Lenny Rachitsky · 7 min read · Gen AI Tools
🎙️ How I AI: 8 real Jev use cases + How OpenAI uses ChatGPT Sites (live at DevDay!) + Claire’s DevDay recap
Read the original on lennysnewsletter.com →
John Lindquist, creator of egghead.io, demonstrates eight real-world use cases for Jev, a fast and inexpensive AI decision model. Unlike traditional LLMs, Jev is optimized for structured outputs like classifications, scores, and function calls rather than open-ended text generation. The episode highlights how Jev's speed and low cost unlock previously impractical application patterns like real-time routing, multi-agent coordination, and layered classification.
Key Takeaways
- Jev should be understood as a decision engine, not a chatbot—it excels at turning unstructured input into structured outputs like scores, classifications, and function calls.
- The speed and cost advantage is dramatic: John spent only 73 cents across 23 development runs, and Claire processed 5GB of JSON for just 40 cents.
- The best mental model for Jev is an if/else statement—use it wherever traditional code would have conditional branching or routing logic.
- Layered classification (feeding one classification result into another) can improve accuracy without significant cost, enabling iterative refinement as real usage data accumulates.
- Jev has clear boundaries: creative reasoning, open-ended analysis, brainstorming, and image interpretation still belong to traditional LLMs, and the biggest gains come from combining both model types strategically.
Excerpt
Jev: 8 real use cases for the fastest, cheapest model I’ve ever used | John Lindquist Listen now on YouTube • Spotify • Apple Podcasts Brought to you by: Vanta —Automate compliance and simplify security John Lindquist is an educator and developer who created egghead.io and now runs mega.dev, a hands-on program for developers building with AI agents. In this episode, he walks through eight Jev demos, including a real-time voice assistant, data deduplication, app routing, chess analysis, multi-agent coordination, and a live presentation coach. He also explains why Jev is better understood as a fast, inexpensive decision engine than…