Complete Course: Claude for PMs (Cowork + Code + Dispatch)

1. 0:00 – 0:54

   Intro

   1. AGAakash Gupta0:00

      There's no good reason to talk to Claude just over a normal web chat anymore

   2. PHPawel Huryn0:03

      I don't use chat at all. I use it sporadically, just if a window is open, I can ask some questions like, "Is this grammatically correct?"

   3. AGAakash Gupta0:10

      Claude is way better at using PowerPoint than it was. There is no excuse to walk into a meeting with a bad presentation anymore. You basically get a McKinsey-level output in a minute or two

   4. PHPawel Huryn0:19

      Cowork is not adjusted to working with code base. As a product manager, you will be working with code bases a lot.

   5. AGAakash Gupta0:25

      Pawel Huryn is the number one AI PM voice in Europe with over 200,000 LinkedIn followers and over 100,000 newsletter subscribers. And today, he's gonna break down everything you need to know to get the most out of Claude as a product manager. Pawel, you have spent more time in Cowork than almost anyone. You use it for a lot of your everyday day-to-day tasks, even though you are a former engineer who could use Terminal just fine. So can you walk us through what a power user, what a master setup looks like in Cowork, and how you use it?

   6. PHPawel Huryn0:53

      In Cowork,

2. 0:54 – 3:08

   Cowork Power User Demo

   1. PHPawel Huryn0:54

      we can organize work in projects or folders. So what you can see here, most people stay in chat forever. That's like using Photoshop only to crop photos. And today, we are going to discuss Chat, which is part of the cloud desktop, but also Cowork, Cloud Code, Dispatch, which allows you to control remote sessions for Cowork and Code, Web Sessions, and how to use them all to 10X your productivity.

   2. AGAakash Gupta1:18

      Before we go any further, do me a favor and check that you are subscribed on YouTube and following on Apple and Spotify podcasts. And if you wanna get access to amazing AI tools, check out my bundle, where if you become an annual subscriber to my newsletter, you get a full year free of the paid plans of Mobbin, Arize, Relay app, Dovetail, Linear, Magic Patterns, DeepSky, Reforge Build, Descript, and Speechify. So be sure to check that out at bundle.aakashg.com, and now into today's episode. [fire crackling] So Pawel is the guest behind our most popular episode ever, with over 65,000 views when he gave an impromptu complete course on AI product management. He wrote- did the n8n video, which absolutely crushed with over 10,000 views. Before that, he did a discovery masterclass. Today, he is back for a record fourth time to help you understand how to get the most out of your Claude Enterprise subscription, when to use Claude on the web, when to use Claude Cowork, when to use Claude Code, when to use Claude Dispatch. Pawel has tried all of the combinations of PM task versus tool, and he is gonna break down when you should use which tool for what. And he is going to break down how you can create a self-improving system that builds upon itself, so that every time you do something, the system gets better at doing that task. Pawel, welcome back to the podcast.

   3. PHPawel Huryn2:47

      Hi, Aakash. Thanks for having me.

   4. AGAakash Gupta2:51

      I love getting your newsletter in my inbox. One of the most interesting things you've been writing about recently is how Anthropic has gone from a billion to 30 billion in 16 months on the back of this insane shipping velocity. You tracked 74

3. 3:08 – 6:07

   Anthropic's Shipping Velocity

   1. AGAakash Gupta3:08

      releases in 52 days, which is pretty mind-blowing. Putting aside what's happening at Anthropic, what I want you to help us understand is, what does this velocity say about the future direction of product management?

   2. PHPawel Huryn3:23

      So what, what I can say by observing Anthropic is that, uh, they are adjusting, and many companies, not just Anthropic, are adjusting their workflows to AI, so they do not use AI to replace steps in their processes, but to redesign their processes around what is currently possible. What I can see [chuckles] in many companies is that those roles, uh, so product manager, product marketing manager, designer, and engineer are maybe not merging, but coming closer together. And many of the things that required a separate role, like maybe prototyping, maybe, uh, testing, uh, writing release notes, even des- designing an interface based on the design system, this is all being automated. And for a PM to be-- to thrive in this [chuckles] new environment and in, in those new workflows, product managers must understand technology. They must get comfortable with, uh, tools that previously were designed for engineers traditionally, so like the terminal. Uh, they also need to go outside their comfort zone and understand strategy, understand how their work, uh, drives revenue for the business, how it connects to, to business goals, product strategy, uh, understand how, uh, it translates to, to, to revenue. The future is super PM [chuckles] or super individual contributor, uh, with maybe a PM focus, maybe engineering focus, uh, but having skills from multiple areas, not just one. You cannot be a product manager who only interviews customers and creates items in the product backlog.

   3. AGAakash Gupta5:08

      Yeah. The nature of the role is shifting, but for my money, this is the most fun version [chuckles] of product management yet.

