Debate with CLI adapters, transcripts, dashboard, and tests
Happy to share some good news: FREE-MAD, the consensus-free multi-agent debate framework for LLMs, has been accepted at ACL 2026 Findings. Full credit goes to the authors, Yu Cui, and the team at…
17810
统一 LLM 服务接入层 — 一套 API 接入 172+ 家 AI 服务商
aimux 0.3.0 更新啦，现在你可以记录、回放和追踪每一次调用，探测 Prompt Cache 的真实命中情况；通过 RouterModel 和 MoA 实现多模型路由、容错与聚合；同时我们完成了大规模稳定性加固，让多语言、多 Provider 场景运行得更加可靠。
Evals and benchmarks for AI models
4/5 They also published the methodology. Models got problem descriptions instead of implementation hints, each one ran three times with default settings, and success was measured by whether the code…
Instruction to setup hidden deepseek's roleplay reasoning.
关于这两天大火的 DS 鲸鱼娘神秘代码，它原理和“你是一个猫娘”没有区别，如果翻译成中文发给它，一样能激活。 写成英语大写+下划线纯粹是想搞大新闻。 其格式来源是 ds 的大佬发的项目，在第一条信息后面附上方括号与人格设定的提示词，连思考链都会遵循人设 这个功能三个月之前已经集成在 HanaAgent 里面，打开过后助手会以你Agents.md里面的信息为设定，第一人称思考，增加推理链的代入感。
AI-HPP-Standard: an inspection-ready architecture for accountable AI systems. Vendor-neutral. Audit-ready. High-risk gated. Developed via structured multi-model orchestration with human oversight. Designed to support emerging international AI governance.
AI-HPP v4.3.0 is still a USABLE_DRAFT. No claims of certification. No claims of universal safety. No claims of independent validation or adoption. So we are asking researchers and engineers for…
wasm powered frappe playground running entirely in browser
Frappe Playground got a fresh UI + Boot Flags 👀 You can now customize the Playground through query params. And #Frappe Vault’s live demo is now powered by Playground itself. Same Playground,…
8.5k1k
📬 Most “CRM setups” die because nobody types the contacts in. Charge a solo consultant to stand up an agentic CRM that fills itself from mailbox evidence, queues the follow-ups, and then pay…
1.8k28
Opinionated Oxlint rules for rejecting low-evidence TypeScript and JavaScript patterns
Fighting AI slop with tools that don’t have opinions. I ignored this problem for way too long, giving agents breathing room to let their slop grow in secret. If you’re still reading the code from…
Custom Parts for a Voron 2.4 Stealthchanger Build
GX12-4コネクタで簡単にツールヘッドを脱着できるVoron Stealthchanger CMYKのアンビリカルブラケットに合わせたHEXA DISTRO FUSIONマウンタの試作一号を作りました。この勢いでアンビリカルケーブルも作ろうとしたら、買ってあったケーブルの径が太すぎてコネクタに入らなかった…。1回休み。
The MIT-licensed infinite-canvas whiteboard SDK. An open-source tldraw alternative — no license fee, free for commercial use. React, React Native, and plain JS.
别急着买付费白板 SDK，先看这个。 它的核心不是画布渲染，而是数据层：文档是扁平不可变记录，每次操作产生一个 JSON diff。持久化、撤销、多人同步共用这条路径，远程 diff 不污染本地 undo 栈，协作回退不会误伤别人。MIT 协议、core 零依赖纯 ESM，React、React Native、原生 JS 都有包，RN 端把引擎装进 WebView…
Solana IMprovement Documents (SIMDs) describe proposed and accepted changes to the Solana protocol.
Solana is moving beyond flat transaction fees with SIMD-0553, introducing resource-based pricing for blockspace. The proposal could lift daily $SOL burn from ~648 to 7,500–9,000, while making…