From Stock Photos to Trillions of Artifacts: Shutterstock's Data Journey with Jefferson Frazer

<p>In this conversation recorded at Fastly Xcelerate, James Governor talks with Jefferson Frazer, Director of AI at Shutterstock, about how a two-decade-old media company reinvented itself for the age of foundation models. Frazer explains how Shutterstock's long history of human-reviewed, meticulously labeled content left it uniquely positioned when hyperscalers began seeking diverse, web-scale datasets for multimodal training. Frazer digs into the technical backbone too: a 100-plus petabyte footprint, cloud-agnostic architecture built on Fastly's private fiber and Wasm-based edge compute, and the discipline of tracking data provenance through standards like C2PA and SynthID. Looking ahead, he makes the case that unified embedding spaces and portable metadata—not any single model provider—will define the next chapter of AI-ready content.</p><p></p><p>This RedMonk video is sponsored by Fastly.</p><p></p><p>Show notes: <a rel="noopener noreferrer nofollow" href="https://redmonk.com/videos/jefferson-frazer/" target="_blank">https://redmonk.com/videos/jefferson-frazer/</a><br /><br />Chapters</p><p>00:00 Introduction to Shutterstock and AI</p><p>04:14 Data Scale and Infrastructure</p><p>07:25 Fastly's Role in Data Management</p><p>10:28 The Future of AI and Business Models</p><p>13:28 Regulation and Provenance in AI</p><p>16:35 Excitement for Unified Embedding Spaces</p>