# Field Guide to Fable — Thariq Shihipar — session 2026-07-01T16:05:00.000Z → 2026-07-01T16:25:00.000Z

_78 transcript lines · 40 slides · source: full recording_

## Transcript

Without the support of our sponsors. So please everybody your hands together for the sponsors of the conference. We've got Microsoft, the presenting sponsor. Keep it going. We've got Microsoft. We've got the lab and platinum sponsors. We've got you've got to keep it going. We've got the gold sponsors. We've got silver and bronze. We've got so many sponsors. And this conference Would not be possible without us, so we're very, very thankful. Now we get to introduce, we get to open the state. This is so cool. Today is gonna be such an incredible jam packed agenda, and I hope all of you can make all that you want. I mean there are quite a few tracks, but don't worry, there's a live stream, there's also videos. We're gonna start introducing our first speaker. Oh, I'm excited about this one. Who saw the announcement about Fable yesterday? Yeah, let's go. I, this is so exciting. So, so, coincident The first talk has changed today. We're gonna This conference moves at the speed of AI. It's so cool. Our first speaker, Tariq comes to us from Entropic. Give it up for Tariq comes comes to us from Entropic. Oh, Oh, I'm excited about this. I was talking to him backstage and I said, what what's this gonna be about? This talk. I think the first time it's ever been given, if I'm not mistaken, is about is gonna teach us all how to work with the new myth Those class of models, of which Fable is gonna be soonly available. So your biggest round of applause for Tariq. Please welcome to the stage, member of technical staff at Anthropic, Tariq Shehpar. Hey, everyone. I'm Tarek. I work at Anthropic on Cloud Code. Before we get started, we have a tradition on Cloud Code where, we take a selfie before a talk. So if you don't mind, if you strike a pose with me, I'll, take a quick selfie at AI Engineer. Okay. Incredible. Well, yeah. To kick things off, like we said, Fable is back. We're rolling it out later today. Keep stay tuned for exact timeline. Me and Kat Wu and Simon Wilson will be doing a fireside chat at 12:30. We might have some updates for you then. But Fable is a model I'm just so, so excited about. It's one of those anthropic models where you just, like, you're just gonna remember it. Like, Sonnet 3.5 new, Opus four, Opus 4.5. It's a model that I just have a lot of, like, effect And excitement for. And the best way to describe Fable to me is, like, the the map is opening up. You know? Like, you are playing, like, an RPG, and you've been on the tutorial. And now you get to the point where the, like, you know, the open world starts. Right? And there's so much that you can do and explore, but there's also it's also a little bit intimidating and confusing, right, because there's so much you can do. And so what I wanted to do in this talk Is give you guys a field guide to Fable. Right? How do you work with this new class of models? So I've got four parts to it. I've been working on this as a series of articles and blog posts, but, you know, when we announced Fable was coming out, I was like, okay, let me do all of this at once at the talk, you know, speed run. So, there are four parts. On hobbling Claude, finding your unknowns, dealing with the Brief and being unreasonable. So, first, unhobbling Claude. I think something we say really often is that the models are grown, not designed. Right? We don't wake up and be, like, we need 99% on Sweebench. Right? Like, the models are, you something we we grow carefully. We give it data and feedback and compute. But ultimately, it's something that we It's a little bit organic, and we sort of figure out and learn with the model as we use it. And so, that what that also means is that what contains them is us. Right? The harness we put them in and the way we prompt them is basically, like, a function of our understanding of Claude. Right? And by un hobbling it, I mean, how can we understand Claude better to unleash it? And we need to understand Fable more. So I think one of my point That, you know, we're still so early, and I think there's a lot more understanding in Fable, to unlock. And, I think I'll give you a quick example about how models get smarter because it's a little bit unintuitive. Right? Like, there I saw this viral tweet couple weeks ago being, like, you know, why can't LLMs say which Pokemon end in a w? There are a thousand Pokemon. Right? And turns out there are two who whose names end in a Like, Croconaw and Dreadnaw. Right? And it turns out if you ask, like, a normal chat model, it can't answer it, which is kinda confusing because, like, you know, it definitely knows all the names of the Pokemon. Right? But if you, ask Claude code, it can. Right? Because what it does is that it fetches every Pokemon and writes a script to filter for AW. Right? And so this is what I mean by, like, unhobbling Claude. We call this capability overhang. Right? Claude gets Harder in spiky ways. So it doesn't just remember every Pokemon and reason through it, but if you give it the code execution tool, it can find the two Pokemons that end with a w. Right? And so this is I think, part of the challenge with Fable is figuring out this capability overhang. What is now possible? And I think this is, like, a discovery that I'm excited to go on with you. To make this a little bit clearer, I'm gonna talk about a few different examples of how models have progressed in the past. One of the big examples, obvious Is, like, chat. You know, the chat models were had to be given context. Right? Like, maybe you paste in your code base, and maybe, naively, you might have thought, like, you know, the way we solve coding is by the context just gets really large, and I can just paste in my entire code base. You know, it'll be a 100,000,000 context window. But it turns out that instead, if you give it arms, like you give it the bash tool and ways to work with the environment, it can build and search its own context. And that's sort of, like, the insight that led to cloud code. Right. And so again And spiky, like, a new, like, innovation, kind