   4. PHPawel Huryn5:14

      Yeah. Yeah. Absolutely. [chuckles]

   5. AGAakash Gupta5:16

      In the 16-plus years I've been in this field, I've seen it evolve a lot, which is one of its constants, is that it's always evolving more than the average profession. And I think this version of it is the closest to the bare metal of the product, which I think is incredibly exciting. So for this future model of product management, I wanna break down for everybody who watches this episode to the end how to embrace it, how to actually redesign their processes from the ground up like Anthropic, how to use the right tooling from Anthropic [chuckles] to behave like a next-generation AI product manager. So that's the goal for this episode today we're gonna break down for folks. You and I spend a good amount of time chatting on DMs and WhatsApp, and one of the craziest things you told meWas

4. 6:07 – 7:48

   Why Stop Using Chat

   1. AGAakash Gupta6:07

      that there is no good reason to use chat anymore. That is, there's no good reason to talk to Claude just over a normal web chat anymore. Can you break this down? Because this is gonna surprise a lot of product managers listening.

   2. PHPawel Huryn6:20

      Yeah. Like I, I personally, I would lie if I told you that I don't use chat at all. Uh, I use it sporadically, just if the window is open, I can ask some question like, "Is this grammatically correct?" or, or something. But mo-most of the time when starting a session, you don't know what exactly you will need, and when you start a session in chat, there are certain restrictions that you will face sooner or later, like you cannot continue. So let's imagine you have started your work, you are in the middle of the work, and you have to need, uh, leave your desktop. You cannot continue. You can copy the entire chat history to another window and maybe start a remote session, but you cannot continue in this, in this window where, where the conversations happened. Uh, you cannot continue on mobile. Um, there are restrictions related to what if you decide that now I want to code something, uh, chat cannot, cannot do that. And you want to create a HTML page, then export this HTML page to infographic, and you need this infographic in your email. Uh, so once again, you need to start a different session with a different context and explain what you have been discussing with chat. So for me, it's easier just to start everything either in with Cowork, with Dispatch, or with Claude Code, and then you can easily navigate between those interfaces. Does it make sense, Aakash?

   3. AGAakash Gupta7:45

      [sighs] Yeah, it does. So I've been

5. 7:48 – 9:56

   Ads

   1. AGAakash Gupta7:48

      building a lot of AI products lately. My job search OS has 16 different agents. My newsletter has a recommendation engine, and I kept running into the same problem. I'd ship something, it would work in my testing, and then I'd get messages from users saying it's hallucinating or picking the wrong tool. The issue wasn't the prompts or the tools, it was that I wasn't actually evaluating anything. I didn't have a way to see what my agent was actually doing step by step. Every tool call, every decision. That's where Arize comes in. Let me show you. I'm going to open Claude Code and install Arize with just one command, npx skills add arizeai, arize skill skill yes. Now Claude Code already knows how to instrument my agent. I tell it, "Set up tracing to Arize," and it automatically analyzes my code base, figures out where the LLM and tool calls are, and adds instrumentation automatically. Now I can see everything, every trace, every span, every decision, and more importantly, I can evaluate it. That's the shift. Trace what's happening, evaluate where it fails, then fix it. This trace right here, my resume feedback agent was supposed to pull the company's tech stack from the job posting, but instead it hallucinated that they use React when the posting said Python. But instead, it hallucinated that they use React when the posting said Python. I never would have caught that without seeing the trace. And here's the part that blew my mind. I asked Claude Code to look at these traces and tell me what I should be evaluating. It came back with four eval criteria I hadn't written, things like picking the right tool and staying grounded in the input. I wrote the evals, ran them, and found that my agent was making the same kind of mistake about 12% of the time. Claude pushed a fix, I reran the evals, and it dropped to under 2%. That whole loop, trace, evaluate, fix, took me about 20 minutes, and now it runs automatically. If you're building AI products and not evaluating them, you're shipping blind. Try Arize free at arize.com and get a year free, a $1,260 value with my bundle. Arize. Check it out. It's one of the top AI evals platforms used by all of the top AI teams for a reason.

6. 9:56 – 18:44

   Cowork vs Code vs Dispatch

   1. AGAakash Gupta9:56

      Let's make this more concrete, though. Most PMs I talk to, they're finally on like a Claude Pro or Max or API Enterprise subscription, and what they really need is a very clear mapping of if this, then that. You've mapped all of this out for us. Can you show us when should a product manager be using Cowork versus Dispatch versus Code?

