From Math Olympiads to Navier-Stokes: How Fast Is AI Progressing?
Sep 29, 2026
The TWIML AI Podcast
Why AI Agents Break the GenAI Security Model with Devvret Rishi
Machine Learning, Artificial Intelligence and Computer Science podcast with Devvret Rishi
The TWIML AI Podcast, Hosted by Sam Charrington
Tuesday 56 min
About
Devvret Rishi discusses how agents with access to tools and external systems change the security requirements of generative AI applications. He examines the limitations of static controls and the need for runtime policies, observability and recovery mechanisms as agent workflows become more capable.
Topics
Related episodes
- Reconstructing how OpenAI agents attacked Hugging FaceJul 30, 2026 · 44 min
- Why Image Generation Needs More Than Bigger Models with Fatih PorikliAug 12, 2026 · 57 min
- From Voice Agents to AI Avatars with Alexander SmolaSep 16, 2026 · 1 h 5 min
- AI poses biosecurity risk, and the ice that doesn’t shatterSep 18, 2026 · 32 min
- Building Durable AI AgentsJul 9, 2026 · 47 min
More from The TWIML AI Podcast
All episodesDo AI Tokenomics Matter More Than Model Benchmarks? with Christopher Potts
Sep 9, 2026The TWIML AI PodcastWorld Models and the Future of Spatial AI with Justin Johnson
Sep 1, 2026The TWIML AI PodcastWhy the Next AI Breakthrough May Come from Physics with Max Welling
Aug 25, 2026The TWIML AI Podcast
More with Sam Charrington
All their workWhy Models Are AI’s Next Training Dataset with Damian Borth
Episode · Jul 27, 2026The TWIML AI PodcastHow AI Learns to Smell with Alex Wiltschko
Episode · Jul 8, 2026The TWIML AI PodcastIs RAG Dead? Lessons from Building AI for Tax Law with Alex Bowcut
Episode · Jun 9, 2026The TWIML AI Podcast
New episodes in Machine Learning
All episodesAI Agents and The Golden Age of Asking Questions with Dimitris Papailiopoulos (MSR/UW-Madison)
Jul 9, 2026The Information BottleneckModels, Harnesses, and Multi-Agent Systems
Aug 6, 2026Practical AI