of, right, in how we think about and work with the model. And then recently, we've rolled out clog tag, and what sort of unlocked clog tag is its ability to work proactively and multi player. Clog code, you know, is something that you have to prompt for it to do work, right? And, this ability for clog to wake itself up and do work is something that we think is unlocking the new wave of agents. But there's there's Or here. So, for example, we recently removed 80% of the system prompt for cloud code. Right? And this is one of the ways in which models, you know, and what they need, changes over time. So originally, like, you know, maybe back in Sonnet 3.5 new, the best practices for a system prompt was a small system prompt, few tools, and lots of examples. Right? And then as the models get smarter, you can give them more information and more instructions and they start following them. And so it's a larger system prompt. With lots of examples and many tools. Right? But most recently, we found this new class of models want fewer want want a smaller system prompt. The examples tend to constrain it because it's actually more imaginative than the examples we give it. And so, and we tried to give it context and not just constraints. We'd really try and avoid being, like, do not do this, which is really necessary for the previous models. And so this is, like, a way that the system Prompt is changing and probably will continue to change. Another feature I really like is the ask user question tool. This is something I worked on when I first got to Cloud Code. And it's, when Cloud, you know, is planning or wants to ask you a question, it can show you a multiple to those dialog. For Opus four, it could barely call it. I had to, like, really tweak the tool to make sure that it was, that it would work, right? And then sometime at Opus four point five, I was like, well, what I asked it to, like, you know, ask me 40 questions about the spec. It could start interviewing me, right? And so, its ability to ask questions jumped, right? And then most recently with Opus 4.8 and Fable, I can now build a whole HTML report with the questions embedded inside of them. And, it's just like a whole new way of interacting with, with Claude. Right? And and so this progression of, like, how Claude can get information from you has also changed. Speaking of which, markdown in HTML is something I've also talked a lot about. You know, it turned Initially, markdown was a a good output for the model. You know, it could show a little bit of rich information. And then, you know, with plan mode, it started to be for you. Like, you could understand what Cloud was about to do. Now, you know, Cloud can build you these in-depth HTML reports. Right? So, again, a way of this the model's getting smarter in a spiky way. I really like to Emphasize that this is closer to a biology than a physics, right? It's still very empirical, very organic. We don't know all the rules, but there is some sort of science behind it, right? Like, there is an intuition to build as well. And so I really encourage you to treat Fable like that. One of my favorite papers, that at Anthropic that we've written is on the biology of a large language model. All of our research papers are meant to be read by people with various degrees of technical expertise, but this is one of my favorites. If you're looking to learn a little bit more, suggest you check it out. But so, yeah, we talked about on hobbling Claude, but it turns out when you're working with Fable, you also need to on hobble yourself. Right? And so one of the things that I think a lot about is that the map is not the territory. Right? When I'm working on a coding problem, the plan and prompt and spec that I have in my mind is the map. Right? But the territory is the actual Base, the real world, the constraints that Claude needs to navigate. Right? And whenever Claude runs into something in the territory that's not in the map, I call that an unknown. Right? Claude has to figure out what to do about it. It's a decision point that I haven't specified. And Fable is one of those first models where I felt that, like, I really have to figure out my unknowns because if not, it's gonna traverse such a large area that, like, it's going to run into a lot of them. So how do you figure out your unknown? Yeah. It I Fables bottleneck my by my ability to match the map and the territory to find my unknowns. So a few, few ways to think about this. I like to think of it in a matrix. So, like, for any problem, I have a bunch of known knowns. This is usually, like, what I write in my prompt. What do I want? Right? Then I have known unknowns. Things that, like, I know I haven't don't really know yet, but I just haven't figured out yet. I can, Yeah. Then I've got unknown knowns. Like, what's so obvious that I just wouldn't write it down? You know? But I'd I'd know it when I see it. Right? And then finally, unknowns and unknowns. What haven't I considered at all? What do I not know? Right? Like, what is something that, if I knew, could change how I prompt Claude? And luckily, you can use Claude, you can use Fable to find your unknowns. So I'm gonna go over a few examples of how I do that with Fable. The first is I like to do what I call a blind spot pass. So I like to say something like, hey. I'm working on a new auth provider that I know nothing about, like, in this code base. Can you do a blind spot pass to help me figure out my relevant unknown unknowns and help me prompt better? Right? And so this, like, might have Claude go through the the auth module and figure out, like, oh, you know, this is kind of like a hairy little, dead end that comes up a lot. Maybe it searches my git diff or Slack. I might tell it where there's context. Right? So that I can learn You know, all the gotchas. And you can use this very broadly, right, you can use it to teach you about new fields. I recently did this for color grading when doing video editing. Because I think it's really powerful and Fable is incredible at it. In many ways, the model knows more about, you know, almost everything than I do. I just need to get it out of it. Then I like to use brainstorms