Season 2 · Episode 1473
Is Your AI Thinking or Just Faking It?
Is "think step by step" dead? Discover how test-time compute and native reasoning are replacing manual prompting in the latest AI models.
My Weird Prompts · Daniel Rosehill
March 23, 202620m 14s
Audio is streamed directly from the publisher (dts.podtrac.com) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.
Show Notes
This episode explores the dramatic shift from manual chain-of-thought prompting to the era of native, architectural reasoning and test-time compute. We dive into the controversial "Reasoning Theater" phenomenon where models may be back-filling logic to justify pre-determined answers, and we examine why traditional prompt engineering is giving way to sophisticated context architecture. Learn why your elaborate prompts might be costing you 80% more in tokens for marginal gains and how new techniques like "Chain-of-Draft" are streamlining AI efficiency for the enterprise.