   2. PHPawel Huryn10:17

      So chat is like, is like a chatbot, like ChatGPT, [chuckles] with various tools. So you type a question, it answers. Sometimes it can execute a simple script and, for example, re-reply with a spreadsheet. But other than that, uh, nothing complex. [chuckles] So it could draft an email, it can pretend to be some person, summarize information, very basic stuff. Uh, when it comes to Cowork, it's about working with real files and, uh, executing workflows. So by real files, I mean, uh, reorganize invoices on my desktop or create a HTML int- infographic. Uh, Cowork can also plan, uh, long-running tasks. So for example, if completing a task requires taking five, six, six, seven or more steps, it can do that, and it can then execute those steps one by one. It can also spawn, uh, sub-agents, so some tasks can be executed in parallel. So for example, you want to write an email, you want to summarize your product strategy. Inside this email, you also want to add some presentation as an attachment, but the presentation should be converted to PDF. So it can, uh, start one agent to, to summarize strategy or whatever, uh, another agent to create a HTML infographic and then maybe convert it to PDF, and then another one to do something else. Maybe, uh, find personal details in your, mm, in some CRM system. And then all those agent, agents [chuckles] finish the work. It can get the results, maybe reply some steps or adjust and, yeah, finish the work. So those are real, real workflows, real, real agentic workflows with the access to real files on your desktop.And Code is, is the same, but it's, uh, coding, so except, o-other than connecting to all those systems and working with real files, uh, it can execute scripts on your real machine because Cowork runs in a virtual machine, and Code can execute, uh, scripts, can execute system commands on your laptop. Uh, and it can also-- It is also adjusted to working with codebases. So it means, uh, it has different set of plugins, not adjusted to, uh, knowledge work, but adjusted to designing front-end, working with databases, debugging, and so on.

   3. AGAakash Gupta12:57

      A PM who's non-technical who's watching this is probably thinking code is scary, the IDE is scary. Can I just skip it?

   4. PHPawel Huryn13:05

      No, you can't, because engineers are going to use code. [chuckles] And even though many features i- wh- when talking about personal productivity, like working with files, uh, analyzing information, even organizing a self-improving knowledge database, you can do that in Cowork. Uh, you will not have the explorer view, which is about presenting the hierarchy of folders and files, but other than that, uh, Cowork can do almost everything. Uh, but there are certain features in Claude that only Cowork has. So for example, uh, Code has, so for example, one, one example can be sub-agents that you define in your solution. Cowork cannot do that. It can call dynamic sub-agents, but you, you can't have those agentic personas that you... So for example, researcher, for example, uh, an agent that tests your solution, an agent that, uh, creates release notes. You, you cannot have this structure in Cowork. And similarly, there are also other features that are like, uh, code-specific or working with codebases specific like, uh, hooks that do not work in Cowork. So eventually, as a product manager working with engineers, you will have to work with code.

   5. AGAakash Gupta14:29

      I can't emphasize this enough. As a product manager, you should be learning Claude Code. You need to get over the initial sort of hump of this doesn't look great. But Clawwork, Cowork is super powerful, and so Pawel, you have spent more time in Cowork than almost anyone. You use it for a lot of your everyday day-to-day tasks, even though you are a former engineer who could use Terminal just fine. So can you walk us through what a power user, what a master setup looks like in Cowork and how you use it?

   6. PHPawel Huryn15:01

      Mm-hmm. Okay. So in Cowork, w- so first thing to understand that, uh, we can organize work in projects or folders. Uh, so what you can see here, here are some of the recent folders that I have been worked in, working in, and also predefined projects that can have predef- custom instructions. So one of those projects is Editor, which is my project that helps me with copywriting, research, generating infographics, um, um, and other things related to content creation. Uh, but, uh, maybe we can first use it without it. So I will just select some random folder on my desktop. It's not entirely random because I pres- prepared it before.

   7. AGAakash Gupta15:54

      [chuckles]

   8. PHPawel Huryn15:55

      [chuckles] So, okay. So here is the folder with invoices. Uh, and yeah, after granting permissions, Cowork can now access real files, uh, in this folder. And the files inside are just random invoices that I collected [chuckles] from two months. So like At-Atlassian invoices. I'm not sure if preview is relevant, but yeah, like different invoices, uh, different, uh, receipts and so on. Uh, so what I can now ask it to do is to organize, analyze PDF invoices in, in this folder and group them by month in folders like, uh, Jan, Jan, Mar, June, et cetera. Uh, yeah. The duplicate, the duplicate files. Some of the files have duplicates, so which it should figure out how to do that. [chuckles] The names might be different, but the content is the same.

   9. AGAakash Gupta17:26

      And so now you guys are seeing immediately why Cowork pretty much always is preferable to Chat. In this case, it's actually manipulating things in your file system.