and prototypes. This helps me figure out my unknown knowns. Right? Things like Actually, for design, for me, it's like know it when you see it. Right? So I might ask it to, create a dashboard, and I tell it I have no visual taste. Make me an HTML page with four wildly different design decisions so I can react to them. Right? And then, you know, you tweak this as you want, but, like, the idea is to sort of get an idea of, like, what are the things that you, you know, you can't describe in words. Right? And, like, work with the model to help figure that out. Then Interviews. So once I have an idea of, like, you know, this is what I want to do, there's probably still a lot of, like, unknowns here, right, where I might not have considered something, I might not have specified it, and so I'll ask Claude to interview me, right, and I'll give it a little bit more context. In any of these questions, like, giving it a little bit more context about you and the work and the stage you're at, like, hey, yeah, prioritize questions that would change the architecture, is extremely helpful. Then References. One of the best ways to give Claude a map is to give it another map. Right? So instead of me writing out the spec, I can just say, hey, here is some code that represents what I want to be done. Right? It could be in a different system or language, but just read this code, understand it, and then use that to start your work. Right? And, again, this can be in a lot of different ways. If I'm making a React component, I might have an HTML mock up that is my map. That I pass in as a reference. I think this is really, really powerful, and Fable is really incredible at it. Something else I've, like, really appreciated is implementation notes. So if, while you're running Fable, and it runs into an unknown, ask it to log it, right, so that, you, you can see where the deviations happened, and then you can sort of figure out why as well. You know, it'll usually give you some context about what happened. And then finally, I like to get Of Fable to quiz me about what happened, just to make sure I understand what I'm doing and I can represent this work, you know, when I'm creating a PR or merging it. This is a really great way of, like, making sure that you're, like, really in the loop with Fable. And I think that's one of the most important parts of Fable is staying in the loop and making sure that you, you get what you want. So, those are some of my tips for working with Fable. I also want to say The first time I used a mythos class model, used Fable, I felt both a huge sense of, like, gain, but also a sense of loss. And I I wanted to talk a little bit about that, you know. When I think about coding before LLMs, it feels like a foreign country. You know, like, I used to run a YC startup about 30 people, and we were just constantly forced into trade offs because of how hard code was, right? Like, we could make the the app fast. Astro, we could try prototyping a new feature, and this might take a month, or this would take two months, and so we had to choose. It was just really, really hard. And now I went back to that codebase a couple weeks ago, and I thought about some of the things that I wanted to do, and, it was just way easier. It was, like, the things that would have taken me weeks, I could do in hours. Yeah. And, at some point, it's like, yeah, like, how can you not laugh? Also, how can you not cry honestly? Like, it's like one of these things. Where, I really, really loved programming and writing code by hand. I love the feeling of, like, seeing the code base in my mind and, like, rotating it. But I also remember just, you know, like, staying up late nights trying to debug, working on things for weeks without working, right. I just remember swimming in failure. I just remember that, like, the most of the projects I've ever worked on have failed. Most startups go bankrupt, you know? I think just overall programming and coding is extremely Hard. And, like, as much as I enjoy those highs, I I cannot go back. Right? And, the way my reflection here is, like, the only way out is through. Right? There's still a lot to learn with the agenda coding. There's a lot to learn with Fable. But I think if we try really hard and if we, like, stay in the loop, we un hobble it, we can get there, you know, and we can come out on the other side, with just, so much more. And so the I wanted to talk about is is the so much more part. Right? I call this being unreasonable. One of my favorite parts of anthropic is that we believe that trade offs are not real. Like, I think that very often, I, like, in my previous company, I was very used to being reasonable. So I'd, like, write down this list of priorities, and I'd be like, well, I guess we can prioritize this against this. Right? And, like, you know, that makes sense. So we'll Well, this will be our priority this quarter. But what if you, just did all of it? You know, what if you forest reality to show you the trade off? Right? This is something I've really valued as our culture in Anthropic, and my reflection going forward is that I'm gonna be a lot less reasonable. I think one of this, like, the math of Claude and Fable really changes how you think about trade offs. And there are so many trade offs that you make implicitly in your head. Right? Like good, fast, cheap. Now it's pick three. Right? I think that, like, the best way to, like, do more ambitious work is to, like, reframe and make make ourselves more ambitious. Because I think the only way to prove that agents work is to do the best work of our lives faster than ever before. You know, for example, I made this deck last night in about four hours with Fable. I feel like it's a it's a deck I really like, and I I really enjoyed it, but I also You know, it did it really fast. And I think that if you're here, you know, at AI Engineer, the world is kind of looking at you to prove that AI works. Right? That it's not just, like, a fad or something, but that it can make us more productive and also save us time. And that's my resolution for this year is to be more productive but work less and spend more time with people I really care about.

