Greg Isenberg spent a week testing Gemini 4 Argon, Google's latest frontier AI model, and shared a video walkthrough of three places it excels: marketing automation, physical AI (3D design and manufacturing), and complex engineering. He demonstrated Argon rebuilding his holding company's website, writing posts in his voice, designing a Shopify product customizer that exports print-ready STL files, and building a free quoting app for commercial cleaning companies.
Key findings: Argon produces strong creative work with reference images and clear direction. It handles engineering correctly on the first but needs multiple rounds for design refinement. Isenberg flagged Google's interface and harness as areas where other tools still lead, but expects Google to close the gap. He shared six techniques—including /plan, /subagents, and /goal commands—that significantly boost output quality.
This is an AI-written summary. The original reporting belongs to Greg Isenberg. How we write these · Spot a mistake?