   10. PHPawel Huryn17:35

       Yeah. And, uh, as, uh, we can see, it created a list of steps, so it needs to extract dates from PDF invoices. It needs to identify and remove duplicates. Maybe it will use a hash function. It needs to create month folders, move files, and yeah, verify that if everything is correct. Uh, what we can see is also the context. So context for this, um, because this is an agent, the context for this agent is, uh, this file CLAUDE.md instructions, um-There are no special instructions inside, but if there-- if this file existed, that would be this Claude.md that everyone is talking about with like custom instructions when working in this folder. Uh, so the next time I could prepare an inbox and just drag and drop new invoices, uh, inside, and every time a, a new file appears, uh, this process is repeated, and the file is sent to the right subfolder. And it also

7. 18:44 – 25:03

   Skills and MCP Connectors

   1. PHPawel Huryn18:44

      dynamically loaded skills, which are like procedures. Uh, skills are activated based on the task that, uh, the agent currently is executing. So skill has a description, and based on this description, an agent can decide that, yeah, this skill is about working with PDFs. I'm working with PDFs, so let's see what is inside, and then it will, uh, read detailed instructions, detailed procedures of, of how to work with PDF files. And, and we call this progressive disclosure. So you can have dozens or, or hundreds maybe, uh, of skills, and Claude will read them. So Cla-Claude will read the detailed instructions only when the skill description matches what you are trying to do. So in this case, only one PDF was identified. Uh, what it did, it created four folders. We can see it. Uh, it removed some duplicates, and the names suggest that, suggest that indeed those were duplicates. Le-let's see the folder. Okay, April, February, Jan, March. If I open January, then I have some invoice from January. It's-- this is not a secret. Maybe this, this one is, so let's try a different one. Um, yeah, like postmark is, it is not a secret. I can present it. So it's February. Is that correct? Yeah, it's, it's an invoice from February. And that's it. And I asked about PDF, but you can also see that, uh, it also processed images.

   2. AGAakash Gupta20:35

      Nice.

   3. PHPawel Huryn20:35

      So yeah, and this is some fuel invoice. Uh, let me verify that it is correct. It's in Polish, but yeah, it's February. I hope you can see it.

   4. AGAakash Gupta20:47

      Yes.

   5. PHPawel Huryn20:48

      So it, it just ident-- it just understood that maybe Pawel didn't know that there are images inside. Let's, let's move them to this folder too. Uh, but-

   6. AGAakash Gupta20:58

      Pretty powerful. Not Gemini is not the-

   7. PHPawel Huryn21:01

      Yeah

   8. AGAakash Gupta21:01

      ... only thing that can read images these days. Claude can too.

   9. PHPawel Huryn21:04

      Yeah, no problem. What else it can do? Um, so it can work with files, so we have presented that. It can load skills, which are instructions, but skill can also have, um, scripts to be executed inside. Not necessarily, but yeah, Cowork can also execute them, uh, inside a virtual machine. Uh, it can connect to external and local services, and the most popular, uh, format is MCP server, so it's like the USB for, for agents. And by MCP servers, I mean in Claude they are called, called connectors. So we have connector for Google's, for Google Drive, for Gmail, for Slack, and for dozens of other apps or hundreds of other apps. Uh, some of those connectors of MC-- older MCP servers are provided by Anthropic and others. You can just download them and configure locally and yeah, talk to your apps. So for example, I can ask, let, let me demonstrate. Um, we are still in this folder. It doesn't matter, but, uh, how many unanswered emails I have? Uh, write code.

   10. AGAakash Gupta22:28

       Oh, this is a go. [laughs] I don't want to run this on my inbox live on the podcast.

   11. PHPawel Huryn22:34

       Count my category. Do not reveal any personal information or emails. Yeah, let's, let's just see. Uh, uh, so it connected to my Gmail account. It can also draft emails, and it can... This one cannot send. It can only create the drafts, but you can also connect a, a connector that can, can send emails for you.

   12. AGAakash Gupta23:05

       How do you process your email inbox today? Do you recommend most PMs do it with Claude?

   13. PHPawel Huryn23:10

       Yeah. I create all the responses. Yeah, this is my current configuration. So yeah, Claude, uh, the system that learns drafts replies, um, but it doesn't s-send them automatically. So I'm using this default connector that cannot send. And similarly on Slack, theoretically, you can, um, ask Claude to respond automatically. If you use this connector, uh, there is a footer sent by Claude, so everyone can see it. Uh, instead, I ask it to draft replies, and I verify them. Uh, in, in many cases, someone asks about the URL, so no reason to find for this, yeah, to, to spend time looking for this. And, uh, I just approve the me-messages or edit them. I send it manually, so there is a button in the interface, Send. And after every session, Cowork or Code, depending co- on what I, what interface I'm using, uh, verifies my responses and tries to learn from them, so the next time it will get better.

   14. AGAakash Gupta24:20

       Love it.

   15. PHPawel Huryn24:22

       Uh, yeah, and that's basically, that's basically it. You-- So, so a bi- a big thing [chuckles] are those skills that you can, you can either use, uh, predefined skills or skills from third-party marketplaces like, like my micro- marketplace. Uh, that's basically it when it comes to Cowork. So-

   16. AGAakash Gupta24:48

       Can you show us the PM skills marketplace? This hit 1,300 GitHub stars in 72 hours, so it's gone pretty viral.

