Why Your AI Model Suddenly Got Worse: Inside AI "Nerfing" | Utilizing AI Episode 53
Educational, News, Talk Show, 18-Nov-2026
Have you ever woken up to find that your favorite AI model suddenly feels lazier, less accurate, or just plain worse than it did the week before?
In this episode of Utilizing AI, host Stephen Foskett is joined by Futurum Group analysts Brad Shimmin and Olivier Blanchard to unpack the frustrating reality of AI "nerfing"—when frontier cloud models mysteriously lose capabilities or shift behavior without warning.
The panel debates whether silent model degradation stems from SaaS "enshittification" and pre-IPO cost-cutting or fragile, constant software tweaks, and why unpredictable outputs create serious accountability risks for professionals.
They also explore how running distilled, purpose-built models locally on next-generation AI PCs and edge hardware can "buff" your workflows and restore consistency.
This and more on Utilizing AI, part of Futurum Media and the Futurum Podcast Network.