LIVE: How Professional Writers Write with AI | Write-along
Description
Every is the most AI-native startup on the internet. Through ideas, software and education, subscribers get the tools to work at the frontier of AI. Start your free trial today: https://every.to/subscribe?utm_source=youtube Links: Compound Writing Guide: https://every.to/guides/compound-writing?&utm_source=youtube&utm_campaign=hwwn&utm_content=hwwnlivestream What Writers Who Use AI Want You to Know: https://every.to/p/what-writers-who-use-ai-want-you-to?&utm_source=youtube&utm_campaign=hwwn&utm_content=hwwnlivestream 13 Beliefs About AI Writing: https://every.to/also-true-for-humans/ai-writing-beliefs?&utm_source=youtube&utm_campaign=hwwn&utm_content=hwwnlivestream Good Writing With AI Starts Before the Prompt: https://every.to/also-true-for-humans/good-writing-with-ai-starts-before-the-prompt?&utm_source=youtube&utm_campaign=hwwn&utm_content=hwwnlivestream
Summary
Generated by gpt-5.6-terraAt-a-Glance
- Verdict: Watch fully
- Core thesis: Professional AI-assisted writing works best as a human-directed, staged system—interview, outline, section-by-section drafting, adversarial review, and layered editing—rather than a one-shot generation task.
- Why it matters: The session provides a concrete operating model for turning personal style, editorial judgment, and accumulated feedback into reusable AI skills without surrendering authorship or letting context become unmanageable.
- Best use: Use it as a design reference for an AI-native content pipeline: persistent-but-pruned context, explicit stage gates, role-specific reviewers, and human approval of every consequential editorial decision.
Executive Summary
Every’s editorial team demonstrates an end-to-end AI-native writing and editing workflow rather than arguing abstractly for or against AI writing. Writer Katie Parrott uses a structured folder of style, voice, examples, ideas, drafts, and skills to turn a spoken idea into a publishable essay. Her process begins with an AI interview that surfaces her own thinking, moves through a lightweight “10” outline and a fuller “30” outline, then drafts one section at a time rather than asking the model to write an entire piece in one pass.
The central practical lesson comes from Katie’s own failure: she told her system to save every output, instruction, revision, and experiment. Over time, formerly separate templates and rules fused into conflicting requirements, producing crowded, flat drafts that feedback could not fix. Her remedy was not more prompting but a full reset: archive the old context, rebuild core files, retain only deliberate checkpoints, and treat context as something to prune continuously rather than maximize.
The workflow keeps the human in control through multiple checkpoints. AI reviewer personas test different dimensions of a planned piece—cold-reader comprehension, Hitchcock-style suspense, Sorkin-like pace, Vonnegut narrative structure, objections, and a deliberately hostile “nemesis” reader. These are not accepted as authoritative criticism; they are prompts for reflection, with the writer deciding whether a recommendation is valid and how to implement it.
Senior editor Jack Chang then shows a distinct, deliberately lightweight editing layer. He first comments on the full draft without prescribing fixes, separating editorial diagnosis from solution generation. He uses AI for alternatives, tightening, jargon removal, and production QA, but retains contextual judgment over suggestions. A final agentized “Kate bench” applies the editor-in-chief’s accumulated copy and editorial preferences as tracked changes, allowing the final human editor to spend attention on higher-order issues.
Key Takeaways
- Claim: Treat AI writing as a staged production pipeline, not a one-shot prompting exercise. | Evidence: Katie’s Compound Writing flow is explicitly organized as brainstorm/interview, outline, draft, review, then feedback; she uses a “10” outline for story beats and thesis before expanding to a “30” outline with sections, promises, and open loops. | Implication: For reliable output, introduce explicit approval gates before prose generation—especially thesis, reader payoff, structure, and section-level intent—so fundamental framing errors are corrected upstream. | Caveat: The presenters do not claim every writer needs every stage; they repeatedly frame the system as adaptable to individual practice.
- Claim: More AI context and retained memory can degrade output when instructions, examples, and templates become contradictory. | Evidence: Katie describes saving every outline, decision, feedback item, and experiment; in her Working Overtime context, roughly five templates intended as alternatives became an impossible combined set of requirements, yielding drafts that were “crowded and flat.” She ultimately archived the entire system and rebuilt it. | Implication: Design context as a curated control plane: separate optional templates from mandatory rules, version core instructions, archive historical material outside active context, and periodically delete stale examples and instructions. | Caveat: The diagnosis is an experiential case study, not a controlled test of model-context limits; model changes and workflow changes could also affect output.
- Claim: Never make major context-system changes under deadline pressure or in a highly activated state. | Evidence: Katie says her system expansion happened amid excitement around a new model and continued while vibe-check deadlines were active; attempts to fix poor output by layering on still more instructions made the context worse. Her stated rule is: “never edit your context under pressure or duress.” | Implication: Operationally separate production work from context maintenance: schedule retrospectives and controlled updates, rather than changing foundational prompts, agents, or style files during an urgent deliverable.
- Claim: AI should elicit and structure the writer’s thinking, while the writer retains agency over the argument. | Evidence: Katie uses Monologue plus Claude to interview her one question at a time because speaking helps her discover what she thinks. When Claude pushes toward an interpretation she does not want—such as framing the story around over-dependence—she redirects it rather than following the model’s lead. | Implication: Use models as adaptive interviewers and structured mirrors, not autonomous authors: permit them to probe, but make human acceptance of framing and thesis explicit. | Caveat: A psychologically insightful model can ask useful questions, but its framing can also take the piece in an unwanted direction.
- Claim: Persona-based reviewers create useful, targeted critique when each persona maps to a specific editorial test. | Evidence: Katie uses a cold “Reader” to find comprehension gaps, Hitchcock to assess suspense via the “bomb under the table” principle, Sorkin for pace, Vonnegut for story mechanics, and Objections/Nemesis to expose weaknesses. In the demo, Reader says the damage arrives too late, while Hitchcock argues that moving it earlier could spoil suspense—creating a real editorial trade-off for the writer to resolve. | Implication: Build reviewers around narrow failure modes rather than generic “improve this” prompts, and expect conflicting reviews; disagreement is valuable because it reveals decisions that require human taste. | Caveat: These personas are AI reconstructions of editorial theories, not substitutes for an actual audience or named writers’ judgment.
- Claim: Section-by-section drafting preserves discovery and produces more controllable prose than end-to-end generation. | Evidence: Katie says she must resist the impulse to one-shot a complete piece even when the model has extensive context. She drafts the introduction first—hook, bridge, thesis, and promise—then iterates through sections because the argument changes during writing. By contrast, the pre-baked one-shot draft Jack edits contains over-explicit structure, undefined concepts, and AI-sounding constructions. | Implication: Route task types differently: use fast one-shot generation for constrained, repeatable formats, but use section-level loops for narrative, analytical, or voice-sensitive work where argument discovery matters. | Caveat: Katie reports that a GPT-6 Astra vibe check was successfully one-shotted in eight minutes, so one-shot generation can work for some formats and contexts.
- Claim: The most effective editorial use of AI is suggestion generation and rule-based QA, with humans retaining diagnosis and final judgment. | Evidence: Jack reads the whole draft and leaves comments about what is wrong before asking AI for fixes. He uses a ChatGPT browser extension to inspect the document and attached Google Docs comments, asks for several variants when a suggestion is only partly right, and uses skills such as “Tighten Draft” and “Jargonify.” His tightening skill targets a 10–15% reduction, inspired by Stephen King’s “second draft = first draft minus 10%” rule. | Implication: Separate editorial diagnosis, AI proposal generation, and implementation review. Do not let an agent silently apply broad stylistic rewrites; require visible suggestions, context-aware evaluation, and human acceptance or rejection. | Caveat: Jack notes that AI suggestions can be better in one way and worse in another, can miss house-style evidence available through connected sources, and may require higher effort settings.
Detailed Brief
The active writing context: what belongs in it
- Claims: Katie’s active folder is the engine of her writing process, with a separate folder per recurring column or format.; Her style file defines the column’s purpose, intended reader payoff, structural patterns, strengths, predictable argumentative weaknesses, and a pre-publication checklist.; Her voice file captures sentence-level tendencies and how ideas progress at the prose level; it is distinct from the higher-level style specification.; The system also holds an idea farm, selected examples, and saved outlines/drafts so the model can work from recurring patterns rather than each task beginning from zero.
- Evidence: For Working Overtime, the style file describes the column as a first-person ethnography and laboratory for documenting AI’s emotional, cultural, and practical effects on work.; Katie identifies a recurring “friction to framework” arc: begin with struggle, conflict, or an aha moment; move through examples; end with a reusable reader framework.; She explicitly lists failure modes in her style file, including straw men, false balance, flattened binaries, hedging, and insufficient explanation of expertise or causal reasoning.
- Caveats: The active set should not be an indiscriminate archive. Katie’s rebuild deliberately reduced what gets saved and exposed to the model.; A style system can inadvertently make output formulaic if all prior structural preferences are treated as mandatory conditions.
- Implications: Represent writing knowledge at multiple layers: format/strategy, voice/prose, examples, and quality checks.; Make model-accessible files legible enough that a human can audit conflicts and distinguish hard constraints from optional techniques.
Editorial mechanics and agent implementation details
- Claims: Jack postpones title and subtitle decisions until after he has been immersed in the piece, rather than optimizing packaging before understanding the actual article.; His first editing pass uses comments that state the problem but often avoid prescribing a solution, preserving the writer’s ownership of experience-driven revisions.; His AI skills are intentionally narrow: tightening, jargon removal, and production proofreading rather than a single universal editing agent.; Before final review, the team can run an agentized benchmark of editor-in-chief Kate Lee’s preferences; it creates tracked changes in Google Docs, which Jack triages into clear accepts, clear rejects, and decisions for Kate.
- Evidence: The “Tighten Draft” skill asks the model to read the full draft before changing selected text, identify removable paragraphs before sentence-level edits, favor concrete detail, remove throat-clearing and bridge language, expose subjects and verbs, reduce adverbs, and report the word-count change.; The “Jargonify” workflow removes jargon, then performs a tightening pass, then rereads surrounding text for continuity because a local simplification can create repetition or damage transitions.; The production proofreader catches formatting defects such as line breaks introduced when copying a Google Doc into the CMS.; Kate bench can take up to roughly 10 minutes depending on draft length, so it is run asynchronously rather than treated as instant interaction.
- Caveats: The team remains on Google Docs for publication work partly because editing requires reliable comments, tracked changes, and undo; their internal AI document tool is used more for internal writing.; Kate bench is described as a way to reduce obvious rule-based work, not as a replacement for the final editor-in-chief review.
- Implications: Choose the interface by task: orchestration agents can work in agent environments, while high-fidelity editorial review may belong in a mature document system.; Turn stable editorial preferences into preflight automation, but reserve senior editorial attention for story, argument, audience, and standards.
Notable Concepts & Terms
- Compound Writing: Every’s adaptation of Compound Engineering to writing: a repeatable pipeline that accumulates useful writing context, skills, and feedback over time.
- style.md: A format-level specification covering purpose, audience payoff, structures, strengths, pitfalls, and publication checks for a recurring editorial product.
- voice.md: A prose-level representation of sentence architecture and language patterns intended to make generated drafts feel recognizably like the writer.
- 10 / 30 outline: An upstream quality-control method: validate a minimal thesis-and-beats outline before investing in a fuller structure, preventing expensive reframing later.
- Schlimmbesserung: German for roughly “worsen-bettering”; Katie’s label for improving a context system so aggressively that it becomes less useful.
- Prune your context like a bonsai tree: The operating metaphor for active context management: intentional subtraction and shaping, not unlimited storage and instruction accumulation.
- Bomb under the table: Hitchcock’s suspense model: reveal an impending problem early enough that the audience anticipates it, rather than merely surprising them at the moment of failure.
- Kate bench: An internal agent benchmark encoding Kate Lee’s editorial and copy-editing preferences, used to generate tracked-change suggestions before her final edit.
Operator Notes / Why Ken Should Care
- Create a versioned active-context policy for agent workflows: define what qualifies for promotion into durable instructions, examples, memory, or skills, and require a periodic pruning review.
- Separate production-mode prompts from system-maintenance work. Freeze foundational context files during deadline periods unless there is an explicit rollback path.
- Implement stage gates for high-value written outputs: idea interview, thesis/outline approval, section drafting, targeted critique, developmental review, line edit, and production QA.
- Build a reviewer library where each reviewer has one measurable job—reader comprehension, pace, suspense, objections, jargon, or house style—rather than relying on broad quality prompts.
- Use an editor-agent interaction pattern in which the human first annotates issues and the agent generates multiple candidate fixes; preserve comments and tracked changes for reviewability.
- Audit whether any current agent memory stores process exhaust that should instead be archived outside the active execution context.
Source/Metadata
- Title: LIVE: How Professional Writers Write with AI | Write-along
- Transcript words: 24033
- Duration seconds: 5419
- Timestamp note: No usable timestamps or chapter markers were present in the supplied transcript; substantial duplicated transcript segments were present.