## Slides

### 00:00:21

- World's Fair
- Microsoft
- OpenAI
- AI Engineer

### 00:00:48

# AI Engineer World's Fair

## Participating Companies / Sponsors
- ORACLE
- paper compute co.
- extend
- RELAI
- Surreal
- SOIOJO

### 00:01:20

- Microsoft
- OpenAI
- AI Engineer World's Fair
- Akamai
[Microsoft logo]
[Akamai logo]

### 00:01:46

- Amazon AGI Lab
- W Fair
- Op
- .ai
- Worl
- Fair
- trust

### 00:02:25

# THARIQ SHIHIPAR
## Member of Technical Staff
### Anthropic

[Ornate, steampunk-style compass or clockwork mechanism graphic]

### 00:02:46

- Amazon AGI Lab
- OpenAI
- AI Engineer - World's Fair
- together
- DATADOG

[Logo of three interconnected circles for "together"]
[Logo of a dog holding a document for "DATADOG"]

### 00:03:25

# Companies and Organizations
- Amazon AGI Lab
- OpenAI
- trust
- Fair
- tog
- DATAD

### Recurring Themes
- AI Engineer
- World's

[

### 00:03:45

# The Map is Opening Up
## In the Land of AI Agents, the Verifiers Are King
Tariq Shaukat / Chief Executive Officer Sonar Makers of SonarQube

[Stylized map showing a central red point with multiple paths branching out to other points, some marked with flags, and icons for mountains and birds.]

### 00:04:23

# A Field Guide to Fable:
1. Unhobbling Claude
2. Finding your Unknowns
3. Dealing with the Grief
4. Being Unreasonable

## In the Land

### 00:04:49

# Models are grown, not designed.

[Line art illustration of a plant growing from the ground, with leaves and a star at the top, next to a grid-like design blueprint]

