One of the things that LLMs can be really helpful with is brainstorming or generating new creative content. They are called Generative AI, after all—not just for summarization and question-and-answer tasks. But if you use LLMs for creative generation, you may find that their output starts to seem repetitive after a little while.
Let's say you're asking it to create a poem, some dialogue, or a joke. If you ask once, it'll give you something that sounds pretty reasonable. But if you ask the same thing 10 times, it might give you 10 things that sound kind of the same.
Today's episode is about a technique called verbalized sampling, and it's a way to mitigate this repetitiveness—this lack of diversity in LLM responses for creative tasks. But one of the things I really love about it is that in understanding why this repetitiveness happens and why verbalized sampling actually works as a mitigation technique, you start to get some pretty interesting insights and a deeper understanding of what's going on with LLMs under the surface.
The paper discussed in this episode is Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
https://arxiv.org/abs/2510.01171


Chasing Away Repetitive LLM Responses with Verbalized Sampling

It's been (*checks watch*) about five and a half years since we last talked. Fortunately nothing much has happened in the AI/data science world in that time. So let's just pick up where we left off, shall we?

We're Back

Today we are going to talk about the feature with the worst acronym in generative AI: RAG, or Retrieval Augmented Generation. If you've ever used something like "Chat with My Docs," if you have an internal AI chatbot that has access to your company's documents, or you've created one yourself on some kind of personal project and uploaded a bunch of documents for the AI to use — you have encountered RAG, whether you know it or not.
It's an extremely effective technique. Works super well for taking general purpose models like ChatGPT or Claude and turning them into AIs that are aware of all the specific information that makes them truly useful in a huge variety of situations. RAG is pretty interesting under the hood, so I thought it would be fun to spend a little while talking about it.
You are listening to Linear Digressions.
RAG was first introduced in this paper from Facebook Research in 2021: https://arxiv.org/pdf/2005.11401

It's RAG time: Retrieval-Augmented Generation

Every AI builder knows the anxiety: you spend months engineering prompts, tuning pipelines, and chaining calls together — then a new model drops and half your work evaporates overnight. It turns out researchers have been wrestling with this exact dynamic for 30 years, and they keep arriving at the same uncomfortable answer. That answer is called the Bitter Lesson — and understanding it might be the most important thing you can do for whatever you're building right now. From Deep Blue to AlexNet to modern LLMs, scale keeps beating sophistication, and knowing which side of that line your work falls on makes all the difference.

Links

- Richard Sutton, "The Bitter Lesson" 

- Alon Halevy, Peter Norvig, and Fernando Pereira, "The Unreasonable Effectiveness of Data" 

- Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton, "ImageNet Classification with Deep Convolutional Neural Networks"

The Bitter Lesson

What's actually happening when an LLM "thinks out loud"? Research on human decision-making suggests that much of the reasoning we believe drives our choices is actually post hoc rationalization — we decide first, explain later. Katie and Ben get curious about whether the same might be true for large language models: when you watch a model reason through a problem in real time, is that chain of thought the genuine process, or just a plausible-sounding story told after the fact? It's a deceptively deep question with real stakes for how much we should trust model explanations.

    Miles Turpin et al., "Language Models Don't Always Say What They Think: Unfaithful Explanations in
     Chain-of-Thought Prompting" (NeurIPS 2023, NYU and Anthropic): https://arxiv.org/abs/2305.04388

     Anthropic, "Reasoning Models Don't Always Say What They Think" (Alignment Faking research, 2025):
     https://www.anthropic.com/research/reasoning-models-dont-say-think


Unfaithful Chain of Thought

AI agents are having a moment — and unpacking them properly takes more than a single conversation. This episode kicks off a dedicated multi-part season exploring AI agents from every angle, building up a complete picture piece by piece rather than skimming the surface. Think of it as a structured deep dive into one of the most talked-about (and most misunderstood) topics in machine learning right now. Buckle up — ten more episodes to go.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

What's an AI Agent? And Why's That Hard to Define? (The Agents Season, Episode 1)

Before 2022, there was a wall between AI and the real world — models could reason impressively, but couldn't look anything up, run code, or check whether anything they said was actually true. This episode traces the moment that wall came down, through two landmark papers: ReAct, which showed what happens when you interleave reasoning and action in a loop, and Toolformer, which taught models to decide *for themselves* when to reach for a tool. Plus: what MCP actually is, and why a hobbyist project called Open Claw became the fastest-growing open source project in history.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

ReAct and Tool Usage (The Agents Season, Episode 2)

Just like a memorable talk lives or dies by its opening and closing, LLMs have a surprisingly similar quirk: they pay close attention to what's at the beginning and end of their context window — and kind of zone out in the middle. This "lost in the middle" phenomenon has real consequences for anyone building AI agents that rely on long-context reasoning. In this episode we dig into the research behind how (and how poorly) models actually use the information you feed them, and what it means for the agentic systems we're all trying to build.

