Summary
Greg Isenberg shares what he found after a week of heavy use with Google’s Gemini 4 Argon, arguing the model is most valuable when pointed at three specific business problems rather than used as a generic chatbot. The episode is sponsored by Google, which gave early access.
The first use case is AI-driven marketing. Isenberg shows Argon rebuilding his holding company’s website, which he says originally cost around $200,000, as an interactive experience with sound effects and small games, and he demonstrates training the model on his own writing so newsletters match his voice. The second is physical AI manufacturing, where Argon helps design customizable products as STL files for Bambu Lab 3D printers, including an idea to embed product customizers into Shopify stores. The third is aiming the model at legacy SaaS products to build better, more niche versions, which he says it handles with tested rather than loosely vibe-coded output.
He is also candid about Google’s Antigravity coding harness, saying it still needs to catch up, and notes Argon can be used in Cursor, OpenCode, and other tools. The video closes with six tips for getting more out of the model.
📺 Source: Greg Isenberg · Published October 15, 2026
🏷️ Format: Workflow Case Study
