Controlling AI Models from the Inside

2026-01-20 · 44 min · episode 343 · 6 entities

episode page audio

Asserted relationships

→ hosted by Daniel Whitenack and Chris Benson person

0.55

evidence rules-v5
Feed author/publisher: Daniel Whitenack and Chris Benson
→ discusses Technology company

0.40

evidence rules-v5
Feed category: Technology
→ hosted by Practical AI LLC company

0.40

evidence rules-v5
Feed author/publisher: Practical AI LLC

Entities found in this episode

companys 4

mentioned Practical AI LLC company

0.70

evidence rules-v5
Feed author/publisher: Practical AI LLC
mentioned Technology company

0.50

evidence rules-v5
Feed category: Technology
discusses Technology company

0.40

evidence rules-v5
Feed category: Technology
hosted by Practical AI LLC company

0.40

evidence rules-v5
Feed author/publisher: Practical AI LLC

persons 2

mentioned Daniel Whitenack and Chris Benson person

0.70

evidence rules-v5
Feed author/publisher: Daniel Whitenack and Chris Benson
hosted by Daniel Whitenack and Chris Benson person

0.55

evidence rules-v5
Feed author/publisher: Daniel Whitenack and Chris Benson

Episode description as stored

As generative AI moves into production, traditional guardrails and input/output filters can prove too slow, too expensive, and/or too limited. In this episode, Alizishaan Khatri of Wrynx joins Daniel and Chris to explore a fundamentally different approach to AI safety and interpretability. They unpack the limits of today’s black-box defenses, the role of interpretability, and how model-native, runtime signals can enable safer AI systems. Featuring: Alizishaan Khatri – LinkedIn Chris Benson – Website , LinkedIn , Bluesky , GitHub , X Daniel Whitenack – Website , GitHub , X Upcoming Events: Register for upcoming webinars here !