Transcript
We're live, Kate. Thank you. Thanks. Hi. Hello, everyone. Thank you to Katie for letting me know that we're on. We are now live. Thank you for joining us. My name is Kate Lee. I'm the editor-in-chief of Every. We are the only subscription that you need to stay at the edge of AI. We publish ideas, apps, and do trainings. And you may have come to us because you know us for our vibe checks, where we get early access to new models and run them through rigorous testing and publish our findings to you, the audience. I'm here today with two members of the editorial team who I will let introduce themselves shortly. But we have done a series on writing with AI, and we'll share a little bit about what we've done there. And we wanted to cap it off with what we're calling a write-along. So before you go any further, Katie, go ahead and introduce yourself. Hi, my name is Katie Parrott, and I am a staff writer here at Every. I write across a variety of columns that we have. My personal column is working overtime, where I write about how AI is changing work. And then I write our vibe checks and our context window daily roundups of the latest you need to know in AI. So that's me. And Jack. Hey, everyone. I'm Jack Chang, senior editor here. I edit all of the kinds of pieces that Katie writes, and write a few of my own as well. I should also mention that Katie and I are both on the frontier team here at Every, which means that we're tasked with experimenting with different tools, different ways of working. I see Betty's comment in the chat about my post on Jev, and trying to make it do really interesting things. So, yes, very excited to be here to share more about my editing process. Awesome. And our colleague Mike will be joining us in a little bit as well to show his process. So just to set this up, we decided to do a series on writing with AI because, first of all, it's something we already write about regularly. That's something that both Katie and Jack and Mike and everyone on our team and Dan have been writing about. But it also felt like this conversation had reached a fever pitch this summer with lots of different news organizations and journalists and others sharing policies, sharing opinions about writing with AI. And we felt like it was a good moment for us to share what we're doing, why we're doing it, what we believe, and how we think you can do your best work with AI when you know how to use the tools well. This has also grown out of a writing camp and a writing workshop that Katie and I have done. I think we did one or two of them earlier this year, where people were incredibly curious about the process and the actual mechanics. And I think one thing we try to emphasize is that this is a process that works individually for us. It works for Katie or it works for Jack, but you can hopefully adapt it for yourself for what works for you. But we thought that we would use this time to literally take you through the process that Katie goes through when she's writing something, the process that Jack goes through when he's editing something, and the process as well that Mike goes through when he's writing something, because his process is entirely different from Katie's in pretty much every way. I also do want to show you that if you haven't seen it yet, we did publish Katie's guide to writing with compound writing, which is a plugin that we have available to you. And that essentially allows you to incorporate her methodology into yours. So I'm going to stop screen sharing and I'm going to go over to Katie. Let's get started. Let's get started. Bear with me while I talk my way through finding my window. Can you share my screen? Can you see my screen? Yes. Okay. So we see Claude. No, we see StreamYard. Oh, we see StreamYard. Oh, that's not as exciting. Let's see. Share screen. Window, Claude, entire screen. That's what I want. Yes. Great. So now we see Claude. Right? We do. Yes. Okay. Great. So we're going to be writing inside Claude. This is a vibe shift for me because backstory: Claude was actually my first love as a writing model. I remember back in the days of Sonnet 4.5, the team was going crazy for it. I started using it for writing, fell in love with the writing style, but somewhere around the Opus 4.7, 4.8, definitely Opus 5 era, Claude models just stopped working for me and I kind of moved over to ChatGPT, but with the release of Opus 5.5, I really think that Claude has gotten its groove back. Shout out to the team at Anthropic for amazing work. So I'm going to be working inside the Claude desktop app in code, which might sound weird to people that I'm not in the main Claude, the form artist formerly known as co-work. But I like this is just where I live and it makes me happy. So before I go into the actual drafting in Claude, I do want to give a little bit of context about the context that is driving this behavior. So for sure, for those that don't know, folders are the driving force for a lot of us here at Every. We even have a weekly session called "show us your folders" at our standup where on Fridays we show each other our folders and how they help us do our work. The drafting folders that I have, I have one for each of the kinds of columns I do, are the engine behind all of the writing that I do. And there are really three or four key things to know about. The first is my style.md. Style defines the column, what it is, how it works and what a good example of the column looks like. So we have the purpose: working overtime is a first person ethnography of the ways AI is changing work. It is a first person laboratory where Katie Parrott documents the emotional, cultural and practical reality of working with AI while actively experimenting with herself. And then I have reader payoff. A reader should leave with a new lens or framework that they can apply to their work and language that can help them better understand their own relationship with work in the era of AI. Oh, we lost it. We lost it. Live demos. Where do we go? My computer is completely freaking out. I apologize for that. Do you need to do a restart? I shouldn't. Let me just chill out for a second. I can voice over a little bit about what else is in the style.md. So that covers structure. There's something in there called the friction to framework arc, which is actually something that I uncovered through the process of just feeding my essays to Claude and asking, what do you notice? What do these have in common? And it found this pattern in my work where I start with a personal experience of struggle, conflict, or an aha moment. And then we go from the friction through the example to a framework that the reader can use. So that's something that we're going to be trying to work through in this example. And then the other thing that lives in the folder that's really important. Oh, well, here it is. That's really important. Oh, here are some like the sex, success equation, narrative driven analysis, personal vulnerability. Everyone who reads working overtime knows I love nothing so much as to overshare, strategic use of humor and wit. And you just see that—this is just big picture guidance, things to watch out for. Equally important in your style md: argumentative weaknesses, straw men, both sides fallacy, flattening binaries. Here's our friend, "not expert why." Let's get that out of there, hedging, and then a pre-publication checklist. So that's everything that's in the style md. And then the voice md—I'm not going to show because it's not as exciting—but it just captures sentence level patterns. The way I like to architect sentences so that you can see how an idea progresses. That's really important. Here are some of the sex success equation, narrative-driven analysis, personal vulnerability. Everyone who reads working overtime knows I love nothing so much as to overshare strategic use of humor and wit. And you just see that, this is just big picture guidance, things to watch out for equally important in your style MD, argumentative weaknesses, straw men, both sides fallacy, flattening binaries. Here's our friend, not expert why let's get that out of there, hedging, and then a pre-publication checklist. So that's everything that's in the style MD. And then the voice MD, I'm not going to show because it's not quite as exciting, but it just captures sentence level patterns. The way I like to architect sentences so that you can see how an idea progresses. And so now that we've done that, we're just going to run through the process and compound writing. Compound writing descends from compound engineering, which is Kieran Claussen's amazing plugin and framework for compound engineering for software engineering that gets smarter and learns with you. And the core engine of that system is this pipeline of steps you go through. And the thing about writing is that it follows similar steps. You start with brainstorming, you go through a planning phase, which in writing is the outline. You go through drafting, which is the work phase, and then you review and give feedback. So I have an idea for a working overtime article. And so where this is going to start, I'm literally just going to double tap monologue and say, I have an idea for a working overtime essay about how I hopelessly messed up my context and basically ruined my life because the models weren't behaving anymore. Can you interview me one question at a time to draw out my thinking on this topic? And we'll take it from there. Okay. So I'm going to get rid of this just so I know it doesn't really matter if you catch a couple straight thoughts. And then we're just going to let Claude take it away with some questions. It's going to check in. It's going to call the skill, which I love. It's looking at the style guide, the idea farm, all of these things. Don't get too nervous when it says a command failed. I don't usually find that that works. Yeah, this sounds exactly like the piece in your idea farm. Okay. We're going to ignore the existing 30%. I'd like to draft this from first principles because we're in a demo. Can you start the interview process over and we will take it from there. But you see it, it found the ideas. All of these things, my idea farm is where I capture ideas that I want to grow into potential columns. And then outlines and drafts are all saved so I can come back to things. But for the purposes of this demo, we don't want to skip the line. Quick question, Katie. Just from Kashuk, our colleague who wants to know why you're using Monologue instead of Claude's voice mode. Honestly, it's just, I love Monologue. It's an amazing product. It's an every product. Shout out to Naveen. The general, the brains behind the process who we actually just shared that he built an entire language model for voice mode. So we love Monologue and that's just where we do it. So it's like habit. You love it. Go on. Yes. I'm very much leaning into the familiar. This is everything is just muscle memory for me at this point. Yep. And this is how I work. So what happened? Tell me about the moment you realized the models weren't behaving anymore. I was trying to write through a draft of an essay about the relationship between work and play in my AI life. And it just went completely off the rails. I couldn't get the models to do what I wanted. Drafts were coming out overly crowded and flat. There wasn't the shape that I was used to seeing. And no matter how much feedback I gave, no matter how much I tried to brute force my way through the drafting process, it just wasn't working. And this was a sudden change. I had one-shotted a vibe check for GPT-6 Astra with no trouble. And then this was literally the next piece I tried to draft and it had gone completely haywire. And the thing that had happened in the middle of that was that I had gone a little bit crazy and started layering way too much stuff into my context. I made the disastrous decision to tell my machine to save literally everything that it produced, every outline, every decision, every piece of feedback, every experiment. And the result was just an absolute catastrophe in terms of my context's ability to support me and do work the way that I wanted to do it. So that's the origin story behind where we find ourselves now. And the thing, the reason that this is labeled Schlimmbesserung, is that I came across this word. I'm a German major, so I speak German, but that's not how I magically knew this word. Emily Campbell, who works at Figma, had tweeted that she just learned this word, which means literally worsen better. So it refers to the process of trying to make something better only to make it worse. And that's exactly what I had done with my context. I had loaded it down with so much stuff well-intentioned to try to make it better, but I absolutely actually wound up making it worse. So that's my first brain dump. And you'll see it's very long. I kind of just let myself ramble a little bit. This is why I prefer speaking out loud, especially at this early stage of the process. I'm a verbalizer. I'm a yapper. And I kind of find my way to what I think through the process of narrating. So I realized I could have hit enter and been talking and had this thinking churning while we go. But look how fast Opus 5.5 is. It's amazing. So take me back to when you told the machine to save everything. What were you hoping it would do for you? This ties to my new role on the frontier team, where we are actively experimenting on the frontier of AI. And I just wanted a methodology for capturing experiments that I was doing, learnings I was accumulating, processes that were evolving. And I just thought it would be great to capture all of that, both for future learnings and retrospectives, and then also for potential writing. Because this is actually something that I do a lot is, when I'm trying to reconstruct the lessons I've learned for a working overtime piece, for example, I will very often go back into my saved context, my saved chat histories and things to reconstruct what I did. So this is something that I've been able to rely on in the past, but I just over-depended on it. And it turned out to be a horrible mistake. You said this was something you've been able to rely on before. What was different this time? Where did ordinary reliance turn into over-dependence? Okay. And this kind of highlights something sometimes, especially Claude models, they will get a little, they are very psychologically insightful and they will ask you tough questions that actually make you wrestle with what you're thinking and the consequences of it and sometimes take it in a direction that you weren't necessarily thinking about. So for example, where did ordinary reliance turn into over-dependence? I don't know if that's necessarily what I want to drill into in this piece. I'm more, because that's a separate question of over-depending on your context versus just having context you can depend on. So I'm going to say that I'm not necessarily as interested in exploring the idea of over-dependence as much as what happens when context that you've historically been able to trust kind of betrays you. And we should state, Katie, that you were working on this earlier this week. You were working on the very vibe check for Opus 5.5 that, That actually makes you wrestle with what you're thinking and the consequences of it and sometimes take it in a direction that you weren't necessarily thinking about. So for example, where did ordinary reliance turn into over dependence? I don't know if that's necessarily what I want to drill into in this piece. I'm more interested in that. That's a separate question of over depending on your context versus just having context you can depend on. So I'm going to say that I'm not necessarily as interested in exploring over the idea of over dependence as much as what happens when context that you've historically been able to trust kind of betrays you. And we should state Katie that you were, this was earlier this week. You were working with what you were working on was the very vibe check for Opus 5.5 that hopefully many of you have read along with some inputs you gave into the GPT soul. I'm not even remembering which number it was, but which just came out on Tuesday as well. So this was literally three or four days ago. How did you figure out the context was the problem? What was it like to realize the thing you trusted was the thing going wrong? I kind of covered this already, but it really was the only variable that had changed in between the successful Astro Vibe check and this work play piece that had gone completely haywire. So it was about just knowing that's what changed. I didn't change. The models didn't change. They've changed since, but they hadn't changed at this point. And honestly, what was it like to realize the thing you trusted was the thing going wrong? It was terrifying. It was really, really worrying because work doesn't stop. You need to keep making progress through your commitments and realizing that this essential infrastructure that I was really relying on had gone haywire. I needed to stop everything and rebuild from scratch, which I actually have done. And we'll talk about that when we get to solutions. I was trying to throw one-off fixes into the context. I was making it worse by layering on more instructions and more contradictory inputs that just made the models more confused because I was just layering on more and more into these documents like my style.md and my voice.md and my agents.md. And the whole thing just got hopelessly tangled. When you finally looked inside style.md and yada, yada, what did the tangle actually look like? Is there a specific instruction or contradiction that sticks with you? This is honestly where I would probably send the model, like exact. This is exactly the kind of thing that I have on my machine. It's not on this machine though. It's