This Week on How I AI: Which AI Model Is the Best Designer? And ChatGPT Agent Mode, the “Little Helper” That Transformed Recruiting

Summary

In this episode, Michal Peled demonstrates how ChatGPT's agent mode can act as a "little helper" for real‑world tasks—automating LinkedIn recruiting, creating interactive AI personas, and building a parking‑avoidance app—by first interviewing humans to capture exact workflows and then refining prompts iteratively. Claire also runs a side‑by‑side test of Gemini 3, Claude Opus 4.5, and GPT‑5.1 Codex on the same design prompt, finding Anthropic's Opus 4.5 to be the clear design winner, while Codex struggles with visual polish and Gemini 3 is merely serviceable. The episode offers practical tips on structuring agent workflows, using prompt‑fixing techniques, and choosing the right model for design work.