ABOUT THIS EPISODE
An AI answers that the animal spinning a web has eight legs. Researchers find an internal representation for “spider,” replace it with “ant,” and the answer changes to six.
Mechanistic interpretability attempts to reverse-engineer the internal computations of neural networks. Follow researchers from numerical activations to features, circuits, and causal interventions—and learn why an AI's spoken explanation is not proof of how it actually reached an answer.
English
United States
TRANSCRIPT 🔗
Are you the producer of this podcast?
Add a podcast transcript
Need Audio-to-Text?
Transcribe with Listen411 in Just 60 Seconds
SEARCH PAST EPISODES
Search past episodes of In Simple Terms with Satish.
OTHER EPISODES IN THIS PODCAST
Two developers use the same AI coding tool, the same project, and the same deadline—but follow very different working methods.
Ship-It Sam optimizes for speed. Verify Veera defines assumptions, acceptance criteria, and evidence before trusting the output. At 9:20, who was really ahead?
Learn how …
Deleting a file does not necessarily remove what an AI model learned from it. Machine unlearning asks a harder question: can we remove a specific training influence while preserving everything the model should retain?
In this episode:
• The difference between deleting data and making AI forget
• F…
A consulting company adopts AI to help its employees research clients and prepare proposals. The pilot saves hours, employees love it, and management quickly expands it—then an unexpected bill arrives.
Follow the company as it opens the hidden receipt behind one AI-generated proposal and discovers…
You watch an extraordinary rescue video, read hundreds of enthusiastic comments, and then notice that almost nothing looks quite real. Was the video made by AI? Were the comments posted by bots? Was there ever a human story behind it?
Follow one suspicious post through the machinery of generative …
Disclaimer: The podcast and artwork embedded on this page are from Satish Choudhary, which is the property of its owner and not affiliated with or endorsed by Listen Notes, Inc.
EDIT
Thank you for helping to keep the podcast database up to date.