on my Mac mini, which is a journey I'm currently on, trying to harmonize the documents available on both of my machines. But it's less specific. What I noticed in my working overtime context, for example, is that there were like five different templates that I had meant to be different templates that the model could choose from that had somehow gotten fused together into an impossible list of conditions that every draft had to meet. You know, so it had to have a story and a cultural connection and an actionable framework and funny jip. And it just completely broke the model's ability to help me decide which of those building blocks made sense for the piece at hand. Okay. Who do you picture reading this piece? What might they be doing in their own context right now that makes the story matter to them? I think context is really something that you kind of come to the realization of how important it is to how you're building the farther you get into working with AI. So for example, I was talking to a friend of mine who's a little bit farther behind me in this AI journey. And I was telling him how I messed up my context. And he was like, wait, what is context? Why is it important? And then I explained to him a brief history of the concept of context engineering, which started in the developer space and caught on around the time that projects became a thing, however many eons ago that was in AI time. I started adding examples to projects then. I created these style guides again just based on my own bottomless curiosity about myself and my writing process and the output. And I found very quickly that the outputs helped shape pieces the way that I wanted them to be shaped. And then when I made the transition over into the wild west of Claude Code and Codex slash ChatGPT, so much more became possible with having these files on your desktop that are interrelated, that you can orchestrate together to combine with a plugin like Compound Writing to get the AI to work the way that you want it to work with. So I think where this becomes interesting to our readers, you know, maybe they don't have context yet and this is an early cautionary tale to them to keep it simple stupid the way that I should have. Or maybe they're a little farther along and they've noticed some of these challenges with unruly context and things getting out of control. And then they can learn from the process of disentangling and some of the principles that I think I've landed on as far as how to keep this from happening. So just if you're just joining us, for those of you who have just come on, we are going through this is the write along hosted by Every. We are going through Katie Parrot's Compound Writing process to actually show literally how the sausage is made, how she writes a piece using her AI tools and systems. She has been going through having her AI interview her to draw out what she's been thinking about in this piece. So Katie, keep going. So the next question is: tell me about the rebuild. What did you actually do and which principles came out of it? Kieran comes into play here again because I actually just asked my Codex like what's going on with my context, and it without me even specifying called the Compound Engineering skill I didn't know existed called Doc Review. And it actually went through my context and analyzed it against five different core principles like feasibility, product experience, three others that I have amnesia about. And it delivered a report to me about what was wrong with my context. That's where I discovered okay, all these instructions and agents.md and style.md are conflicting and fighting each other and confusing each other. So what I did honestly just earlier this week is I nuked the entire thing. I moved all of my folders into a historical folder and rebuilt them from scratch. So I kind of went through the Compound Engine, the Compound Writing setup process again for my writing folders. I reset the examples that AI had access to. I rebuilt my style.md and my voice.md and just kind of reset everything in a Marie Kondo kind of way, let go of what was not bringing me joy. And the principles that came out of it, there's a couple and I remember them because you know I cheated a little and went through this process before. But you can make something, you can mean to make something better and not make it better. More memory is not always a good thing. Sometimes it pays to let your model forget things that no longer serve you. Prune your context like a bonsai tree. You're kind of always in the process of taking things away, shaping things to the way that you want them to be shaped. So that's kind of some of the examples of the takeaways. And then just because we're a little short on time and I want to make sure other people have time, I'm just going to run through how we go from this to the outline a couple reviewers quickly. And then one note Katie, we do have some time. Mike, Mike just for the audience, Mike is at a conference and unable to find a quiet spot to stream from, so it'll be the three of us, it's the Katie and Jack show. Yes, okay cool, so we'll just keep rocking then. Um, so since the rebuild, what has it been like to work with the models? Has anything come back or stayed broken? So far we're rocking and rolling because we're going through this process and it's going smoothly. I've written some pieces for my personal newsletter that have gone well. I test drove the process for vibe check through a model that we had not reviewed and it went smoothly. So I'm kind of ready to trust my context again. That trust had been broken. The trust in myself had honestly been kind of broken. And that's another thing that I want to say. This is another important lesson: never edit your context under pressure. conference and unable to find a quiet spot to stream from so it'll be the three of us. It's the Katie and Jack show. Yes, okay, cool. So we'll just keep rocking then. So since the rebuild, what has it been like to work with the models? Has anything come back or stayed broken? So far we're rocking and rolling because we're going through this process and it's going smoothly. I've written some pieces for my personal newsletter that have gone well. I test drove the process for vibe check through a model that we had not reviewed, and it went smoothly. So I'm kind of ready to trust my context again. That trust had been broken. The trust in myself had honestly been kind of broken. And that's another thing that I want to say. This is another important lesson: never edit your context under pressure or duress. So at the time that I was doing this, I was kind of in the midst of AI psychosis. Astra had come out. I felt like I was flying and could do anything, and that was really the origin of all these amazing ideas that I thought were so amazing to change my context. And then deadlines happened. Vibe checks had to go out, and I was trying to fix it while I was working, and that's just not a good environment to do these things in. So finding a quiet state of mind from which to work with your context is also important. And so we're getting toward the end. I can tell from the questions that it's asking—like, what do you want the audience to take away? What did you learn? And it says I think there's enough here for an outline. How do you want to go from here? Reflect, generate possibilities, keep asking. So I'm going to have it reflect back what it knows. Let's reflect back what I've said, and then you can tell me, and then we'll shape it from there. I'd say Katie, you've done this enough times that you see you have a sense of the questions and the back and forth and when it's going to say essentially I've gotten what you need. Or I've gotten what I need—excuse me. Yeah, and the thing that I'll say is you're kind of in charge. You're very much in the driver's seat. I think a failure mode of this is that you give the model too much agency and too much control over things. So there are times when it will ask me a question that I'm not interested in, and I'll be like, let's not talk about that. Or I'll just get to the point where I feel like I know what the piece is about and I'll cut it off. But a lot of the time, and this is tied to the fact that it has this context about the structure that my pieces often follow, it naturally kind of pulls out the information that it needs to find the friction, find the action steps, find the lessons, and then we can watch a drafting. So what happened? You one shot a piece with no trouble. The very next piece, an essay about your life, went completely off the rails. Why you saved everything, what you want to explore, how you knew what the tangle looks like. And now obviously I'm not going to stop and read all of this right now because that would make for bad TV, but I would stop and review all of this. And this is something that I always have—just open loops that I haven't resolved. Your original pitch said this basically ruined my life. You haven't yet said how much of that personal toll belongs in the piece. That's where my chronic oversharing, pathological oversharing, comes in. So that is probably something that I would do, but I can always add that. You would go back and specify, is what you mean? Yeah. I would specify how much of the personal toll—like, we can be honest about the personal toll and the fact that I was already not doing so great and getting underwater, and then the fact that my context wasn't helping me anymore put me that much farther behind. And I'm always open to share that kind of stuff. I've written about my mental health on the channel and the column before, so that's fair game to a certain extent. Kieran comes into play here again. Yes, it did. Oh, wait, no. That was on the call before I started interviewing. So Kieran—we have to establish the context that Kieran built something called Compound Engineering, which has tools that can be applied to non-engineering work like doc review. And the three doc review things—I don't remember which—we're not going to dwell on what's wrong or missing. I think we're ready at this point, honestly. I think I'd rather see a 10. So a 10 comes from my background in a content marketing agency called Animals. Our founder instituted this principle of the 10 frame outline and the 30 outline. The idea is there are certain problems with a piece of writing that if you wait until later in the process, they become harder to untangle. So if you wait until you have a whole blog post and the framing is completely wrong or the reader takeaway is completely off, the whole thing falls apart. It's really better to fix those issues upstream. So I took that concept of the 10 and 30, which is like this is a 10—just the story beats and the working thesis. And so I review this, make sure all of this makes sense, and then I move it to a 30. So I'm going to look at the thesis here because that's really the important part. Context you've learned to trust can betray you, and it happens through good intentions. The effort to make it better is what makes it worse. The way back is to prune your context, not add more to it. So that's a good place to start. I might do some things. I actually have something in mind for this. So I want to start with the concept of fresh and investment. I got to use the degree that I got somehow—it was very expensive. So something else that I will often do with these pieces is at the beginning of the process, rather than just the brain dump, I have existing notes. Here I've got all of this stuff that I had written out, and this is what I initially gave the model when I did this pre-bake that Jack is going to edit. So I'm going to just give it this part—the opening. I think I have an opening or a direction for the opening that I'm really happy with. So I'm just going to give you that, and then we can rework the 30 based on this hook. A hook, for those who don't know, is the opening of the piece, and it's called a hook because it needs to pull you in and reel you in like a fish. So what I'm always looking for in a hook is something that opens a curiosity gap. It needs to make the reader wonder, "What does that mean? What happened? How did you deal with it?" And a lot of times the hook or the intro will just come to me fully formed or very close to it, and I'll start there. So you don't have to use AI for every part of the process if it doesn't serve you or if you want to get in there and trad right, if you will. You can totally do that. So here we have our 30 outline for Slim Besolung. And the structure here is very structured. This is actually something I really appreciate about Opus 5.5—it follows this structure because I've had some very know-it-all models lately who don't like to be told how to format things. And that makes it really difficult for me to see the information that I need to see in order to confirm, yes, this is the direction that I want. So here we see the main point—name what you did to your context with the word and establish what the thing you broke is, how you stay sane. The beats. The thesis, which I might workshop a little bit. The promise. And then what's in the folders: worse and bettering, round one. The betrayal, worse and bettering, around two. Diagnosis and rebuild after conclusion. And then kicker's not set yet. That's the conclusion. And then open loops. So this is another checkpoint where if I wasn't live, I would spend some time reviewing this output, changing things. Maybe I want to move some things around. Maybe I think it's getting a little too long to happily fit inside 1500 words, which is very often a problem of mine, so I'll cut some things. It says the crash here—the crash comes after the explainer, the save everything, which is chronological. Do you want a quick flash of the crash earlier so the readers feel the stakes before the backstory? Like, that is actually something that I would say yes to. I want to show the stakes and the consequences. So that's something that I wish I had just hit monologue and told it, because that would be efficient. But so this is what the outline stage is. Diagnosis and rebuild after conclusion and then kicker's not set yet that's the conclusion and then open loops. So this is another checkpoint where if I wasn't live I would spend some time reviewing this output, changing things. Maybe I want to move some things around. Maybe I think it's getting a little too long to happily fit inside 1500 words, which is very often a problem of mine, so I'll cut some things. It says the crash here—the crash comes after the explainer, the save everything, which is chronological. Do you want a quick flash of the crash earlier so the readers feel the stakes before the backstory? That is actually something that I would say yes to. I want to show the stakes and the consequences. So that's something that I wish I had just hit monologue and told it, because that would be efficient. But so this is the outline stage, and this is the point in the process where I start bringing in reviewers. Reviewers are my favorite part of the compound writing plugin. They are a group of personas that I have built based on different writing principles. Some of them are tied to specific writers and storytellers I admire. Some of them are just principles of good writing. So I think what I'd like to run here first of all is reader. Reader is a first time reader—a cold reader who doesn't have the context on the piece. And I just want to check if there's anywhere in this outline where the reader might get lost, where we need to explain things a little bit more. I have this skill set up with this definition. This is what I want you to look at: things like where do you get lost, what's disappointing, what's missing that you would want to see. Oh, it's not here. So let's do Hitchcock instead. I was going to say, can you explain? You've got names for some of your readers—like Hitchcock, Sorkin. I don't know if you've Sedaris. And essentially, what this panel of readers is— Yeah. So we've got Hitchcock is for suspense. So that's really the lean-in principle that's built on. There's a saying that Hitchcock has about a bomb under the table. If you have a bomb under the table and it blows up, okay, you surprise people for one second. But if you show them the bomb and say this bomb will go off in ten minutes, you have ten whole minutes to draw that out, and the reader leans in. So that's the principle behind Hitchcock. The principle behind Sorkin—that's inspired by the famous walk and talk in West Wing. In general, Sorkin's dialogue moves very quickly. And I want to see if the piece is moving along or if it's getting slowed down anywhere. And then who else? Hitchcock is just for concision—keeping it tight, killing our darlings. And Mom is one that I love quite a bit. Mom is somebody who cares about you very much but is confused by everything you're saying. That's a person we all have. Mom reader? Kitty? How do you decide which personas you're going to throw out a piece with? At this point I have a kind of consistent set. Reader I always run because I always want to see where are people getting lost. Hitchcock I very often run when it's a piece that has a story, because I want to know that I understand clearly, okay, are there stakes? Oh, Vonnegut is another one that I really love. That's like Vonnegut has eight characteristics of story—things like start close to the end, respect the reader's time, give them someone to root for. So that's another one that I would run. I don't run Vonnegut on a vibe check, but I run Vonnegut very frequently on working overtime pieces that are narrative driven. So that's some of them. And then there's things like Objections. I will very often run which is just what could somebody say no to? There's also a meaner version of that which I recently renamed from asshole to nemesis—so that's imagining your nemesis reading