Castuner

Linear Digressions

Katie Malone

In each episode, your hosts explore machine learning and data science through interesting (and often very unusual) applications. 896520

Podcastaren webgunea

Entzun zure aplikazioan

Atalak

332 atal

  1. GPU Architecture for LLM Inference: Memory, Bandwidth and Hardware Explained 19:38
  2. KV Cache Fundamentals 24:16
  3. AI Inference 101 27:37
  4. Back Next Week 0:18
  5. Back in Two Weeks 0:21
  6. Mechanistic Interpretability: How Researchers Try To See Inside Models 29:21
  7. LLM As A Judge 30:12
  8. The Impact of AI on Podcasting (Harvard Data Science Review Cross-Post) 29:38
  9. Better Know A Benchmark: ExploitGym 32:34
  10. Constitutional AI 31:57
  11. A Data-Driven Reality Check on AI in Business (Interview with Tom Davenport, Babson College) 40:28
  12. Understanding AI Text Watermarking 29:57
  13. Better Know a Benchmark: Humanity's Last Exam 23:21
  14. A Scientific Deep Dive into Overconfident LLMs: Interview with Kaitlyn Zhou (Cornell) 33:54
  15. Reasoning Models: When LLMs Went Beyond Fancy Autocomplete 25:00
  16. Distillation, or, How to Steal a Model 23:38
  17. Invisible LLM Failures and AI Fluency with Chris Potts (Stanford) 41:23
  18. Still summer break: back next week 0:25
  19. Summer break: back soon 0:36
  20. Interviewing the Linear Digressions Agents (The Agents Season, Episode 11) 37:39
  21. Agent Economics (The Agents Season, Episode 10) 24:24
  22. Agent Trust, Oversight and Control (The Agents Season, Episode 9) 25:41
  23. Many Agents, Many Problems (The Agents Season, Episode 8) 28:26
  24. How Do You Evaluate An AI Agent? (The Agents Season, Episode 7) 31:45
  25. AI Agent Failure Modes (The Agents Season, Episode 6) 32:42
  26. Agentic Planning (The Agents Season, Episode 5) 24:00
  27. Memory Management for AI Agents (The Agents Season, Episode 4) 24:41
  28. Lost in the Middle (The Agents Season, Episode 3) 19:44
  29. ReAct and Tool Usage (The Agents Season, Episode 2) 23:41
  30. What's an AI Agent? And Why's That Hard to Define? (The Agents Season, Episode 1) 19:03
  31. Unfaithful Chain of Thought 24:32
  32. Benchmark Bank Heist 12:36
  33. Benchmarking AI Models 29:55
  34. The Hot Mess of AI (Mis-)Alignment 22:32
  35. The Bitter Lesson 19:17
  36. From Atari to ChatGPT: How AI Learned to Follow Instructions 25:53
  37. It's RAG time: Retrieval-Augmented Generation 17:14
  38. Chasing Away Repetitive LLM Responses with Verbalized Sampling 19:12
  39. We're Back 2:58
  40. A Key Concept in AI Alignment: Deep Reinforcement Learning from Human Preferences 19:13
  41. The Impact of Generative AI on Critical Thinking 25:33
  42. So long, and thanks for all the fish 35:44
  43. A Reality Check on AI-Driven Medical Assistants 14:00
  44. A Data Science Take on Open Policing Data 23:44
  45. Procella: YouTube's super-system for analytics data storage 29:48
  46. The Data Science Open Source Ecosystem 23:06
  47. Rock the ROC Curve 15:52
  48. Criminology and Data Science 30:57
  49. Racism, the criminal justice system, and data science 31:36
  50. An interstitial word from Ben 5:59