   17. PHPawel Huryn24:56

       Yes, of course. So it's this one. It's, uh, P-H-U-R-Y-N, pm-skills/pm-skills.

8. 25:03 – 29:06

   PM Skills Marketplace

   1. PHPawel Huryn25:04

      It's currently 10,000 GitHub stars, and what I have done is that I, I, I created a set of, uh, plugins. So a plugin is like a collection of skills and commands, and for different domains like data analytics, execution, go-to-market, market research, and you can load each of those plugins separately. Next, inside each plugin, you have skills. So for example, if we open, let's say, product discovery, inside there are... Ah, there is a documentation. So [chuckles] there are analyze feature request, uh, brainstorm ideas, plan experiments, mm, create metrics to track, um, your feature, and so on. And also I have defined, uh, workflows that aggregate more than one skill. So for example, product discovery, this can be, um... Okay, this is a better example. So discover, it is from ideation, then we map assumptions, then we think about, uh, we think about, uh, uh... So we analyze... Actually, this description is not correct. So it analyzes [chuckles] customer needs, then based on those needs, it, uh, will, uh, map the opportunities and how important, uh, uh, certain problems are for the customers and how satisfied they are with what they already have. Then it will ideate, so how we can solve those problems, uh, map the assumptions related to value, usability, feasibility, viability, maybe some ethical considerations, and it will pra- plan experiments to prove or disprove our assumptions. Uh, so this is like the entire product, [chuckles] product discovery workflow in one command. Uh, how to use it? You can go to the homepage, and there is a, a script, but basically you can just take this URL, go to Claude, uh, open Customize. Let me see. Personal plugins. Add Marketplace. Yeah, like, uh, yeah, this interface is really changing all the time. So, uh, it's, uh, let's see this one. But I, I think that owner/repo doesn't work despite the description, so let's, let's try this one. That's my gut feeling that this, this can work. Okay. And, uh, those are not mine. So yeah, we can see it here. [chuckles] So pro- data analytics, execution, go-to-market, market research, and so on. And if I click product discovery or product strategy, yeah, maybe this one. Now it is added, installed, and ready to use. What is inside? Analyze. Ah. Strategy. We have Ansoff Matrix, pricing strategy. So now when I go to, to the chat, uh, let's start a new task in Cowork. Help me design. And theoretically, I can count on, on skill being activated automatically. But yeah, if you want really, uh, be sure that the skill is activated, uh, yeah, product strategy.

   2. AGAakash Gupta29:03

      Use the slash command, which I do recommend. [chuckles]

9. 29:06 – 35:14

   Strategy Canvas Demo

   1. PHPawel Huryn29:06

      Yeah, it's, it's kind of messy, and sometimes if Claude has general knowledge about a certain area like, uh, yeah, product strategy, it thinks it knows, uh, something, but you want to override this knowledge, and without doing it explicitly, it can default to, to the training data. So yeah, i- if you know that there is for Amazon 2.0, whatever it means. So it loaded the skill, and now it will interview me to get more information, but it is, it is all part of the skill. So what is the core concept? Output format, and then it will create product, product strategy canvas, by the way, based on my product strategy canvas. [chuckles] When you type in Google Product Strategy Canvas, it will be my, uh, the one I created.

   2. AGAakash Gupta30:02

      Nice.

   3. PHPawel Huryn30:03

      So it looks like this, and, uh, of course, there are those, uh, famous plugins that Anthropic created. A GitHub repo with, uh, plugins and skills. Uh, the main plugin was le- legal. Uh, hundreds of millions of dollars evaporated from, uh, from the stock market.That's because people understood how easy it is to, to describe certain processes and automate them.

   4. AGAakash Gupta30:30

      Yeah. With a proper markdown file written by a subject matter expert, these days Opus 4.5 and Opus 4.6, they can really use the tools, whether it's PowerPoint or Excel or for PMs Notion or Google Docs, and they can operate it like a competent person if they have the right instructions. It's pretty crazy. So can you walk us through a couple of these skill files and how you constructed them and how somebody constructs a good con skill file?

