AI multi-agent pipeline built specifically for academic papers. Upload a PDF or LaTeX source, get an editable PowerPoint with original figures preserved. Three-role architecture (Strategist, Executor, Critic) with 4-pass deep reading, SVG-to-PPTX conversion, and optional icon semantic search. Early-stage but architecturally impressive. MIT license.
An AI Skill that guides academic paper framework figure design through a 13-step collaborative prompt workflow. Features an Atlas reference gallery of common figure types and an 'explore wide, refine locally, lock final' design logic. Primarily built for ChatGPT Web and Codex. Open-source.
The Claude Code Telegram plugin extends your desktop AI assistant to mobile devices via MCP. This guide covers four-step installation, secure pairing with allowlist configuration, group and multi-bot setup, so you can remotely operate your desktop Claude session from your phone with full skill and tool permissions.
MoE models frequently crash mid-training without warning. DeepSeek V4 identified the cause from a differential geometry perspective: manifold tearing. Three stabilization techniques make training predictable.
Jakob Nielsen argues that AI is triggering the first major UI paradigm shift in 60 years: users are becoming supervisors instead of operators, delegating intent rather than issuing commands. This rewrites the definition of usability and introduces three crises: the articulation barrier, the credibility trap, and the cognitive wheelchair.
Local-first agentic HTML editor from nexu-io that turns natural language into complete HTML pages via AI coding agents. Too early to adopt, worth watching.
Curated list of 163 design systems from enterprise, government, and open-source communities. Each entry annotates component library, design tokens, and Storybook availability. Best starting point for design system selection and comparison. Unlicense.
23 production-grade engineering skills for AI coding agents by Addy Osmani (Google Chrome). Covers the full Define-to-Ship lifecycle with Doubt-Driven Development for adversarial self-verification, Anti-rationalization Tables to prevent shortcutting, and 3 specialist personas for parallel review. Pure Markdown, zero dependencies. MIT licensed.
Autonomous ML research framework using cross-model adversarial review — Claude Code executes experiments while GPT reviews and finds weaknesses. Covers the full research lifecycle from literature review to submission-ready manuscript. Requires dual LLM subscription. MIT licensed, Markdown-only, zero framework dependencies.
A single AI Agent is reliable. But your cognitive bandwidth caps at 5 parallel threads. Skill Graphs 2.0 introduces a three-layer architecture — atoms (deterministic), molecules (fixed sequence), compounds (strategic autonomy) — turning 5 compound-level tasks into 500 work units. This article breaks down the framework, practical principles, and known limitations.
Please do not include confidential, unpublished or personal data. Messages are relayed to our team via LINE and kept for 30 days.