your piece. But so here we have the reader report, which it actually found despite saying it didn't exist. So the first time reader, it's reading as a working overtime reader—that's a knowledge worker using AI, anywhere from my friends' what-is-context level to someone with their own sprawling setup. What I think the piece is saying: the writer tried to improve the files, let me create AI tools, work broke them, and learned that pruning beats piling on. Overall first impression: the reader waits a long time to see the damage. So maybe we should have taken that guidance to move that damage up the way that the outline step had actually opened up. The reading experience—schadenfreude is one of my favorites that are in the draft. The jackal review is, backpfeifengesicht, which means a face worthy of slapping. And there's another one that involves bacon somehow, so we'll see that. Used in a sentence: I have my context, which means I have ruined my context. I have worsened bettered my context. This causes a stumble. Causes a stumble, causes a stumble. So we're noticing where things are lost or missing, what may put the reader off. AI psychosis—if it isn't calibrated, that's a good call out. That is a very specific experience that is topically sensitive, so I might not want to use that language in the finished piece. What works on first contact: fix first. The question that I still have. So that's reader, and if I'd thought that it was going to work I would have stopped and said okay we need to make this, this, and this change. But I went ahead and did another one. So suspense analysis: current tension medium. The hook plants the bomb well, but the outline then diffuses it for two sections. The bomb under the table is the instruction to say everything. Your hook already had told readers you're worse than bad—you worse and bettered your context—so they know the disaster is coming. Right now the outline covers that scene quickly instead of letting it tick out of the box. So there's more suggestions, and I've got to sit and wrestle with these and be like—so here's a suggested revision of sequence. The hook, the peak, a bomb gets planted, the explosion, round two. So what changed? Where this disagrees with the reader report is interesting. Showing the crash earlier would turn suspense into surprise. Reordering around the astro peak fixes the momentum problem without spoiling the fall. That's your call. So again, this gets really deep a lot of the time. And obviously this is an AI's reconstruction of these kinds of theories, so to speak. But it's a mechanism for self-reflection and saying do I agree with this? Is this something that I would not have caught, that I want to incorporate into my outline at this stage? And so that's the kind of thing that I would do at this stage in the process. But for the purposes of this session, I think we're going to go ahead and move forward to drafting. Well, there's one more thing. Okay, we're going to draft, but we're going to go section by section. This is a point in the process where I have to fight my own impulses because I want to go fast. I want to think this thing knows me so well, it has so much context, I've talked it through the whole story—surely it can write one shot a piece. And you will actually see when Jack edits what happens when I try to one shot things, because it's not great. So the process that I try to follow when I'm behaving myself is section by section. I'll start with the intro. I'll often spend a lot of time on the intro because that really sets up the whole thing. And then I'll go section by section because even with an outline that I've approved through the writing process, you discover things that change. And this is true of traditional writing too. In fact, many of the people who object to the concept of writing with AI argue that the piece emerges from the drafting, and they sort of argue that writing with AI robs you of that process of discovering the piece through the process. But this is my version. I still find that it's just that oftentimes instead of having to write the sentence out myself and decide that I hate it, I have AI write out the first draft, and then that's something for me to react to and be like okay, is this right or is this not quite right? So I'm going to say let's go ahead and start drafting just the intro. Make sure we have the hook, the bridge, the thesis, and the promise. For people not familiar with writerly language, we've talked about the hook—that's what pulls the reader in. The bridge is my shorthand for zooming out to the stakes and why the reader should care and creating that connection from the hook into the thesis. The thesis is your main argument. And the promise is what the reader stands to gain from reading the piece. discovering the piece through the process but this is my version. I still find that oftentimes instead of having to write the sentence out myself and decide that I hate it, I have AI write out the first draft and then that's something for me to react to and be like okay is this right or is this not quite right. So I'm gonna say let's go ahead and start drafting just the intro, make sure we have the hook, the bridge, the thesis, and the promise. Which, those kinds of—for people not familiar with writery language, we've talked about the hook so that's what pulls the reader in. The bridge is kind of my shorthand for zooming out to the stakes and why the reader should care and kind of creating that connection from the hook into the thesis. The thesis is your main argument, and the promise is what the reader stands to gain from reading the piece. It's a gesture toward what the piece contains to give the reader a reason to keep reading. So we're gonna see how our model friend does at producing the draft. It's calling the skill, it's reading the voice guide. It may have read some examples in the background—I'm not 100 percent sure—and hopefully it'll come out with my opening line, which I'm actually very fond of. But we're still thinking about how long does it take then to one shot a whole piece end to end? Honestly, I can tell you the Astra, the Astra vibe check took eight minutes. Yeah, and I just kind of set it up and was like make a goal. I probably didn't need to tell it to make a goal at this point. But honestly, I'm a little surprised that it's thinking so much at this stage because I haven't really seen it take this long, but sometimes it does. And this is the point where—hello, here we go. So we're just gonna read a little bit. "It's a well-known fact of the universe that the Germans have a word for everything. Traubenzucker literally means grease bacon. Fremdschämen—panic bacon—is there. Torschlusspanik—panic gate closing panic—is the fear that time is running out to do the thing you're supposed to do with your life. Oh my God, that's amazing. Recently, this gem came across my feed: Schlimm-Besserung, literally worse and bettering. It refers to the most human of impulses, our desire to make something better only to wind up making it worse. Used in a sentence: I have mined in context for Schlimm-Besserung. I have worsened-bettered my context reader. That's precisely what I have done onto the folders on my desktop, which is inconvenient because those folders are the main way I stay sane out here on the decidedly unpaced here. Inside them is what people who build with AI call context—the style guides, examples and instructions that tell Claude, ChatGPT, who I am, how I write and what I'm working on. When my context is good, the models hand me drafts that already have the shape I'm after. Then mine went bad. And this is an example of something I might change because when my context is good, I expect—and I think this is an expectation that I would want to lean into—is when my context is bad, come after that. So this is just—and then I think the model is probably trying to avoid repetition there because that's something that usually I'm looking out for because models will often state and restate the same idea over and over again. So I think it has a good impulse here, but this is something I might steer, might push back on and be like, let's construct it as when my context is bad, yada yada. "So drafts come back crowded and flat, feedback bounced off them. I was already a little underwater, and the system I built to keep me afloat started pulling me down with it. For a while, I stopped trusting my context. Then I stopped trusting myself. The worst part is that I had done this to myself on purpose, with a clear conscience and a lot of enthusiasm. I gave my context more memory, more rules, more of everything. And each addition made it a little worse at the job I built it to do. Context you've learned to trust can turn on you. Good intentions are how it happens. The way back for me was subtraction. If you haven't built context of your own yet, consider this a cautionary tale: keep it simple, stupid. I don't know that that was more like on-the-fly writing, which sometimes I'll come up with some good lines and I get happy when they get included, but I don't need keep it simple stupid, especially if you have and you're watching yours get unruly. I'll show you how I figured out what went wrong, how I tore the whole thing down, rebuilt it, and the principles I'm using now to keep it from happening again." Honestly, pretty good. This—I was able to read it really naturally because this sounds like me to me. And that's all you can ask for in a first draft: something that's close enough that you want to get in there and tweak here and there. You know, I in a piece I wrote called "Keep Your How to Keep Your Writing Weird in the Age of AI," I talked about the need to kind of get it up and get in there, rough it up, and take some of the smooth prose and make it weirder. But that was in an era before I had this whole comprehensive system that does a lot of that roughing up for me because it uses the language that I used in the interview and it has examples and it has guidance about my sense of humor, my use of examples, and all of that kind of stuff. So things to check: the German examples both are real—that's good. Typo fix, thank you. "I have" is not right. Self-trust—I really like that quite a bit. Thesis wording is pretty good. "The way back for me was subtraction"—we do need to credit Emily Campbell because I don't want it to seem like I came across this on my own. And then it says I haven't saved this anywhere. Tell me if you want it in drafts or we can keep working on it here. And that's something that I instituted as a lesson from this whole disaster that we're sort of reconstructing: I don't want every single outline saved. You know, I want certain checkpoints saved so I can go back, re-examine choices, pull things back in if I like, you know, if I draft something and then draft it different and then decide I want to revert back to the old thing—version control, as we say, as we call it—gets a little tricky when you're working on your desktop versus in a Google Doc where it just naturally saves all that stuff. But so I try to be deliberate about what I save and don't save. So I'm gonna say let's go ahead and save this intro. I'm quite pleased with it. Yeah, I think we can go on and draft section one, and this almost never happens, by the way. Usually, I have to wrestle with it to get the thesis right. I have a thesis skill that gives you three different constructions of the thesis so you can pick which one you agree with, you go back and forth to get the argument right. But for whatever reason, the AI writing gods are with us today. Probably because we're on Opus 5.5, which is just a lovely model to work with. And it's gonna think because it's got to save the dot MD file to the folder. Let me just do a quick interaction and just say where we're at because we are now 55 minutes into this livestream. And Katie, you've taken us through quite extensively your process. Where we are now is Katie has taken us through the compound writing process. Her AI, in this case it's Opus 5.5, has interviewed her based on an idea that she initially had, created two outlines—one a 10 outline, one a 30 outline—that Katie checked both of those. Katie also had her reviewers, her panel of viewers which are based on personas like Hitchcock, Sorkin, Vonnegut, review the piece and the outline for certain qualities. And then she had a one-shot the introduction and went through that introduction and thought it was pretty good, thought it was sort of up to snuff. So I think next we would, mindful of where we are with time, where would we get to next to then get to a full piece and then that Jack would then be taking over to edit in his own completely AI-native way? Yeah, so as I'm writing through section by section and having it save things, usually I will be working—I will have the Google Doc open in the in-app browser. I didn't do that here just because I wanted to keep the screen big and have people be able to follow what's going on. But as I'm drafting and beginning getting content that I'm happy with or almost happy with, I'm porting it over into the Google Doc. I'm getting in there and tinkering with what I want to tinker with manually. And then when the piece gets fully drafted, there are some additional checks that I run. I will do Sorkin again very often to make sure things aren't dragging in places. I will—I have some—there are some—there are official steps in that in the pipeline. So we've got brainstorm, which Get to next, to then get to a full piece and then that Jack would then be taking over to edit in his own completely AI native way. Yeah, so as I'm writing through section by section and having it save things, usually I will be working. I will have the Google Doc open in the in-app browser. I didn't do that here just because I wanted to keep the screen big and have people be able to follow what's going on. But as I'm drafting and beginning getting content that I'm happy with or almost happy with, I'm porting it over into the Google Doc. I'm getting in there and tinkering with what I want to tinker with manually. And then when the piece gets fully drafted, there are some additional checks that I run. I will do Sorkin again very often to make sure things aren't dragging in places. I have some official steps in that pipeline. So we've got brainstorm, which is the inner or in the interview, outline draft. And then there's a line edit stage and a developmental edit stage, which I sometimes do and sometimes I just swap in the specific reviewers I want. But the developmental stage looks at the argument and makes sure it's continuous and logical and supported. And then the line edit goes in and fixes some sentences. Jack actually has a lovely skill for line editing called Titan Draft, which I use quite a bit because it does a really good job. It has a really good sense that Jack has baked into it with his editorial intuition and experience. It kind of knows the kinds of things that we would want to cut. And then I make sure all the links and if I'm behaving myself, the screenshots get in. And that's the point at which I would pass it to Jack, which I will do now. Excellent. So yeah, I'll share my screen here. Let me see. All right, so yeah, you should be seeing my screen here. So what I have here is basically the pre-baked version of this that Katie created. And what I would typically do from here, when a draft from Katie or another writer, whether it's like our other staff writer Laura or someone outside of every, an outside contributor—what I would typically do is the first time through I just want to read it from top to bottom and just kind of have these bigger picture thoughts and comments. For the sake of time and the stream, I'm basically going to kind of combine that. Pretend I did that and you know, the big picture it looks okay and it is fundamentally solid. And then I'm going to do my line edit. So even when I do my line edit, I think you'll find that my setup is very different from Katie's. I'm much more of an ultralight hiker in that sense, in that I just try to keep everything minimal. But the first time I go through, what I'm actually doing is I want to not actually be making suggestions or making edits on the page. What I'm going through and doing is basically reacting to things and putting in comments about what needs to change without necessarily prescribing solutions. So just kind of going from the top, I usually kind of save the title and the subtitle for last, once I've been so immersed in the piece that I really understand it and really understand what we might title it. So I'll just start from here. "It is a well-known fact of the universe that the Germans have a word for everything. Kummerspeck, literally grief bacon, is the weight you gain from eating your feelings. Back five and Geist. I don't know if I'm pronouncing that as a face that's begging to be slapped." I—this is hilarious. So you know, I'm just going to comment a smiley face. And you'll find that as I'm going through, because I'm working with Katie here, she knows that in some cases I'll be leaving notes for her and in other cases I'll be leaving notes for either her AI or my AI. So I tend to be a little more brisk with my notes, and for someone who's contributing from the outside who might not actually be using an AI agent to help them write, I tend to be a little more gentle. Basically, you're not worried about offending the AI? Yeah, I'm not worried about offending the AI. I'm not worried about offending Katie because she knows that sometimes my comments are more meant for an AI. So okay, so recently this gem came across my X feed courtesy of Emily Camp Campbell, Director of AI Model Design at Figma. So like sometimes I'll pick up little style things like we usually bold names and I'll kind of make little changes like that there. So Versuch Schlimmer, "breast rung," literally "worsen bettering," refers to that most human of all impulses: the desire to make something better, only to wind up making it worse. Can definitely relate to that. Use in a sentence: Not going to try to pronounce that. I have worsen bettered my context. I love this. Great transition into the AI topic. So okay, next year. "Reader, that's precisely what I've done to the folders on my desktop." So I'm curious about this because I can recall a few instances where Katie does kind of direct address, saying "reader" like this in Working Overtime pieces. So I'm not entirely sure, maybe that's more common than I think. But it's like, you know, do we typically do this kind of direct address in Working Overtime pieces? So I'm basically just trying to articulate what I'm feeling, what my reactions are as I'm going through. "Reader, that's precisely what I've done to the folders on my desktop, which is inconvenient because those folders are the main way I stay sane out here on the decidedly un-paste AI frontier." This is cute, but maybe needs some more context or at least a link to a post about pacing the frontier, because I feel like if you're not immersed in AI, you're coming to this and you're not aware of what's been in the news, then it might not land. So the next one: "How badly did I worsen better than? Badly enough that one essay ate roughly 135 hours of my life and produced 91 complete drafts. Yikes. And I couldn't get any of them right." So I see a couple things here. How badly did I worsen better than? And there's only a "badly" here, and I think it might be neat to actually follow the structure of this phrase. So maybe I'll highlight this and say I wonder if it works to have lines that follow the worse and better structure. So like "worse enough" or "badly enough that TK." Well enough that TK. Love a TK. Yeah. And for those who aren't familiar with the term, TK is sort of the publishing industry placeholder for something that goes here. And the reason it's those two letters is because they don't typically appear in words next to each other, so it's easier to spot, and you won't actually see a TK in a real phrase. "So I did it to myself. I've been handing AI bigger and bigger pieces of my work because I trusted my context to carry it through." This sounds like AI, and the reason it sounds like AI to me is I feel like AI really loves and overuses handing stuff to other stuff. And then something about "such a good observation." It's always handing things over, handing things off. Yeah. But it's not that alone. It's the handing in combination with "I trusted my context to carry it through." Like something about those in the same sentence just strikes me as AI sounding. So I just mark it like this. "But I built much of that context during a hypomanic stretch." So this sticks out because we should give a little context about what is a hypomanic stretch, how does it pertain to Katie and Katie's writing. So we should quickly define or give context here. Yeah, and this is the kind of thing that if I had not just one shot at this from the previous interview, I did this probably wouldn't have made it in, or if it had, it would have been a little more grounded. But I think like if I did this piece for real, I would keep it more broadly relatable about like hectic, crazy change, things moving fast. We don't need to get into my particular pathology. Yeah, yeah, yeah, for sure, for sure. And yeah, I think it's fascinating, Katie, to actually see your process live because I feel like I've heard and seen bits and pieces of it, and for you... So this sticks out because we should give a little context about what is a hypomanic stretch and how does it pertain to Katie and Katie's writing? We should quickly define or give context here, and this is the kind of thing that if I had not just one shot at this from the previous interview I did, this probably wouldn't have made it in, or if it had, it would have been a little bit more grounded. But I think if I did this piece for real, I would keep it more broadly relatable about hectic, crazy change, things moving fast, and we don't need to get into my particular pathology. Yeah, yeah, for sure. And I think it's so fascinating, Katie, to actually see your process live because I've heard and seen bits and pieces of it, and for you to actually show how you got to this state or how you got to something similar to this was really enlightening for me. So just very quickly continuing, but I built much of that context during a hypomanic stretch and it absorbed my state of mind. A comment that I often leave is: can we say this more directly? Every experiment got saved, every court correction got promoted to rule. The system I was trusting was in effect a transcript of me at my least steady. Systems effectors—there's something about this line that's maybe a little too lyrical. I don't know. I think it leans too much on knowledge that the reader doesn't have yet. Yeah, like "promoted"—like correction got promoted to a rule. What does that mean? Here, yeah, like right. This is about compound writing, but we haven't set it up yet. And then so this is a story of what broke, how I dug my way out, and what I learned. I learned about shaping your contacts while you depend on it. Avoid AI overuse of shape. Also broke—I love to write about things that break all the time, and I think this construction—the story about broke, how I dug my way out, and what I learned—this sounds like AI. It's very like that. I'm very mad that AI took away my rules of three. Like, it's even more devastating than that. You naturally did that yourself. Is what you're saying? Yes. This is a structure I very often follow, and it is a little content marketing. I think my background being in content marketing, not journalism, you see the—I'm going to make it super clear for you. I'm going to map the structure of the piece into the promise. And we want to be more artful than that. Yeah. So kind of like, you know, this intro is an example of how I would go through and do my line edits. I'm just going to say, let's see. I'm just going to kind of speed things up. There's a passage here that I spotted earlier. So like, for instance, just for the sake of demonstration, let's tighten this. Or actually, I'll save that. I'll save that for another thing. There's another one here that I spotted earlier. Let me see if I can find it. We have a comment from Kashuk asking: do you ever flag something as "sounds like AI" but decide to keep it? For sure. Yeah. I think given the right context, or sometimes we have another AI kind of clone of Kate who will do a top edit, and there'll be some things that, you know, that bot will flag that I'm like, actually, I don't mind the use of the word "shape" here. And so there's definitely things that I end up wanting to keep. Yeah, so for sure. Let me see if there was a—let me see. Oh, I'm actually going to delete this one and I'm going to say sounds jargon-ish. Okay. So basically, like, if we do that, then, and like, I'll go through the entire document and kind of leave my comments all the way through. And then from there, depending on a few things—one is depending on how much time we have. In some cases, if we have plenty of time to edit it, I'll kind of like let Katie address these comments without necessarily providing suggestions of my own, unless there's something I feel isn't clear, and I also give an example. Sometimes when we're on deadline crunch, I'll basically try to leave the ones that I think need Katie's specific personal experience to fill in, and for the other ones, I'll try to provide suggestions to help create those suggestions. Sometimes that's when I will use my AI agent. What I've found that I really don't like is editing documents in the cloud code or Codex browser, just because I feel like I want separate apps for different types of activities—like, things that happen in the orchestration apps. Whereas I like to do my editing just like in this—I'm using the DIA browser here—but just straight on the browser. And recently, in the past couple of months, I've really been using the ChatGPT browser Chrome extension. What this does is if you have Codex or the ChatGPT app on your computer, it uses that account and pulls in everything that it has—so all your connections, all your chat transcripts, and things like that. So what I'll do here is basically say, okay, review the comments in this document and let me know your suggestions here. So you'll see I'm using Claude at medium. I haven't really played around with Opus 5.5 enough on editing tasks that I really trust yet, but ever since the Claude models and even GPT 5, I've found that 5.6 on medium does a great job of editing. So lately I've been more using Claude on editing. What it does is this: whatever tab you have open, it knows what you're looking at in the open tab. So if you open a different tab, then this chat context gets separate, and that new tab also has its own chat that is unrelated to this one. So it says, okay, I've read all 14 comments. Keep the German word opening line transition. Both are positive. So this is it reacting to my smiley face here, and sometimes the agent will think I'm addressing it when that's mostly meant for Katie, but no harm, no foul here. Okay. So in place of the paragraph beginning "I did," it says, okay: "I had been giving AI more of my work because the instruction examples in my folders had served me well, but I built much of that material during a hypomanic stretch." So what it's doing here is seeing all these comments and basically trying to address them together. And so what I would be doing is reading this and saying, okay, do I want that? Do I actually want it to tackle these individual comments? And if I do want it to tackle individual comments, one neat thing about the extension is I can actually just highlight something, and when I highlight something, the selection gets passed in as context automatically. Okay. So in place of the paragraph beginning, I did it to myself. It says, okay. I had been giving AI more of my work because the instruction examples in my folders had served me well, but I built much of that material during a hypomanic stretch. So what it's doing here is it's seeing all these comments and trying to address them together. And so what I would be doing is I would be reading this and say, okay, do I want that? Do I actually want it to tackle these individual comments? And if I do want it to tackle individual comments, one neat thing about the extension is I can actually just say highlight something. And when I highlight something, the selection gets passed in as context automatically. So along with the comment, it so it knows, it has access to the Google workspace connector. And so it's able to also figure out what comment is attached to whatever it is highlighting. So maybe I just want to address this highlight for now. So look at the comment here or address. So now it's basically reading the selected passage and its comment and suggesting a revision here. So instead of, I'd been handing AI bigger and bigger pieces of my work because I trusted my contacts to carry it through. It's saying I'd been giving AI more of the writing because the instructions and examples in my folders had worked so well before. This is close, but it's not, I feel like it's better in some ways, but worse than other ways. So I might be, give me a few more variations. Big fan of using AI for options. Yes, for sure. I should just say, we do also have someone in the chat who says no love for proof writing. We are referring to our document, an AI document tool we have called proof that we all use for a lot of internal docs. A lot of internal memos and things like that. But we, when it comes to writing and editing things for publication, we are wedded to Google Docs. Go on. Jack we're wedded to Google Docs. In part because it's familiar. And also I think proof currently doesn't handle undo very well. And so there's sometimes where I'll leave a comment and then I'll change my mind and I'll have to undo it. And so as an editor I use, I rely on that so much that Google Docs is second nature to me. Yeah. Editing the editor. Yeah. And so it gives me more suggestions. Just to jump back a little bit, the other comments, you know, replace reader. That's precisely what I have done. That's what I did to the folders on my desktop. It avoids. So this is an interesting thing. Cause I feel like before, five, six wouldn't have done this. Where five, six would have been, let me look at the working overtime pieces because I have the every MCP. I know the site, let me read the site, figure out. And so this is a case where maybe with six soul, I actually need to crank up the effort level because it seems like it's being a little lazy here. It should have told you that already. It should know what was typical of working overtime. Yeah. It should have told me that already. So what I would do, as I'm going through, it's basically I'm reviewing each of these suggestions one by one and deciding whether or not it's something that I still want to implement or still want to suggest. Whether or not I can come up with something better or whether or not I want to take the AI suggestion, basically doing that through the entire document. And probably right now, editing a piece like this from top to bottom, I would guess it would take me three to four hours. Following this pattern. I'll show you a couple other things that I have. So Katie mentioned a skill that I've set up to tighten drafts. So what that skill does, and I'll show you the skill text, but I can just demo this. So let's highlight this. And so I'll just tighten draft. And that again, it pulls in the selection. And basically the skill is applying the Stephen King rule, which is that a second draft should be a first draft minus 10%. It's partly doing that. And then just partly trying to, there's also suggestions for ways to tighten that I have in the skill. So let's see. Yeah. So keep the three instructions, instruction excerpts and the Great Gatsby minus the parties line. And then change the records piled up by the time I stopped my desktop held 91 separate instruction files just to, by the time I stopped my desktop held, to cut that first part. Get to the failure sooner. What I didn't see was that my system couldn't tell a record from a rule. You can just state that more directly. My system couldn't tell a record from a rule. So it does that. And then it tells you the revised version is the original minus, cut from 172 words to 145 words. And then it's trying to shoot for that 10 percent target. So the other one that I have set up is one that is called the jargon of I. So let's see the example I had here. Is that every experiment, every cat. So let's just do it to this one. I'll have but basically what this does, there's a little bit of overlap with tighten, but it looks for technical language. Like maybe here it would pick up corrections getting promoted to rules. And it tries to state it in plain English. And then I saved our experiment and the system treated each correction as a rule for future drafts, which is a lot more legible to a reader. Yeah. So I'll show you these skills here. Let me share. So basically here's the tighten draft skill. And so it says aim for a 10 to 15% word count reduction when the draft supports it. Stop sooner or further cuts would weaken meaning, voice, rhythm, or necessary context. This is saying it's don't make direct edits to the document. Just tell me the suggested changes in chat. And there's a workflow. It's read the full draft before editing, because you want to know where the passage sits in the context. Cut structural bloat before tightening sentences where sometimes a paragraph might not even need to be there. So I want it to do those larger levels. And then it applies the rules below and then gives me a readout of, so first, it'll flag paragraphs that are worth cutting altogether. And then what characterizes a paragraph that's worth cutting. And then here are editing rules, prefer concrete detail over abstract labels, remove throat clearing sentences, bridge sentences. Like it's worth noting, or this means, or in other words, expose the subject and verb, get rid of there is, and there are, and it is, make the verbs more active. Cut the adverbs, prefer present or past tense instead of gerunds or participles. So I want it to do those larger levels. And then it applies the rules below and then gives me a readout of, so like first, it'll flag paragraphs that are worth cutting altogether. And then what characterizes a paragraph that's worth cutting. And then here are editing rules, like prefer concrete detail over abstract labels, remove throat clearing sentences, bridge sentences like it's worth noting, or this means, or in other words, expose the subject and verb, get rid of the there is, and there are, and it is, make the verbs more active, cut the adverbs, prefer present or past tense instead of gerunds or participles. And then abstract closing sentences that are like from X to Y when you can end on a concrete claim. So that's the Titan draft and then the jargonify also has its own workflow where first I have it remove jargon from the sentence. And then sometimes when I would ask my agent to remove the jargon, it would introduce a bunch of other phrases that would then need to be tightened. So then it has a tightening pass. And then what I found doing that was that sometimes it would end up repeating an idea that's elsewhere in the sentence or it would change that particular sentence, but then the flow and transitions in and out of that sentence would not be that great. So then I have it reread the whole text for flow, which is basically what I would do is if I'm changing a sentence after I make that change, I'm going back and rereading the sentence in context to