   5. PHPawel Huryn30:59

      Yeah, sure. Uh, so the skill is basically how to do something. [chuckles] So some procedure or a domain knowledge about doing something. Uh, so you, you don't really have to create those skills manually. You can describe this process to Cowork or Claude Code and ask it to create a skill. [chuckles] So, so basically if you can teach a graduate how to do something, just repeat the same to Claude and it will create a skill. That's, that's the easiest, the easiest way. We can of course look inside and, uh, I usually don't do that, [chuckles] so I just use the chat interface to, to talk to Claude. But yeah, uh, it's not this repo, it's this one. So if we open, for example, product discovery skill, uh, identify assumptions for an existing product, we have this skill MD file. Mm, and you, you can also have additional files in this folder, but yeah, basically the main file is skill.md and, and the format is markdown. Uh, we can see the preview in GitHub, but if I switch to the code view, there is this intros. Mm, there is this section that, uh, agents read, so they do not load the entire skill. They load, um, name and description and when to use the skill, what it is for. And as I said, when doing a specific work like, uh, use when stress testing a future idea. So if the user will s- write in the chat that, "Hey, I want to test this idea and, uh, assess the risk," the agent will see the skill, it will see the description. It, it will just, um, trigger it. It will use it. Uh, other than that, this is just a prompt, so it can have instructions, it can be, uh, general knowledge without step-by-step, but it can also be like step one, uh, get information from place A. Step B is format it in a specific way. Step, step three is something. So it's, it's just a prompt. So in this case it is, uh, yeah, it's the context. So you are a devil's advocate. We are stress testing idea, and there are those arguments. So you-- we have seen that Claude, uh, when working on strategy, asked me a few questions, so those are the arguments. And then instructions, what it should do step by step. And that's all. And, [chuckles] and by the way, this is, uh, if someone wants to learn more, they can, uh, either they can see it in as a markdown file or if they keep talking to an agent, uh, the agent will suggest my articles about identifying assumptions. So this is like, like [chuckles] marketing inside skills, but relevant to what the person was trying to do. So they ha- all the knowledge is inside skill, but then the agent will also have those articles, uh, in a short-term memory.

   6. AGAakash Gupta34:36

      I would tell everybody, like iterating on your skills is one of the highest ROI activities I personally have done. So take Pawel's skills as a baseline, so at least you have a baseline. But then as you encounter some feedback for his skill, give that feedback to Claude and say, "I want you to improve my assumption existing skill. Read our chat and see the feedback I gave you. Understand the root cause of what drove you to give a output that I had to give feedback, and rewrite the skill from first principles so that it doesn't

10. 35:14 – 40:46

    Skill Iteration Cycle

    1. AGAakash Gupta35:14

       make that mistake again." I have found this to be the single highest ROI activity I do. So I think once you get the initial skill, you've got to really iterate on it to really get the ROI.

    2. PHPawel Huryn35:25

       Like I, I agree, Aakash. I, I don't want to memorize prompts. Um, it's like in evals, and you also have been writing about evals, that you need to-- when building some AI system or s- or some AI pipeline, you need to see how the system performs in real life and then identify failure modes. And in this case, you, you don't, you don't create evals, but you can, uh, just give feedback to, to Claude, uh, say what was wrong. It, it will understand the context because it already has this context. And, uh, yeah, what, what, what were your expectations? It, it will fix that. And you, you test it again, test it again, and eventually you will eliminate like maybe 99% of the, of the failures. So that's, that's the, [chuckles] the only way. And, uh, you, you cannot just see it and, yeah, use some magic, magic technique to, to get it right on the first try. It doesn't work like that.

    3. AGAakash Gupta36:23

       So Pawel, now that people understand about self-improving skills, can you show us what our skill developed?

    4. PHPawel Huryn36:28

       So we already presented. I just asked to design product strategy, and it loaded this built-in, uh... So it loaded two skills. One skill was, uh, designing presentations, and we should see it, uh, yeah, PPTX. So this is a skill by Anthropic, and another one was my skill about product strategy from my plugin. And it used it to, to create, uh, this slide deck. We can see it directly in Cowork. Um, okay, mm, I can see it in Google Drive or display it here, directly here. So we have Product Strategy Canvas for Amazon. We have a product vision. Market segments like conscious customers, independent brands, um, relative costs. So, [chuckles] uh, it adjusted colors. Uh, it, uh, figure out what the layout should be. This is not my branding. It's like, uh, Cowork's invention. [chuckles] Also icons and, yeah, the, the difference between orange and green. This is interesting. So buyers have this, whatever it means. [laughs] I would have to think about.

    5. AGAakash Gupta37:55

       [laughs]

    6. PHPawel Huryn37:56

       But it, it looks, it looks [chuckles] like something smart. Trade-offs, so what we are not going to do. So we have this focus on our strategy and, um, the things that we are-- we want to say no to. So, uh, makes sense. Key metrics, so how we are going to track that our strategy is working, like NPS, order value, seller retention. Seems to make sense. Uh, North Star, a star icon. [chuckles] Guideline, it, it even suggested guideline metrics. So, uh, when we focus on, on our North, North Star, how-- what should we monitor to make sure that our, our other areas do not degrade? Um, and [chuckles] carbon neutral de-delivery rate. Uh, this can have an effect on, yeah, I'm not sure, uh, like public relations. Our growth strategy, why, uh, in different market segments and unit economics, like this is pretty advanced, and [chuckles] it's not a single layout. It's not just, uh... Those are not just tables. It's, uh, yeah, diverse layouts, diverse icons, diverse, uh... And it makes sense. Uh, if I ask a graduate to do that, like I am not sure I would get that in, in a few hours, hours. [chuckles]

    7. AGAakash Gupta39:17

       Yeah. So there's two really mind-blowing insights for everybody here. One, Claude is way better at using PowerPoint than it was a month or two ago. And so you need to look at this output and see, like it can use PowerPoint to create great presentations. There's no excuse to walk into a meeting with a bad presentation anymore. And the second is that it used this skill, specifically some of the things Pawel had defined in this skill around have a North Star metric, have guardrails, and it has implemented those. And so that's why if you have a good skill, you can now create things. This is why people say RIP McKinsey, right? You basically get a McKinsey-level output in a minute or two.

