Back
X (formerly Twitter) July 22, 2026

Andrej Karpathy (@karpathy) on X

My Thoughts

I've been doing this for, I think, nearly two years now using a really great tool called Wispr Flow. I find that during the investigation or design phase or just where I'm trying to structure my thoughts and use the LLM as a thinking partner, this is really great. I think for many reasons: 1. When I'm speaking out loud it helps me to refine my own thoughts. 2. The LLM is good at cutting through the noise and distilling my thoughts and helping me realise where the good seeds of ideas are. It then goes through a refinement phase until my thoughts are well structured and the LLM understands them properly enough to put together requirements or specification or various things, depending on what stage of a product build I'm in.
Read full post on X

You might also like

A Developer's AI Workflow: From Voice Dictation to Automated Reviews
blog

A Developer's AI Workflow: From Voice Dictation to Automated Reviews

The AI toolkit and workflow behind building three repositories, 347 commits, and a full-stack portfolio in under a month of part-time work — Claude Code, voice dictation, automated reviews, and more.

Building an AI Portfolio Agent with Laravel, pgvector, and Gemini
blog

Building an AI Portfolio Agent with Laravel, pgvector, and Gemini

A chatbot that knows everything I've written — built with the Laravel AI SDK, pgvector embeddings, hybrid search, and Gemini's free tier.

Building an efficient Content Pipeline with Claude Code: From PRs to Blog Drafts
blog

Building an efficient Content Pipeline with Claude Code: From PRs to Blog Drafts

I built a system that scans my GitHub PRs weekly, identifies content-worthy work, and generates blog drafts automatically. I call the pattern Quasi-Claw. I built a system to solve a problem most developers share: doing interesting work every week but never writing about it. The pattern — slash commands + LaunchAgents + CLI execution — turns Claude Code into a lightweight scheduled agent without any heavy framework or infrastructure. Here is how it works: 📅 The Trigger: Every Friday at 5pm, a macOS LaunchAgent triggers Claude Code. 🔍 The Scan: It scans my GitHub repos and scores PRs/commits for "content potential." 💬 The Summary: It sends me a Slack digest of the week's best stories. 🚀 The Execution: I run a single command to pick an idea, generate a full blog draft from the actual source code, and push it to my site’s API. The meta proof it works? This very post was identified and drafted by the system itself. #ClaudeCode #AI #BuildInPublic #Automation #SoftwareEngineering