make sure it continues to flow. And then this is a very short skill, so the output format for each suggested edit. So those are basically the two skills that I use. The only other one that I use personally is one that is a proofreader, that's basically when it comes to production. Sometimes we're copying and pasting the Google document into our CMS and it introduces these line breaks for whatever reason, and so the proofreader will catch that. And then before we end, I can show you one more thing, which is basically what I would do before handing this to Kate for her top edit, which is the final edit before it goes live on the site. Kate edits is the last person to basically touch every single piece. And so this is our thread that Katie initially shared in our company Slack for posting the draft. So let's pretend I've done my pass, Katie's done her edits, the address and we've basically like it's ready for Kate. So what we do before sending it to Kate, if we have time, sometimes we don't always have time, but I will tag every, which is our every agent, and I'll tell every to run Kate bench on this document. And what Kate bench is, is basically Kate's editorial tastes, her copy editing tastes, that have been created kind of the skill and this benchmark to try to replicate as much of that as we can with the thinking that this way, by the time it gets to Kate, some of the obvious things that need changing, I'll have picked up on those that I might've missed in my editing pass. So that when it finally gets to her, she can focus on big picture questions, like does this meet our editorial standards? Are there any big things that need to move rather than these smaller, nitpicky things that are very much rule-based like we don't put spaces around our em dashes and things like that. We do not. And so I think we're getting close to the end of the time, but what this does is it'll take all what it finds here. And Kate bench actually has access, there's a Google doc employee called Kate assistant that Kate bench drives, and then Kate assistant will actually go into the document and as track changes, basically put all those suggestions as track changes throughout the document. And then what I'll do is I'll mark the ones that are either clear rejects or clear accepts and the ones I'm not sure of I'll leave for Kate. And then I'll let Kate know that this draft is ready for her to take a look. You're muted. Okay. I'm unmuted. Thank you. Just a couple of notes. One, you saw that Jack tagged every, meaning the every agent that is a company agent that we have in beta that we're going to be releasing in a couple of weeks. So all of you can get it in your own Slack as well. So look out for that. And also just a note on Kate bench, which I think we've talked about at various points, it probably takes about up to about 10 minutes to run depending on the length of the piece. So it is something that isn't, I think we're all used to AI working immediately, this is something that just takes a little bit longer. So it's the kind of thing we set off on a run and go do something else. And so then you come back to it and we see the document. So I thank you all. Thank you, first of all, Jack and Katie for taking us through your process. We know that narrating that is a lot of work and really appreciate it. Thank you everyone for joining. As a reminder, we are every, every dot to the only subscription you need to stay at the edge of AI. We have one last question I should say is how happy are you with the results? I don't know if either of you wants to chime in on that, knowing that this was a test case, not necessarily a real one. Well, you will get to read a version of this article soon, hopefully. Right? That's the plan, right? I'm not wasting all this work. No, I mean, I was really happy with the intro that we wrote was so much better than the one that I just yeeted out in order to get Jack something to edit. But this is how it works when it's working well. So yay. I figured out my context. Great. Well, thank you. Thank you everyone for joining. And we will see you again soon. the whole thing falls apart and it's really better to fix those issues upstream and so i took that concept of the 10 and 30 which is like this is a 10 which is just the story beats and the the working thesis um and so i review this make sure all of this makes sense and then i move it to a 30 so i'm going to look at the thesis here because that's really the important part um context you've learned to trust can betray you and it happens through good intentions the effort to make it better is what makes it worse the way back is to prune your context not add more to it so that's that's a good place to start um you know i might um i might do some things so like i actually like have something in mind for this a little bit um so i want to start with the concept of fresh and investment so i'm just going to go ahead and tell it pronunciation by the way but go on yeah german i i i gotta use the the i gotta use the the um that the um the i gotta use the the degree that i got somehow it was very expensive but so like something else that i will often do with these pieces as i will at the beginning of the process rather than just the brain dump like i i have existing notes so here i've got all of this stuff um that i had written out and this is what i initially gave the model when i did this pre-bake that jack is going to edit um so i'm gonna just give it this part that's the opening um i think i have an opening or a direction for the opening that i'm really happy with so i'm just going to give you that and then we can rework the we can build the 30 based on based on this hook and a hook for those who don't know is the the opening of the piece and it's called a hook because it needs to pull you in and reel you in like a fish um so what i'm always looking for in a hook is something that opens a curiosity gap so it needs to make the reader wonder oh what does that mean um what happened and like how did you deal with it um and so a lot of times the hook or the intro will just come to me fully formed or very close to it and i'll start there um so you don't that's another point is you don't have to use ai for every part of the process if it doesn't serve you or if if you like want to like get in there and and trad right if you will you can totally do that um so here we have our 30 outline for slim besolung um and the the the structure here is very um very structured and this is actually something i really appreciate about opus 5.5 is that it follows this structure because i've had some very we've had some very know-it-all models lately who don't like to be told how to format things um and that makes it really difficult for me to see the information that i need to see in order to confirm yes this is the direction that i want so like here we see the the main point um name what you did to your context with the word and establish what the thing you broke is how you stay sane um the beats um the thesis which you know i might workshop a little bit um the promise and then what's in the folders uh worse and bettering round one um the betrayal worse and bettering around two um diagnosis and rebuild after conclusion um and then kicker's not set yet that's the conclusion and then open loops so this is another checkpoint where if i wasn't you know live i would spend some time reviewing this output changing things you know maybe i want to move some things around maybe i think it's a little getting a little too long to happily fit inside 1500 words which is very often a problem of mine so i'll cut some things um uh you know it says the crash here the crash comes after the explainer the save everything which is chronological do you want a quick flash of the crash earlier so the readers feel the stakes before the backstory like that is actually something that i would say yes to like i want to show the stakes and the consequences um so that's something that i wish i had just hit monologue and told it that because that would be efficient but so this is what the outline stage and this is the point in the process where i start bringing in reviewers so reviewers are my favorite part of the compound writing plugin they are a group of a bunch of personas that i have built um based on different writing principles um that um come from some of them are tied to specific writers and storytellers i admire some of them are just principles of good writing so um i think what i'd like to run here first of all is reader um so reader is reader is a first time reader this is a cold reader who doesn't know that have the context um on the piece and i just want to check like is there anywhere in this outline even that the reader might get lost where we need to explain things a little bit more um and so i have this skill set up with this definition of this is what i want you to look at i think it's things like where do you get lost what's disappointing like what's what's missing that you would that you would want to see oh what what oh it's not here so let's do hitchcock instead i was gonna say can you explain katie you've got names for some of your readers like hitchcock sorkin um i don't know if you've sedaris um and essentially what this panel of readers is yeah it's it's so we've got hitchcock is for suspense so that's really like the lean in the principle that that's built on is there's a there's a saying that hitchcock has about a bomb under the table where if you have there's a bomb under the table and you're seen and it blows up okay you surprise people for one second um but if you show them the bomb and say this bomb will go off in 10 minutes you have 10 whole minutes to sort of draw that out and the reader leans in so that's the principle behind hitchcock um the principle behind sorkin that's inspired by the famous walk and talk in in west wing in general sorkin's dialogue moves very quickly um and i want to i want to see if the piece is moving along or if it's getting slowed down anywhere um and then um who else hitchcock is just for concision um keeping it keeping it tight killing our darlings um and mom is one that i love quite a bit mom is somebody who cares about you very much but is confused by everything you're saying that's a person we all have a mom reader kitty kitty how do you how do you decide like which personas are you're gonna throw out a piece um at this point i have a kind of a consistent set like reader i always run because i always want to see where are people getting lost um hitchcock i very often run when it's a piece that has a story um because i want to know that this i want to feel like i understand clearly understand okay are there stakes um oh vonnegut is another one that i really love that's like monogut has eight characteristics of story um things like start close to the end respect the reader's time give them someone to root for so that's another one that i would run i don't run vonnegut on a vibe check um but i run vonnegut very frequently on working overtime pieces that are um that are narrative driven um so that's some of them and then there's things like objections i will very often run which is just what could somebody say no to um there's also a meaner version of that which i recently renamed from asshole to nemesis so that's imagining your nemesis reading your piece but um so here we have the reader report which it actually found despite saying it didn't exist um so the first time reader it's reading as a working overtime reader that's a knowledge worker using ai uh you know anywhere from my friends what is context level to someone with their own sprawling setup what i think the piece is saying the writer tried to improve the files let me create ai tools work broke them and learned that pruning beats piling on overall first impression the the reader waits a long time to see the damage so maybe we should have taken that um taken that that guidance to move that damage up the way that the outline step had actually opened up um the reading experience schadenfreude schadenfreude is one the favorite my favorites that are in the draft the jackal review is um uh uh backpfeifen gesicht which means a face worthy of slapping and um and there's another one that's like uh uh it involves bacon somehow so we'll see that um used in a sentence i have my context which means i have ruined my uh context um i have worsened bettered my context this causes a stumble um causes a stumble causes a stumble so we're we're seeing here that we're kind of um like noticing where things are lost or missing um what may put the reader off ai psychosis if it isn't calibrated that's a good call out that is a very specific experience that is topical sensitive so i might not want to use that language in the finished piece um what works on first contact fix first the question that i still have so that's a reader and if i'd thought that it was going to work i would have stopped and said okay we need to make this this and this change but i went ahead and did another one so suspense analysis current tension medium the hook plants the bomb well but the outline then diffuses it for two sections um the bomb under the table is the instruction to say everything um your hook already had told readers you're worse than bad you worse and bettered your context so they know the disaster is coming um right now the outline covers that scene quickly instead of letting it tick out of the box so um there's just more certain more suggestions and like i've got to sit and wrestle with these and be like um so here's a suggested revision of sequence um the hook um the the peak um a bomb gets planted the explosion round two so like um and what changed um where this disagrees with the reader report that's interesting um showing the crash earlier would turn suspense into surprise reordering around the astro peak fixes the momentum problem without spoiling the fall that's your call so again this gets really it gets really deep a lot of the time and like obviously this is an ai's reconstruction of these kinds of theories of different theories of mine so to speak um but it's so it's really it's really a mechanism for self-reflection and saying do i agree with this is this something that i would not have caught that i that i want to incorporate into my outline at this stage um and so that's the kind of thing that i would do um at this stage in the process but for the purposes of this session i think we're gonna go ahead and move forward to drafting uh well there's one more thing okay we're gonna draft but um we're gonna go section by section so this is a this is a point in the process where i have to fight my own impulses because i want to go fast you know i want to be i want to i want to think this thing knows me so well it has so much context i've talked it through it the whole story surely it can write it can one shot a piece um and you will actually see when jack edits what happens when i try to one shot things because it's not great um so the process that i try to follow when i'm behaving myself uh is section by section so i'll start with the intro i'll often spend a lot of time on the intro because that i find that really sets up the whole thing and then i'll go section by section because even with an outline that i've approved through the writing process you discover things that change um and this is true of trad writing too in fact many of the people the objection they have to the concept of writing with ai is that what the piece emerges from the drafting and they sort of and and they argue that um writing with ai kind of robs you of that process of discovering the piece through the process but this is my version um i i still find that it's just that oftentimes instead of having to write the sentence out myself and decide that i hate it um i um i have ai write out the first draft and then that's something for me to react to and be like okay is this right or is this not quite right so i'm gonna say let's go ahead and start drafting just the intro make sure we have the hook the bridge the thesis and the promise um which that's that those those kinds of for people not familiar with kind of writery language we've talked about the hook so that's what pulls the reader in the bridge is kind of my shorthand for the larger for zooming out to like the stakes and why the reader should care and kind of creating that connection uh from the hook into the thesis the thesis is your main argument uh and the promise is what the reader stands to gain from reading the piece it's like a a a a gesture toward what the piece contains to give the reader a reason to keep reading so we're gonna see how our model friend does at producing the draft it's calling the skill it's reading the voice guide um it may have read some examples in the background i'm not 100 sure and hopefully it'll come out with my opening line which i'm actually very fond of um but we're still thinking and about how long does it take then to to one shot what i mean this is a new model so it may be quite fast but yeah it's uh to one shot a whole piece end to end honestly like i can tell you the astra the astra vibe check took eight minutes wow yeah and like i just kind of set it up and was like make a goal and i probably didn't need to tell it to make a goal at this point um but um and like honestly i'm a little surprised that it's thinking so much at this stage because i haven't really seen it take this long but sometimes it does and this is the point where hello here we go um so we're just gonna read a little bit so it's a well-known fact of the universe that the germans have a word for everything commerce back literally grease bacon is bacon from emotional eating bacon is there trossless panic gate closing panic is the fear that time is running out to do the thing you're supposed to do with your life oh my god that's amazing