### 00:05:22

# What contains them is **us** — the harness we put them in, and the way we prompt them.

## In the Land of AI Agents, the Verifiers Are King

Tariq Shaukat / Chief Executive Officer

AI Engineer World's Fair
Presented by Microsoft

### 00:05:52

# The Pokémon that end in “aw”

## Broski (@broskiFGC)
> its cool how half the global economy is contingent on this

*Search query: pokemon

### 00:06:24

# Fetch every Pokémon. Write a script. Filter for "aw."

```bash
Claude Code v2.1.143
Opus 4.7 (1M context

### 00:06:49

### CAPABILITY OVERHANG
Claude gets smarter in spiky ways

[Illustration of a stick figure climbing a steep cliff, with a dotted line indicating a path from the top of the cliff to the climber]

### 00:07:27

# The evolution of agents

- **Chat**
  Has to be given context
- **Claude Code**
  Given arms to execute
- **Claude Tag**
  Acts proactively

### 00:07:48

# The evolution of agents

- **Chat**
  Has to be given context
- **Claude Code**
  Given arms to execute
- **Claude Tag**
  Acts proactively

### Seeing Like an Agent
Thariq Shihipar / Member of Technical Staff ANTHROPIC

[Diagram illustrating three stages of agent evolution: a document and chat bubble, a terminal window, and a document with a highlighted section.]

### 00:08:22

# System Prompt Design

Small system prompt, few tools, lots of examples

Large system prompt, lots of examples, many tools

Smaller system prompt, tool search, no examples

[Diagram showing three stages of system prompt design: a small prompt with few tools and many examples, evolving into a large prompt with many tools and examples, then simplifying to a smaller prompt with tool search and no examples.]

### 00:08:50

# System Prompt Design

Small system prompt, few tools, lots of examples

Large system prompt, lots of examples, many tools

Smaller system prompt, tool search, no examples

[Diagram illustrating three stages of system prompt design, showing changes in prompt size, number of tools, and examples]

### 00:09:21

# AskUserQuestion

- **Opus 4**: Could call it
- **Opus 4.5**: Could interview you
- **Opus 4.8**: Could build the interview

[Three illustrations showing a progression: first, a single UI card with a question mark and progress bar; second, multiple stacked UI cards; third, a UI card with a bar chart and a network diagram.]

### 00:09:54

# AskUserQuestion

**Opus 4**
Could call it

**Opus 4.5**
Could interview you

**Opus 4.8**
Could build the interview

## Engineering the future of AI

[Three icons showing a progression: first a single document with a question mark and text input, then multiple stacked documents, then a document with a bar chart and network graph.]

### 00:10:24

# How Claude communicates: Markdown to HTML
Markdown for itself
Markdown for you
HTML for decisions

## Engineering the future of AI

[Diagram illustrating a three-step process: a document with a checklist, transforming into a document with text and a user icon, then into a dashboard with charts.]

### 00:10:49

# It's closer to a **biology**, than a **physics**.
On the Biology of a Large Language Model
Anthropic · transformer-circuits.pub

[Icon of a stylized

### 00:11:22

# The Map is not the Territory

### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROP\C

[Stylized illustration of a map with a winding path, a triangle, and circles within a circular view]

### 00:11:50

# The Map is not the Territory
### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROPIC

[Abstract illustration with a winding path, a triangle, and circles within a large white circle on a dark background]

### 00:12:22

# The Unknown Matrix

## KNOWN KNOWNS
What do I want?

## KNOWN UNKNOWNS
What haven't I figured out yet?

## UNKNOWN KNOWNS
What's so obvious I'd never write it down?

## UNKNOWN UNKNOWNS
What haven't I considered at all?

### 00:12:51

# The Unknown Matrix

## KNOWN KNOWNS
What do I want?

## KNOWN UNKNOWNS
What haven't I figured out yet?

## UNKNOWN KNOWNS
What

### 00:13:29

## Blindspot pass
> "I'm working on adding a new auth provider but I know nothing about the auth modules in this codebase. Can you do a **blindspot pass** to help me figure out my **relevant unknown unknowns** and help me prompt you better."

### 00:13:50

# Blindspot pass
> "I'm working on adding a new auth provider but I know nothing about the auth modules in this codebase. **Can you do a blindspot pass to help me figure out my relevant unknown unknowns** and help me prompt you better."

### 00:14:26

## Brainstorms and prototypes

> "I want a dashboard for this data but I have no visual taste and don't know what's possible. Make me an **HTML page** with **4 wildly different design directions** so I can react to them."

### 00:14:50

## Interviews
> "Interview me one question at a time about anything ambiguous — **prioritize questions where my answer would change the architecture**."

### 00:15:21

## References
> "This Rust crate in vendor/rate-limiter implements the exact backoff behavior I want. **Read it and reimplement the same semantics** in our TypeScript API client."

### 00:15:52

# Implementation notes
> "Keep an `implementation-notes.md` file. If you hit an edge case that forces you to deviate from the plan, pick the conservative option, **log it under 'Deviations'**, **and keep going**."

### 00:16:22

the most important part of working with Fable is *staying in the loop*
[A small orange dot and a grey arc graphic]

### 00:16:52

# The past is a foreign country
2019-2024 · a 30-person YC startup, ~12 engineers

[Line drawing illustration of two people

### 00:17:23

# The past is a foreign country
2019-2024 · a 30-person YC startup, ~12 engineers

[Illustration of two people working at desks with laptops, and a document icon]

### 00:17:46

You do in **hours** what might have taken weeks, how can you not **laugh**, how can you not almost **cry**?

## Field Guide to Fable
[Illustration of an hourglass with blocks falling from the top bulb to the bottom bulb, where a red lightning bolt is depicted]

### 00:18:24

> The only way out is through

### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROPIC

[Line drawing of a person swimming in water towards a rising sun]

### 00:18:49

## Tradeoffs are not real

### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROP\C

[Line graph showing two dotted trends: one decreasing and one increasing, with an intersection point. A point on the increasing trend is highlighted in red.]

### 00:19:18

## Tradeoffs are not real
### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROP\C

[Line graph with two dotted lines: one decreasing with grey points, and one increasing with grey points, where the final point on the increasing line is highlighted in red.]

### 00:19:46

The only way to prove that agents work is to do the best work of our lives — *faster than ever* before.

### Field Guide to Fable
Thariq Shihipar / Member of Technical Staff ANTHROPIC