In this episode, I speak with Luca Belli about his forthcoming Manning book, “Hidden Influences,” which explains how algorithmic recommendation systems work from socio-technical and ethical perspectives rather than teaching readers to build them. Belli describes how recommenders and users shape each other, why debates about societal harms (including misinformation) often lack nuance, and why “working properly” is hard to define because value is personal and difficult to measure, leading platforms to optimise for engagement as a proxy with harmful side effects like clickbait.
Drawing on his background in math, data science, and his work at Twitter, where he co-founded a machine learning ethics, transparency, and accountability team, he outlines the book’s framework-based, timeless structure and its coverage of moderation, measurement, societal impacts, and individual and policy responses.
https://www.manning.com/books/hidden-influences
00:00 Welcome and Episode Setup
01:32 Meet Luca Belli
02:28 What the Book Is
03:47 Purpose and Nuance
06:16 Luca’s Background at Twitter
07:51 Why Recommenders Matter Now
11:23 Upsides and Community
13:59 Engagement vs Value
17:05 Incentives and Externalities
19:03 Book Structure and Frameworks
22:21 Why Readers Should Care
24:48 Ethical Building Patterns
26:56 Release Timeline and Writing Journey
28:59 Where to Find Luca
29:34 Wrap Up and Thanks
Published: 20 August 2026