uh recently this gem came across my ex feed for schlimm besserung literally worse and bettering it refers to the most human of impulses our desire to make something better only to wind up making it worse used in a sentence i have a mine in context for schlimm besser or i have worsened bettered my context reader that's precisely what i have done onto the folders on my desktop which is inconvenient because those folders are the main way i stay sane out here on the decidedly unpaced here uh inside them is what people who build with ai call context the style guides examples and instructions that tell clod chat gp2 who i am how i write and what i'm working on when my context is good the models hand me drafts that already have the shape i'm after then mine went bad and like this is this is an example of something i might change because like that when my context is good i expect and i think this is an expectation that i would want to lean into is like i want when my context is bad to come after that um so this is just you know and then like i think the the the the model is probably trying to avoid repetition there because that's something that usually i'm looking out for because models will often state and restate the same idea over and over again so i think it's it has a good impulse here but this is something i might steer i might push back on and be like let's construct it as um when my context is bad yada yada so drafts come back crowded and flat feedback bounced off them i was already a little underwater and the system i built to keep me afloat started pulling me down with it for a while i stopped trusting my context then i stopped trusting myself the worst part is that i had done this to myself on purpose with a clear conscience and a lot of enthusiasm i gave my context more memory more rules more of everything and each addition made it a little worse at the job i built it to do context you've learned to trust can turn on good intentions are how it happens the way back for me was subtraction if you haven't filled context of your own yet consider this a cautionary tale keep it simple stupid um i don't know that that was more like on the fly writing which sometimes i'll come up with some good lines and i get happy when they get included but i don't need no keep it simple stupid especially um if you have and you're you watch yours get unruly i'll show you how i figured out what went wrong how i tore the whole thing down rebuilt it and the principles i'm using now to keep it from happening again um honestly pretty good man um like this it's like this i was able to read it really naturally because this sounds like me to me um and that's that's really what i that's all you can ask for in a in a in a first draft is something that's close enough that yeah you want to get in there and tweak here and there um me and like um you know i in a piece i wrote called uh keep your how to keep your writing weird in the age of ai i talked about the need to kind of get it up and get in there rough it up and like take some of the smooth prose and make it weirder but that was in an era before i had this whole comprehensive system that does a lot of that roughing up for me because it uses the language that i used in the interview and it has examples and it has guidance about my sense of humor my use of examples and all of that kind of stuff so things to check uh the german examples both are real that's good typofix thank you i have is not right um self-trust i really like that quite a bit thesis wording is pretty good um the way back for me was subtraction we do need to credit emily campbell um because i don't want it to seem like i um came across this on my own um and then it says i haven't saved this anywhere tell me if you want it in drafts or we can keep working on it here and that's something that i instituted as a lesson from this whole uh disaster that we're sort of reconstructing is i don't want every single outline saved you know i want certain checkpoints saved so i can go back re-examine choices um pull things back in if i like you know if i draft something and then draft it different and then decide i want to revert back to the old thing version control as we say as we call it um gets a little tricky when you're working on your desktop versus you know in a google doc where it just naturally saves all that stuff but um so i i try to be deliberate about what i save and don't save so i'm gonna say let's go ahead and save this intro uh i'm quite pleased with it um yeah i think we can go on and draft section one and this almost never happens by the way um um usually i have to wrestle with it to get the thesis right i have a thesis skill that gives you three different constructions of the thesis so you can pick which one you agree with you you go back and forth to get the argument right but um for whatever reason the the writing the the ai writing gods are with us today probably because we're with we're on obis uh 5.5 which is just a a lovely model to work with um and um it's gonna think because it's got to save the dot md file to the folder uh let me just do a quick interaction and just say where we're at because we are now 55 minutes into this live stream and katie you've taken us through um quite extensively your process um where we are now is um katie has taken through the compound writing process uh her ai in this case it's opus 55 has interviewed her um based on an idea that that she initially had um created two outlines one a 10 outline one a 30 outline that um katie checked both of those katie also had her um her reviewers uh her reviewers uh her panel of viewers which are based on um personas like hitchcock uh sorkin vonnegut um review the piece for and the outline for certain for certain qualities and then one she had a one shot the introduction um and went through that introduction and thought it was pretty good thought it thought thought it was sort of up to snuff so i think next we would um mindful of where we are with time what where would we get to next to then get to a full piece and then that that jack would then be taking over to edit in his own completely ai native way yeah so as i'm writing through section by section and having it save things usually i will be working um i will have the google doc open in the in-app browser um i didn't do that here just because i wanted to keep the screen big and have people be able to follow what's going on um but as i'm drafting and beginning getting content that i'm happy with or almost happy with i'm porting it over into the google doc i'm getting in there and tinkering with what i want to tinker with manually um and then when the piece gets fully drafted um there are some additional checks that i run i will do sorkin again very often to make sure things aren't dragging in places um i will i have some i have some there are some there are official steps in that in the pipeline so we've got um brainstorm which is the inner or in the interview outline draft and then we there's a line edit stage and a and a or there's a developmental edit stage which i off i sometimes do and sometimes i just swap in the specific reviewers i want but the developmental stage looks at the argument and then and make sure it's continuous and logical and supported and then the line edit goes in and fixes some sentences jack actually has a lovely skill uh for line editing called titan draft which i use quite a bit um because it it does a really good job it has a really good sense that jack has baked into it with his editorial intuition and experience um it kind of knows the kinds of things that we would want to cut and then i make sure all the links and if i'm behaving myself the screenshots get in um and that's the point at which i would pass it to jack which i will do now um excellent uh so yeah i'll i'll share my my screen um here let me see all right so yeah you should be seeing my screen here um so what i have here is basically the um the you the the pre-baked version of of this um that katie created um and what so what i would typically do from here when uh you know when a draft from katie or another writer whether it's like our other staff writer laura or someone outside of every um an outside contributor what i would typically do is um the first time through i just want to read it from top to bottom um and just kind of have these like bigger picture uh thoughts and and comments um for the sake of you know time and and the stream um i'm basically gonna kind of like combine that like pretend i'm gonna pretend that like i did that and you know the the big picture it looks looks okay and it is fundamentally solid um and then i'm gonna go in and i'm gonna do um my line edit um so even when i do my line edit um i i think you'll find that like i my setup is like very very different from katie's um i'm uh i i think like i'm much more of like an ultralight hiker in that in the sense and that like i i just like try to keep everything minimal but the first time i go through what i'm actually doing is i want to not actually be making like suggestions or making edits on the page what i'm going through and doing is basically like reacting to things and putting in comments about what needs to change without necessarily prescribing solutions um so you know so just kind of like going from the top um i usually kind of like save the title and uh you know the subtitle for for last once i'm like so immersed have been so immersed in the piece that i i really understand it and really understand you know what we might title it so i'll just start from here you it is a well-known fact of the universe that the germans have a word for everything uh kummer speck literally grief bacon is the weight you gain from eating your feelings back five and geist i don't know if i'm pronouncing that as a face that's begging to be slapped um i i this is yeah this is hilarious so you know i'm just gonna comment a smiley face um and uh uh you'll find that like it's like as i'm going through because i'm working with katie here like she knows that in some cases i'll be leaving notes for her and in other cases i'll be leaving notes for either her ai or my ai um so i tend to be a little more like brusque with my notes and and and for someone who's like contributing to the outside who might not actually be using an ai agent to help them write i'm tend to be a little more gentle basically you're not worried about offending the ai yeah i'm not worried about offending the ai i'm not worried about offending katie because she knows that sometimes like my comments are more meant for like yeah an ai um so okay so recently this gem came across my x feed courtesy of emily camp campbell director of ai model design at figma um so like sometimes i'll pick up like um you know little style things like we usually like bold um you know bold names and i'll kind of like make little changes like that there um so uh versh schlimmer breast rung literally worsen bettering it refers to that most human of all impulses are desired to make something better only to wind up making it worse um can definitely relate to that use in a sentence uh not going to try to pronounce that i have worsen bettered my context i love this like um like like great transition into you know ai topic um and so uh so okay next year like reader that's precisely what i've done to the folders on my desktop so i'm curious about this because like i can recall a few instances where katie does kind of like direct address saying reader like this um in working overtime pieces so i'm not entirely sure maybe that's more common than than i think but it's like it's like you know do we typically do this kind of direct address in working overtime pieces so i'm i'm basically just like trying to articulate you know what i'm feeling what what my reactions are as i'm going through um reader that's precisely what i've done to the folders on my desktop which is inconvenient because those folders are the main way i stay sane out here on the decidedly unpaste ai frontier um this this is like this is cute but maybe needs some more context or at least a link to uh a post about pacing the frontier because i feel like yeah if you're not immersed in ai then and you're coming to this and you're not aware of what's been in the news then then it could you know it it might not land um so the next one okay how badly did i worsen better than badly enough that one essay ate roughly 135 hours of my life and produced 91 complete drafts yikes um and i couldn't get any of them right um so i i see a couple things here like um like any of them right seems a bit vague uh what does right mean in this context can you be more specific i probably don't need that last bit um the other thing i noticed is like uh how badly did i worsen better than and there's only a badly here and i think it might be neat to actually like follow the structure of this like phrase and so um you know like like maybe i'll highlight this and say you know i wonder if it works to like have lines that follow the worse and better structure so like worse enough or like badly enough that tk um well enough that tk love a tk yeah um and and for for those who who aren't familiar with the terms tk is sort of the the kind of like publishing industry like a placeholder for like you know something goes here um and the reason that it's those two letters is because they don't typically appear in most in words um like next to each other so it's it's easier to to kind of like spot and and and you won't actually like you know see a tk in in in a real phrase um okay so i did it to myself i've been handling handing ai bigger and bigger pieces of my work because i trusted my context to carry it through um this sounds like ai and the reason it sounds like ai to me is i feel like ai really loves like over uses like handing stuff to to other stuff um and then like something about such a good observation it's always handing things over handing things off yeah um so so but but it's like it's not that alone it's the handing in combination with i trusted my context to carry it through like something about those in the same sentence like just strikes me as like ai sounding and so i just like mark it like this um um and then uh but i built much of that context during a hypomanic stretch um um uh so this sticks out because we um we should give like a little context about you know like like what is a hypomanic stretch like what how like how does it pertain to katie and katie's writing so um we should quickly define or give context here about yeah and this is the kind of thing that like if i had not just one shot at this from the previous interview i did this probably wouldn't have made it in or if it had it would have been a little bit more grounded but i think like if i did this piece for real i would keep it more broadly relatable about like hectic crazy change um things moving fast it and we don't need to get into my particular pathology yeah yeah yeah for sure for sure um and uh yeah i i think you know i think it's like it's it's so fascinating katie to actually like see your process live because because i feel like i've heard and seen like bits and pieces of it and and for you to actually like show you know kind of like how you got to this um this state or or like how you got to something similar to this um was really like uh enlightening for me i think um and so just like very quickly continuing um uh uh but i built much of that context during a hypomanic stretch and it absorbed my state of mind so a comment that i often leave is can we say this more directly um um uh every experiment got saved every court correction got promoted to rule the system i was trusting was in effect a transcript of me at my least steady hmm systems effectors like there's something about this line that's maybe like maybe a little too lyrical um i don't know um i think it leans too much on knowledge that the reader doesn't have yet um yeah like what like promoted to like correction got promoted to a rule like what does that mean yeah um here uh yeah like right this is about compound writing but we haven't set it up yet um and then so this is a story of what broke how i dug my way out and what i learned and i learned about shaping your contacts while you depend on it um uh avoid ai over a use of shape also broke the i loves to write about things that break all the time and i think i think it's also like this construction like the story about broke how i dug my way out and what i learned and this sounds like AI. It's very like that. I'm very mad that AI took away my rules of three. Like it's even more devastating than that. You naturally did that yourself. Is what you're saying? Yes. This is a, this is a structure I very often follow and like, it is a little content marketing. Um, like I think like my, my background being in content marketing, not journalism, you see the, like, I'm going to make it super clear for you. Like what the, I'm going to map the structure of the, of the piece into the promise. And we want to be more artful than that. Yeah. Um, yeah. So, so kind of like, you know, this intro is an example of how I would kind of like go through and do my line edits. Um, I'm just going to say, let's see, I'm just to kind of like speed things up. There's a passage here that I, I think like spotted earlier. Um, so like, for instance, like this, you know, just, just for the sake of demonstration, um, um, let's tighten this. Um, and then, or actually, actually, I'll, I'll save that. I'll save that for, for another thing. Um, there's another one here that like, uh, that I spotted earlier. Let me see if I can find it. We have a comment to Kashuk is asking, do you ever flag something as quote, sounds like AI, but decide to keep. For sure. For sure. Yeah. Um, I think, you know, I think like given the right context, um, or like sometimes, you