By Alistair Barr
Publication Date: 2026-04-11 13:40:00
I recently interviewed Peter Gostev, AI Capability Lead at Arena.ai. He developed what I call the defecation test, officially known as the BullshitBench. However, something else he mentioned caught my attention.
It’s a striking example of AI agents in the wild. While editing photos in Adobe Lightroom, he had to denoise 50 images, a tedious task that was usually done one at a time. Instead of doing it manually or even knowing how to batch it, Gostev let OpenAI’s Codex AI coding service…
