Issue 349: Professional skeptics

Nowadays, a little skepticism goes a long way. In fact, it may be one of the best skills a dev can have in the age of AI. BrowserStack’s David Burns joins us on the pod to discuss just that, plus the importance of applying test-driven development to your agentic workflows. But a real skeptic is also a creative one—someone who can be skeptical of what people tell them they can’t do. For instance, we chat with AMD's Anush Elangovan about how their open-source ROCm toolchain is proving that agents can write low-level GPU code. And what’s more fun than sorting fact from fiction with all your skeptic friends? We’ve got a conversation with Slack’s Rob Seaman about bringing multiplayer to your team’s AI development.

This week’s skepticism is carrying over into our favorite reads from around the web. We’ve got one dev’s take on why AI has not solved coding, Nvidia’s latest attempt at containing wily AI models, and a philosophical look at the dangers of hyperbolic AI metaphors. Plus, we’ve got a technical deep dive on streaming from your PS5, how your pirated movies can help film restorationists, and whether being vague can save you from a lawsuit. Nothing to be skeptical about—we really do have all of those stories and more, down in the links below.