Claire Vo · 2 min read · Gen AI Tools
How I run autonomous coding agents from my phone with OpenAI Symphony + Linear | Alessio Fanelli (Kernel Labs)
Read the original on lennysnewsletter.com →
Alessio Fanelli demonstrates two AI-driven workflows: a fully autonomous coding pipeline using OpenAI Symphony and Linear, and a Codex-powered bot that autonomously browses eBay to scout underpriced Pokémon cards for his shop. He clarifies a practical agent-management mindset over prompter approaches, explains scaling with a cloud VPS, and shares tips on wiring Symphony and Linear as an agent state machine, tracking costs, and improving autonomous runs.
Key Takeaways
- Agent manager is a better mental model than agent prompter for scalable autonomous work.
- Local Mac Minis don’t scale; a cloud VPS unlocks scalable AI agent deployments.
- Wire OpenAI Symphony and Linear to form an agent state machine across the dev lifecycle.
- Track token costs per task to understand spend and the potential of large token budgets (e.g., 221M tokens).
- Purge CLAUDE.md and use Glimpse to improve agent sensing and reliability during autonomous runs.
Excerpt
Alessio Fanelli , founder of Kernel Labs and co-host of Latent Space podcast, walks us through two very different AI workflows: (1) a fully autonomous coding setup using OpenAI Symphony + Linear, where Linear acts as a state machine and Symphony manages agents through the whole dev lifecycle with zero babysitting; (2) Codex with browser access searching eBay for underpriced Pokémon cards—autonomously browsing, extracting PSA certificate numbers, and flagging deals on $10K–$20K cards for his San Carlos card shop, Merlin Games. Listen or watch on YouTube , Spotify , or Apple Podcasts What you’ll learn: Why “agent manager” is a…