    8. PHPawel Huryn39:56

       Yeah. And for Cowork I, I have defined 60-something skills, but you can define hu- you can [chuckles] find hundreds of skills or thousands of skills defined by others. Uh, I would only recommend you to, to verify every skill that is not, uh, defined by Anthropic, uh, and make sure that it really fits your specific scenario and specific use case. But other than that, yeah, the knowledge is, [chuckles] uh, there are many free repositories. I can also share those links with Aakash, so.

    9. AGAakash Gupta40:26

       So we'll include those in our newsletters when we talk about this podcast so that you guys get all of those. Be sure to subscribe to both. That's the Cowork segment. So you guys just got the wow moment. Cowork can generate amazing presentations for you. We showed you how Cowork can j- organize your file system for you. Now talk to us about Claude Code, Pawel.

11. 40:46 – 44:43

    Why PMs Need Code

    1. AGAakash Gupta40:46

       Why does a PM even need P-Co-Claude Code? At this point, they think, "Oh, Cowork is good enough."

    2. PHPawel Huryn40:51

       Cowork is, uh, is not adjusted to working with codebases. And as a product manager, you will be working with codebases a lot. And also, if you are building complex systems, mm, that involve multiple files, then, um, this view that, that what we see in Cowork is, is not adjusted to it. So for example, if I want to find this, this presentation, I can click Show in Folder, and it is somewhere. Pro-probably in our... Yeah, we've been working with this invoices folder, so it's placed it there.

    3. AGAakash Gupta41:30

       [chuckles]

    4. PHPawel Huryn41:30

       Uh, but imagine I have, uh, like 100 files or a few hundred files, like i- different invoices, different contractors, um, contracts, uh, maybe some, some article drafts, some marketing strategies, uh, pictures, uh, brand guide, and so on. So those files will be organized in some hierarchy, and y-you cannot browse them from here. You can just drag and drop or select single files, like just by, by browsing your, your lo-local folder or browsing Google Drive. Mm, but there is no way to, uh, like easily work with, with those large codebases. One example, but also a, a list of contractors, a list of invoices, like a, a lot of graphical files, some templates, um-You need to upload them, find those files, upload them here, and once the agent delivers something, you need to find that in the folder or, or save it to the folder. [chuckles] So even though this is my real folder, I still need to find this file, uh, to do something with it. Uh, so especially in case of, uh, code bases, uh, you need this, this view in which you will see folders. You can expand folders, and you can see what is inside. So for example, I have infographics folder, and inside infographics folder I have, um, Cloud Code pricing, and inside Cloud Code pricing I have some files generated by Claude. So let's see it in, um... I will just view it here. So this is not the best picture that it generated. [chuckles] Uh, yeah, but this was published today in my newsletter and, uh, also generated by Claude. Uh, another one can be this picture with calendar that went viral. This was also, uh, generated by Claude, not by me. So this one. And this was-

    5. AGAakash Gupta44:01

       This one is so good. So a lot of people, they struggle with getting faces and images and logos. So you've gotten the Anthropic logo, and you've gotten those eight people's faces.

    6. PHPawel Huryn44:14

       Mm-hmm.

    7. AGAakash Gupta44:14

       Also, you've actually gotten the data from Twitter. So can you show us all three components of how you did that?

    8. PHPawel Huryn44:20

       Mm, yeah. I f- I think I need to explain how my system works and how it is organized [chuckles] to do that. Uh, otherwise it will be... I can repeat the process, but, but to explain how it works, I need to just explain the system.

    9. AGAakash Gupta44:32

       Yep.

    10. PHPawel Huryn44:33

        Um, so recently Karpathy presented this, um, system in which you

12. 44:43 – 56:00

    Building a Second Brain

    1. PHPawel Huryn44:43

       use LLMs to build a personal wiki or knowledge base for humans. So you upload some information, give them, give random articles and random attachments to, to an agent documents, and agent organize them, and then you can browse those files, browse those information and how different, uh, facts are connected. Uh, I've been doing it since February 2026, and, uh, instead of building the second brain for myself, I started building second brain for my agents. Uh, [chuckles] so, so, uh, I'm the curator of the information. Um, and what I do is I send articles, I send, uh, infographics that I find in, on social media. I can also ask, "Hey, let's analyze the last 10 posts by Aakash, um, above 200 reactions. What... Why they worked? Voice, um, hooks, um, emotions." Hmm.

