Tag
Matt Shumer shares a technique for adapting skills to Claude Opus 5: a loop that updates skills, benchmarks them against Opus 4.8 with old skills, and iterates until Opus 5 wins.
Seedance 2.5 is now integrated into Arcads, enabling users to move from research to finished video with unlimited access for a limited time.
Nicolas Zullo demonstrates a workflow using Codex and img2threejs to generate 3D game assets from text prompts, automatically integrating them into a game engine.
Elon Musk shares a workflow using Grok 4.5 for various development tasks, along with Fable 5 and GPT-5.6 Sol, highlighting a real-time research and coding setup.
Introduces the use of the Matrix platform to build an unattended TikTok viral content factory, achieving full automation from product selection, replication to video output, with only two manual steps: product finalization and video review.
The author built an autonomous AI system that runs a live product, generating work, quality-gating it, opening pull requests, and self-improving based on analytics. The main challenge was making the system trustworthy rather than making the model smarter.
A breakdown of how a 4-agent AI workflow was built to automate market intelligence, covering architectural design and implementation details.
This is a tweet recommending Eugene Yan's AI workflow guide, detailing how to efficiently collaborate with AI and achieve compound work gains by organizing context, configuring CLAUDE.md, creating skills, and more.
Demonstrates a workflow that uses Granola for note-taking and Lovable to automatically generate a scoped project proposal, timeline, and pricing from a discovery call, giving an agency a competitive edge by sending the proposal before the client returns to their desk.
A product manager shares how she uses Claude for high-leverage product work beyond busywork, including automating optics tasks and using AI for all three levels of product work: impact, execution, and optics.
A workflow strategy for capturing and reusing AI model outputs, including setting quality bars, strategy roadmaps, knowledge notes, and reasoning records to reduce reliance on frontier models.
Introducing the Karpathy Loop concept, where AI automatically executes experimental loops without real-time user interaction, boosting work efficiency by 10x.
This post details the author's setup and benchmarks for running NVFP4-quantized image and video generation models on a GIGABYTE AI TOP ATOM (DGX Spark) workstation, achieving impressive performance with models like FLUX.2, Qwen-Image, and LTX-2.3 for video with synchronized audio.
A GitHub project with 5.7K+ stars built a Claude Code workflow that automates job applications by scraping job boards, ranking positions, and generating tailored CVs and cover letters with review loops.
This article introduces the concept of Loop Engineering, which involves building repeatable AI workflows by defining goals, execution rules, and verification standards. Using a U.S. stock investor as an example, it demonstrates how to upgrade AI from a chat tool to a continuously collaborative research assistant.
A developer recounts how a cryptic hint from an ex-Anthropic engineer ('check your score') led him to create a harness of files and commands (e.g., CLAUDE.md, init.sh, feature_list.json, etc.) that assess and improve Claude AI agent performance, turning the hint into a practical tool.
A tweet shares nine practical techniques for working effectively with Fable, emphasizing upfront prompting to avoid mistakes, including blindspot passes, asking for options, and letting the AI ask clarifying questions.
Anthropic's telemetry from Claude Code shows users approve 93% of permission prompts, leading to consent fatigue. The article advocates designing human checkpoints at the plan level rather than per-action to improve output quality, citing Anthropic's internal data, Chip Huyen's advice, and Microsoft's crawl-walk-run framework.
The article introduces the advantages of git worktree for parallel multi-task development in the AI era, and how the cc-launch tool developed by the author addresses worktree usage pain points through automatic initialization and unified management, improving development efficiency.
Introducing an AI-assisted development workflow that combines OpenSpec (specification and memory management) with Superpowers (design and execution), using TDD and unified context to solve the two biggest pain points in AI development: lack of memory and lack of discipline.