AI Agents: $85,000 in Tokens for 293 Pull Requests per Week

One developer, 293 pull requests in one week, and zero defects in production. This is how Alexander Lebedev documents the agentic workflow that transformed his work at Lovable, spending $85,000 on AI tokens over 6 months.

We analyze what really changed: how code review was redistributed, why pull requests remain essential in the age of agents, and the real risk when the risk classifier fails. We also discuss the gap this creates between funded startups and small teams without a budget for tokens.

If you’re interested in the future of AI-powered software development, subscribe and hit like so you don’t miss the next episode.

🤖 AI-generated content: the script, voices, and images for this episode were produced using artificial intelligence tools.

📷 Images:

• “San Antonio data center” — Robert Scoble (CC BY 2.0) — https://www.flickr.com/photos/35034363287@N01/2340202215
• “Free computer server room image” — Unknown (CC0) — https://www.rawpixel.com/image/5912401/image-public-domain-black-technology
• “Free computer server room image” — Unknown (CC0) — https://www.rawpixel.com/image/5906639/photo-image-light-desktop-wallpapers-public-domain
• “Quantum Computing for Google Goggles” — jurvetson (CC BY 2.0) — https://www.flickr.com/photos/44124348109@N01/4171280876
• “Startup Business People Working on Laptop Copy Space” — Rawpixel Ltd (CC BY 2.0) — https://www.flickr.com/photos/147875007@N03/26666483447

#ArtificialIntelligence #SoftwareDevelopment #AIAgents #AIProgramming #AIEngineering #CodeWithAI #FutureOfWork