know, sometimes, uh, we have, we have another, like, uh, uh, an AI kind of, uh, clone of, of Kate, who will kind of do a top edit. And there, there'll be some things that like, you know, that bot will flag that I'm like, actually, you know, I don't mind, you know, I don't mind the use of the word shape here. Um, and so, so there's definitely things that like, I end up, you know, uh, wanting to keep. Um, and so, yeah, so for sure. Um, let me, let me see if there was a, let me see. Um, oh, I'm actually gonna hear, I'm gonna delete this one and I'm gonna say, um, sounds jargon-ish. Um, okay. So, so basically like, it's like, if, if, if we do that, then, and like, I, you know, I'll go through the entire document and kind of leave my comments all the way through. And then from there, depending on a few things, like the one is like, depending on how much time we have. Um, in some cases, like if, you know, if we have plenty of time to edit it, I'll just kind of like. Um, let Katie address these comments without necessarily like providing suggestions of my own. Um, if it, unless like, unless, you know, there's something that I feel like isn't clear, unless I also give an example. Um, sometimes when we are like on deadline crunch, um, I'll sort of basically like try to leave the ones that I think need Katie's specific personal experience, um, to fill in. Um, in order, you know, for her and then like the other ones I'll try to like provide suggestions and to, to help me create those suggestions. Sometimes that's when I will use, um, my AI agent. And so what I've found that like, I really, really don't like, um, editing documents in the cloud code or codex browser. Um, just because like, it's like, I, I, I feel like I want separate apps for different types of activities. Um, and so, you know, more like code, like things that happens in the, the orchestration apps. Whereas like, I like to do my editing just like in this, you know, I'm using the DIA browser here. Um, but just straight on the browser. And, um, recently, um, in the past couple of months, I've really been using the, uh, chat GPT, um, browser, Chrome browser extension. And so what this does is if you have codex, uh, or the chat GPT app on your computer, it uses that account and pulls in everything that that has. So all your connections, all your kind of like, you know, like, like, uh, chat transcripts, um, and, and, and things like that. Um, and so, uh, so what I'll do here is basically say like, okay, you know, like review the comments in, uh, in this document. And let me know your suggestions here. Um, so you'll see, I'm using, uh, soul at medium. Um, I haven't really sort of, uh, played around with Opus five, five enough, um, uh, on editing tasks that I like, you know, really trust it yet. Um, but ever since kind of like, you know, the, the, the, the soul models and even GPT five, six, I've just found that like five, six on medium. Does a great job of editing. Um, and so, so lately I've been more using soul on editing. So, um, what it does is this, whatever tab you have open, um, it knows what you're looking at in the open tab. So if you open a different tab, then like this chat context gets, uh, it, that, that new tab also has its own chat that is like unrelated to the, to this one. Um, so it says, okay, I've read all 14 comments. Um, keep the German, uh, word opening line transition. Both they're positive. So this is like it reacting to my smiley face here, you know, and sometimes it'll like, you know, uh, the agent will think I'm like, you know, uh, I'm like addressing it when that's mostly meant for Katie, but it, you know, no, no harm, no foul here. Um, okay. So, uh, in place of the paragraph beginning, I did, um, it to myself. Um, it says, okay. I had been giving AI more of my work because the instruction examples in my folders had served me well, but I built much of that material during a hypomanic stretch. So what it's doing here is it's, it's seeing all these comments and basically like trying to address them together. Um, and so what I would be doing is I would be like reading this and say, okay, you know, it's like, do I want that? Do I actually want it to tackle these individual comments? Um, and if I do want it to tackle individual comments, one neat thing about the extension is I can actually just say like highlight something. And when I highlight something, the selection gets passed in as context automatically. So, uh, along with the comment, um, it, so, so it knows, you know, it, it has access to, um, the Google workspace connector. And so it's able to also figure out like what comment is, um, attached to, uh, whatever it is highlighting. So maybe I just want to like address this, this, uh, highlight for now. So, um, uh, look at the comment here or address. Um, so, so now it's basically reading the selected passage and its comment and suggesting a revision here. So, uh, instead of, I'd been handing AI bigger and bigger pieces of my work because I trusted my contacts to carry it through. It's saying I'd been giving AI more of the writing because the instructions and examples in my folders had worked so well before. This is like close, but it's like not, it's, it's like, I feel like it's better in some ways, but like worse than other ways. So I might be like, give me a few more variations. Big fan of using AI for options. Yes, for sure. I should just say, we, we do also have someone in the, in the chat who says no love for proof writing. We are referring to our document, uh, an AI document, uh, tool we have called proof that we all use for a lot of internal docs. Um, a lot of internal memos and things like that. But we, uh, we do when it comes to writing and editing things for publication, we are wedded to Google docs. Go on. Jack we're wedded to Google docs. Um, in part because it's like familiar. And also I think like, um, proof currently doesn't handle like undo very well. And so, um, like there's sometimes where I'll leave a comment and then I'll change my mind and I'll have to like undo it. And so as an editor I use, I rely on that like so much that, yeah, that it's sort of just, you know, it's like Google docs is like second nature to me. Yeah. Editing the editor. Yeah. Um, and so, yeah, so, so it kind of, you know, it gives me more suggestions. Um, just to kind of like jump back a little bit, like, like the other comments, um, you know, replace reader. That's precisely what I have done. Um, too. That's what I did to, to the folders on my desktop. Um, it avoids. So, so see, this is an interesting thing. Cause I feel like, like before, like five, six wouldn't have done this. Where five, six would have been like, let me, let me look at the working overtime pieces because I have the, every MCP. Like, like, I know the site, let me read the site, figure out. And, and so this is a case where like, you know, maybe with six soul, um, I actually need to like crank up the effort level because it seems like it's being a little like lazy here. Um, like it should have told you that already. It, it should know what was typical of working overtime. Yeah. It should have told me that already. So, so what, what I would do, you know, what, as I'm going through, it's basically like I'm reviewing each of these suggestions one by one and deciding, you know, whether or not, like, like, it's something that I still want to implement or still want to suggest. Whether or not, like I can come up with something better or whether or not I want to like take the AI suggestion, um, basically like kind of like doing that through, you know, the entire document. And probably like right now, like, like editing a piece like this from top to bottom, um, would probably, I would guess it would take me like three to four hours. Um, like kind of like, you know, following this, this kind of, um, pattern. Um, I'll show you a couple other things that I have. So, um, Katie mentioned, um, uh, a skill that I've set up to, um, tighten, uh, drafts. So, um, what that skill does, and I'll show you the, the skill text, but I can just kind of like demo this. Um, so let's like, let's highlight this. And so I'll just like. Tighten draft. And that again, you know, it kind of like pulls in the selection. Um, Um, and basically like the, the, the, the skill is kind of, it's like applying the, um, the Stephen King rule, which is that like, uh, you know, from Stephen King's book on writing, which is saying that like a second draft should be a first draft minus 10%. Um, it's like partly doing that. And then like, just like partly trying to, you know, there's also kind of like suggestions for ways to tighten that I have in the skill. Um, so let's see. Yeah. So keep the three instructions, uh, instruction excerpts and the great Gatsby minus the parties line. Um, and then like, you know, change the records piled up by the time I stopped my desktop held 91 separate instruction files just to, by the time I stopped my desktop held, you know, to cut that first part. Get to the failure sooner. Like what I didn't see was that my system couldn't tell a record from a rule. Um, you can just state that more directly. My system couldn't tell a record from a rule. So it kind of does that. And then that tells you like, you know, it's like the, the revised version is like, you know, the original. Minus, uh, yeah. Uh, uh, cut from one 72 words to one 45 words. And then like, it's trying to shoot for that, like 10 ish percent target. Um, so the other one that I have, um, set up, um, is one that is, um, called the jargon of I. Um, so let's see the example I had here. Um, is that like every, every experiment, every cat. So, so let's just do it to this one. I'll have like, but basically what this does it's it's, there's a little bit of overlap with Titan, but it looks for technical language. Like maybe like here it would pick up, you know, corrections getting promoted to rules. Um, and it tries to state it in plain English. Um, and then, you know, I saved our experiment and the system treated each correction as a rule for future drafts, which is like a lot more legible to like a reader. Yeah. Um, so I'll, I'll kind of show you these, um, these skills here. Let me, um, share. Um, so basically here's the Titan draft skill. And so it says aim for a, you know, 10 to 15% word count reduction when the draft supports it. Stop sooner or further cuts would weaken meaning, voice, rhythm, or necessary context. Um, this is saying it's like, don't make direct edits to the document. Just like, you know, uh, tell me like the suggested changes in chat. Um, and there's a kind of a workflow. It's like read the full draft before editing, because you want to know where the passage sits in the context. Cut structural bloat before tightening sentences where, you know, sometimes like a paragraph might not even need to be there. So I want it to do those like larger levels. Um, and then it kind of like applies the rules below and then gives me like a readout of, um, so like first, like, you know, it'll flag paragraphs that are worth cutting altogether. Um, and then like, you know, like what characterizes a paragraph that's worth cutting. Um, and then here are editing rules, like pre pre prefer concrete detail over, you know, abstract labels, um, remove like throat clearing sentences, um, bridge sentences. Like it's worth noting, or this means, or in other words, um, expose the subject and verb, you know, get rid of the, there is, and there are, and it is, um, make the verbs more active. Like cut the adverbs, um, prefer like present or past tense instead of, uh, gerunds or, you know, participles. Um, and then like, you know, abstract closing sentences that are like, like from X to Y when, um, you know, when you can end on a concrete claim. Um, so that's kind of the, uh, you know, the, the, the, the Titan draft and then the, the jargonify, um, also has its own kind of like workflow where first I have it like remove jargon from the sentence. Um, and then basically like, like then I noticed that sometimes when I would ask my agent to like remove the jargon, it would introduce a bunch of like other phrases. That would then need to be tightened. So then it has a tightening pass. And then like what I found doing that was that sometimes it would end up like basically like repeating an idea that's elsewhere in the sentence or that like it would change that particular sentence. But then the flow and transitions in and out of that sentence would not be that great. So then I have it kind of like reread the whole, whole text for flow, which is basically like what I would do is like, if I'm changing a sentence after I make that change, I'm going back. And like rereading the sentence in context to make sure that like it, it continues to flow. Um, and then, so, you know, this is a very short skill, like then like, you know, the output format for each suggested edit. Um, so, so yeah, so those are basically like the, the, the, the two skills that I use. The, the only other one that I use personally is one that is like a proofreader. That's basically when it comes to production. Um, sometimes like we're copying and pasting the Google document into our CMS and it introduces these like line breaks, um, for whatever reason. And so the proofreader will like, uh, catch that. Um, and then I think like before we end, I can show you one more thing, um, which is basically, um, what I would do before handing this to Kate for her top edit, which is like kind of the final edit before it goes, uh, live on the site. Kate, you know, edits is like the last person to basically like touch every single piece. Um, and so this is, um, so this is our, um, this is the, the thread that, um, uh, Katie initially like shared in our company Slack, um, for like posting the, the draft. Um, and so like, like, let's pretend I've done my pass, you know, Katie's done her edits, the address and we've, we've basically like, it's like ready for Kate. So what we do before sending it to Kate, um, you know, if we have time, sometimes we, we don't always have time. Um, but, um, I will tag every, which is our every agent. Um, uh, and I'll tell every to run, uh, to run Kate bench on this document. And what Kate bunch is, is sort of, um, basically like Kate's editorial tastes, uh, her like copy editing tastes. Um, that have been, um, you know, that we, we've created kind of the skill and this like benchmark to try to like replicate as much of that as we can with the thinking that like. Like, you know, this way, by the time it gets to Kate, like some of the obvious things that need changing, well, you know, I'll have like, kind of like picked up on those that I might've missed in my editing pass. So that when it finally gets to her, you know, she can focus on like the big picture, like questions, like, you know, does this like meet our editorial standards? Like, are there any like big things that need to move rather than like, kind of like these like smaller, like, um, you know, nitpicky things that are very much like rule-based. Like, you know, we don't put spaces around our em dashes and things like that. We do not. Um, and so, uh, I think like we're getting close to the end of the time, but, but what this does is it, it, it'll take all what it finds here. And, um, Kate bench actually has, um, uh, access. There's a, we have a, um, uh, a Google doc employee called Kate assistant that, um, Kate bench drives. And then Kate assistant will actually go into the document and as kind of, as track changes, um, basically like put all those suggestions as track changes throughout the document. Um, and then what I'll do is I'll go in and I'll kind of like mark the ones that are either clear rejects or clear accepts and the ones I'm not sure of I'll, I'll leave for Kate. And then I'll like, I'll, you know, I'll let Kate know that, um, that this draft is, uh, is then, um, ready for her to take a look. You're muted. Okay. I'm unmuted. Thank you. Um, just a couple of notes. One, you saw that Jack tagged every, meaning the every agent that is, um, an agent, a company agent that we have in, in beta that we're going to be releasing in a couple of weeks. So all of you can, uh, can, can, can, can get it in your own Slack as well. Um, so look out for that. Um, and also just, uh, just a note on Kate bench, which I think we've talked about at various points. It probably takes about, um, it can take up to about 10 minutes to run depending on the length of the piece. Um, so it is something that you're, you know, isn't, you know, I think we're all used to sort of AI working immediately. Um, this is something that just takes a little bit longer. So it's the kind of thing we set off on a run and go do something else. Um, and so, uh, and so then you come back to it and we see, we see the document. Um, so, um, I thank you all. Thank you, first of all, Jack and Katie for taking us through your process. Uh, we know that that's, uh, narrating that is a lot of work, uh, and really appreciate it. Thank you everyone for joining. As a reminder, we are every, every dot to the only subscription you need to stay, um, at the edge of AI. We have one last question I should say is how happy are you with the results? Uh, I don't know if either of you wants to chime in on that, knowing that this was a test case, not a, not necessarily a real, a real one. Well, you, you will get to read a version of this article soon, hopefully. Right? Like that, that's the plan, right? I'm not wasting all this work. Um, no, I mean, I was really happy with the, the intro that we wrote was so like, live was so much better than the one that I just yeeted out, you know, in order to get Jack something to, to edit. But, um, this is, this is how it works when it's working well. So yay. I figured out my context. Great. Well, thank you. Thank you everyone for joining. Uh, and we will see you again soon.