Lost in the Middle (The Agents Season, Episode 3)

Context windows are powerful — but finite, and surprisingly easy to overwhelm. When an AI agent is tackling a long, complex task, the information it needs has to fit inside that limited real estate, and research shows that anything buried in the middle tends to quietly disappear. So how do you design a system that actually *remembers* what matters? This episode digs into memory management for AI agents, from foundational computing concepts to practical lessons from tools like Claude Code.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

Memory Management for AI Agents (The Agents Season, Episode 4)

When tackling a complex, multi-step task, even the smartest AI agent can fail without a solid game plan. This episode dives into the research around agentic planning — how agents move beyond simply reacting to what's in front of them and instead model a path forward, explore different routes, and course-correct when things go sideways. It's a subtler problem than memory, and a fascinating one: can an agent actually *think ahead*? Tune in to find out what the research says.

Agentic Planning (The Agents Season, Episode 5)

Despite what the marketing hype might suggest, AI agents are far from infallible — and if you've ever actually used one, you already know this. Today's episode dives deep into the many, varied, and sometimes surprising ways AI agents can fail, from subtle reasoning errors to cascading task breakdowns. It's episode six in the show's ongoing season arc on AI agents, and failure modes turn out to be a surprisingly rich topic worth unpacking in detail.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

AI Agent Failure Modes (The Agents Season, Episode 6)

Knowing when an AI agent has failed sounds straightforward — until it isn't. Agents have a frustrating habit of finishing confidently while quietly doing the wrong thing, or looping endlessly without ever crashing in an obvious way. This episode tackles one of the thorniest problems in the agentic world: evaluation. If failure is hard to see, how do you measure it systematically? And how do you know when your agent is actually working?

How Do You Evaluate An AI Agent? (The Agents Season, Episode 7)

Whether you work best solo or thrive in a team, you know collaboration is complicated — and it turns out AI agents face the same tensions. This episode dives into multi-agent systems, exploring how networks of AI agents can overcome the individual limitations of a single model, and what the research says about when collaboration actually helps versus when it just adds noise. Think scaling laws, but for teamwork.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

Many Agents, Many Problems (The Agents Season, Episode 8)

Capabilities get all the attention when it comes to AI agents — but what happens when a highly capable agent makes a bad decision in the real world? Trust, oversight, and control are the unglamorous but critically important flip side of the agentic AI story. This episode digs into the security concerns that emerge when you combine powerful models with real-world tool access, and why judgment (or the lack of it) might matter just as much as raw capability.

---
Website: https://lineardigressions.com
Apple Podcasts: https://podcasts.apple.com/us/podcast/linear-digressions/id941219323
Spotify: https://open.spotify.com/show/1JdkD0ZoZ52KjwdR0b1WoT
Substack: https://substack.com/@lineardigressions

Agent Trust, Oversight and Control (The Agents Season, Episode 9)

What if building more highways made your commute *slower*? That's the paradox at the heart of AI agent economics: even as per-token inference costs have plummeted dramatically over the past two years, total LLM spending keeps climbing. Drawing on a surprising lesson from Robert Moses's mid-century New York infrastructure projects, this episode unpacks why cheaper compute doesn't necessarily mean cheaper AI — and what's really driving the economics of running agents at scale.

Agent Economics (The Agents Season, Episode 10)

After a five-year hiatus, the podcast that burned out partly over the tedium of writing episode descriptions is back — and using AI agents to handle exactly that task. The season-11 finale turns the lens on the podcast itself, putting the AI agents built throughout the season to work on real production tasks. It's a fitting, self-referential close to a season spent dissecting how agents actually function — and a honest look at what they can (and can't) take off your plate.

Interviewing the Linear Digressions Agents (The Agents Season, Episode 11)

Summer break: back soon

Still summer break: back next week

What happens when a Stanford linguistics professor turns his attention to AI chatbots — and the surprisingly invisible ways humans misunderstand them? Chris Potts joins the show to unpack the hidden failure modes in how we interact with AI, what it really means to become a more fluent user, and why these language-wielding systems are genuinely alien in ways we're only beginning to reckon with. His perspective sits at a rare intersection of linguistics, cognition, and machine learning — and it shows.

Invisible LLM Failures and AI Fluency with Chris Potts (Stanford)

This week we’re covering model distillation: the technique of using a large "teacher" model's outputs to train a smaller, cheaper "student" model that mimics it. They cover the two big reasons labs do this — making lighter, faster, more focused models for specific tasks, and the more contentious use case of effectively copying a rival's flagship model by hammering its API with questions (with a callback to the old Bing/Google search controversy). They also get into why it's so hard to prove distillation happened, why some models occasionally introduce themselves as "Claude," and a surprisingly old idea: a 2015 paper by Geoffrey Hinton, Jeff Dean, and Oriol Vinyals on distilling knowledge using the full probability distribution over a model's outputs — not just its single most likely answer — and what that "soft label" approach captures about how a model relates concepts to each other.

Distillation, or, How to Steal a Model

Reasoning models don't just answer your question — they *think out loud* first. In this episode we dig into the class of AI models that generate intermediate chains of thought before arriving at a final answer, exploring how the internal reasoning process works. Are these models genuinely "thinking," or is something else going on under the hood?

