https://aiconversations.blog/podcast en-gb Conversations with the people building, breaking and governing intelligent systems. Segun O. Solar A fortnightly interview show about the work behind capable systems: how they are evaluated, where they fail, who is accountable for them, and what it takes to put them in front of real people. false https://aiconversations.blog/podcast/the-middle-of-the-task https://aiconversations.blog/podcast/the-middle-of-the-task Wed, 28 Oct 2026 08:00:00 GMT Amara Nwosu on why agent benchmarks reward the ending and ignore the forty steps that produced it. 47:00 https://aiconversations.blog/podcast/opening-the-black-box https://aiconversations.blog/podcast/opening-the-black-box Wed, 14 Oct 2026 08:00:00 GMT Ravi Anand on mechanistic interpretability, sparse autoencoders, and what a feature actually tells you. 55:12 https://aiconversations.blog/podcast/the-compliance-surface https://aiconversations.blog/podcast/the-compliance-surface Wed, 30 Sep 2026 08:00:00 GMT Nadia Okafor on AI governance, audit trails, and the gap between a policy document and a running system. 44:05 https://aiconversations.blog/podcast/breaking-models-on-purpose https://aiconversations.blog/podcast/breaking-models-on-purpose Wed, 16 Sep 2026 08:00:00 GMT Tom Whitfield on red teaming as an engineering discipline rather than a launch-week formality. 50:10 https://aiconversations.blog/podcast/model-risk-in-a-regulated-room https://aiconversations.blog/podcast/model-risk-in-a-regulated-room Wed, 02 Sep 2026 08:00:00 GMT Priya Raman on shipping AI inside financial services, where the model card is the least interesting document. 41:30