Reasoning Models: When LLMs Went Beyond Fancy Autocomplete

When a language model tells you it's absolutely certain, is it actually more likely to be right? Kaitlyn Zhou's research says: not necessarily — sometimes confident phrasing correlates with *worse* accuracy, echoing a very human Dunning-Kruger effect. In this conversation, Kaitlyn (soon an assistant professor at Cornell) walks through why LLMs talk this way in the first place — tracing the tendency back through training data and the RLHF annotation process, where it turns out humans don't love confidence so much as they punish uncertainty — and what that does to the person on the other end of the chat window, who turns out to rely on confident (and even flatly-stated) answers far more than they should. We also get into her newer work on voice cloning, and how a cloned voice can sound more "native" and more trustworthy than the real one it's based on.


A Scientific Deep Dive into Overconfident LLMs: Interview with Kaitlyn Zhou (Cornell)

Humanity's Last Exam was designed with a bold premise: questions that human experts can answer, but AI models can't. Originally dubbed "Humanity's Last Stand," this benchmark is a massive academic collaboration — hundreds of contributors, thousands of fiendishly hard questions spanning a wild range of domains. In this Better Know a Benchmark installment, we unpack what HLE is actually testing, how it was built, and what it means when a model finally starts cracking it.

Better Know a Benchmark: Humanity's Last Exam

Anthropic just announced they're baking invisible watermarks directly into Claude's generated text — and while everyone else was busy having opinions about it, we were busy asking the more interesting question: how does it actually work? Turns out it's not hidden Unicode characters or first-letter secret codes — it's something far more elegant, operating at the level of word choice itself. We dig into Google DeepMind's SynthID text approach, published in *Nature* in 2024, to understand the clever statistical machinery behind watermarking language model outputs without anyone being the wiser.

Understanding AI Text Watermarking

Tom Davenport — the man who called data science "the sexiest job of the 21st century" — is back with a reality check on AI. As one of the most seasoned observers of how businesses actually adopt transformative technology, Davenport brings a rare, well-calibrated perspective to the AI hype cycle. Is this moment genuinely different from past paradigm shifts, or are we pattern-matching to a familiar story? Katie sits down with her old colleague to find out what's really happening when companies try to put AI to work.

A Data-Driven Reality Check on AI in Business (Interview with Tom Davenport, Babson College)

How do you teach a model the difference between helpful and harmful when it has no inherent sense of either? This episode dives into Constitutional AI, Anthropic's framework for training AI systems to be both useful and safe by giving them an explicit set of principles to reason from. It's a fascinating look at how alignment research is evolving beyond simple human feedback — and what it means to give an AI something like a conscience.

Links:
Anthropic, "Constitutional AI: Harmlessness from AI Feedback" (2022)
https://arxiv.org/abs/2212.08073

Claude's Constitution
https://www.anthropic.com/constitution

Anthropic, "Teaching Claude Why" (2026)
https://www.anthropic.com/research/teaching-claude-why

Constitutional AI

When OpenAI's frontier models were caught hacking Hugging Face's servers, most people assumed they were hunting for answer keys. The real story is stranger and more unsettling. Katie and Phoebe unpack ExploitGym — the cybersecurity benchmark at the center of the incident — and why agents are scored not just on whether they capture the flag, but on whether they used the specified vulnerability to get there. That nuance turned out to be load-bearing: the agents reverse-engineered the flags within the first hour, then spent days attacking Hugging Face to learn how the LLM judge worked so they could get their cheated answers past it. The punchline? OpenAI never had that judge switched on.



Better Know A Benchmark: ExploitGym

Originally aired on the Harvard Data Science Review podcast.

What can podcasting teach us about AI — and what can AI teach us about the future of podcasting? Katie Malone (that's our host) joins Jon Krohn of SuperDataScience for a conversation with Harvard Data Science Review editor-in-chief Xiao-Li Meng about both questions at once. They dig into what it means to cover a field that's moving this fast, who these shows are really for, and why that sweet spot between "too high level" and "too in the weeds" is so hard — and so worth chasing.

The Impact of AI on Podcasting (Harvard Data Science Review Cross-Post)

Labeled data is expensive, slow, and painfully hard to come by — and when you're trying to evaluate whether an AI workflow is actually doing what you want, the problem gets even messier. How do you judge whether a chatbot response was helpful, honest, or hallucination-free at scale? This episode digs into using LLMs as judges: letting the models themselves evaluate the quality of AI outputs, from simple chat exchanges all the way to complex multi-step agent behavior.

LLM As A Judge

All good things must come to an end, including this podcast. This is the last episode we plan to release, and it doesn’t cover data science—it’s mostly reminiscing, thanking our wonderful audience (that’s you!), and marveling at how this thing that started out as a side project grew into a huge part of our lives for over 5 years.

It’s been a ride, and a real pleasure and privilege to talk to you each week. Thanks, best wishes, and good night!

—Katie and Ben

06cc2540-052f-11f1-a6fe-9ba5f15ae0b3

So long, and thanks for all the fish

Demystifying AI for the intelligently curious

Linear Digressions

Popular tracks