    2. AGAakash Gupta46:14

       And a lot of people-

    3. PHPawel Huryn46:15

       Not-

    4. AGAakash Gupta46:15

       ... struggle with editing their terminal prompts. You were using the-

    5. PHPawel Huryn46:18

       Yeah

    6. AGAakash Gupta46:18

       ... AI tools to do that, right?

    7. PHPawel Huryn46:20

       Yeah, yeah. You, you... There is the second, uh, way to use Claude Code. I, I'm using CLI inside Visual Studio Code, but you can also use Visual Studio extension. Uh, they are both similar and [chuckles] they are not super user-friendly.

    8. AGAakash Gupta46:37

       Okay.

    9. PHPawel Huryn46:37

       Uh, okay. So yeah, in February I s- started, mm, making screenshots on social media, and I just-- I was giving, uh, different types of information to my agents. So back then it was Cowork. And, uh, I asked, [chuckles] "What made this post or what made this infographic work?" And then the agent, uh, uh, Cowork replied with, with some information. In some cases, it knew the answer, so I was able to, to note it. In, in other cases, it res- replied with a hypothesis. "So this worked probably because something." Uh, I decided that instead of noting this myself, I will ask an agent, "Hey, build a knowledge database, and can you please, every time I give you some article or infographic, organize it by domain." So in this case, the domain is social media, so this can be X, it, it can be LinkedIn, it can be Substack. "And then, uh, write the rules for which you have a lot of information. Um, if you see the repeating patterns, save this pattern as a rule. And if you are not sure, uh, save it as a hypothesis that you can later prove by analyzing more data." So what, uh, we have ended up with is this knowledge database, and I have, um, yeah, for example, sound bites. So what creators use to, uh, yeah, core patterns, uh, that work across pa- platforms based on data. Uh, what are the core techniques like interpretation layer, uh, uncomfortable clause, credibi- credibility before claim, and this is confirmed across multiple, many platform, all platforms and all creators and all successful posts.How to-- what is the word choice? [chuckles] Like, uh, uh, list of hypotheses across platform specific to, to index rejected. So also things that we have been considering in the past, but now we know they are not correct. Like let's say on cross-platform, this is, mm, yeah, achievement as, as proof hooks outperform achievement as point hooks or [chuckles] emotional diversification correlates with higher average engagement. Those are not, uh, hypotheses that I formulated. I have not even seen them. I'm seeing it for the first time, this one. It's the numbers 46, and it's like not even on the file. Uh, so it, it extracts this information and, uh, yeah, every time it analyzes more posts, more graphics, it tries to confirm or, or reject those hypotheses. And, uh, yeah, at the end of the day, uh, I have information how to write, uh, the files beliefs for X. What are the hooks that work on X? What are the accounts that we monitor? What are the rules? [chuckles] Hard rules like, um, probably more technical ones, uh, and some templates that we know we have tested, and we know that this template will work. Uh, this doesn't mean that it writes for me. Um, I previously demonstrated in the newsletter that, uh, like the conversation that, uh, I suggest ideas, I suggest takes, so like raw knowledge and raw opinion, what I think about the specific news. But then Claude adjusts the format, adjusts the style, adjusts the hook, um, to the information, to the platform, and yeah, what I'm trying to say to make it resonate with others. Uh, about your-- Okay. About your posts. I'm not sure why-

    10. AGAakash Gupta50:53

        So is the X API free that you've hooked into? Or how did you-- Are you paying for it? I've n- I haven't seen somebody connect into it before.

    11. PHPawel Huryn51:00

        Yeah, I am paying for X API. So in, in many cases, you can fetch posts for free. Uh, like it's this FX Twitter, something like that. I'm not sure why we are seeing errors. Usually, it is not the case. So this, um-

    12. AGAakash Gupta51:21

        Yep, I saw it. API FX Twitter.

    13. PHPawel Huryn51:23

        Yeah. So this one is free, but, uh, it, it is limited. It, it doesn't always work. So just to save, save costs, the agent tries to use it by default. Uh, if it doesn't work, it uses this custom tool that we developed together. Like I just gave it documentation of the Twitter API-

    14. AGAakash Gupta51:44

        Yeah

    15. PHPawel Huryn51:44

        ... and it created the tool for itself. So, uh, uh, it wraps this, this API to something that is easy to use.

    16. AGAakash Gupta51:52

        Nice.

    17. PHPawel Huryn51:52

        Uh, uh, okay. I'm not sure what, what, what... [chuckles] One formula if, um... Let me see what it told about you. Pure analyst voice, radical topic diversity. He doesn't need to stay in PM. He can surf whatever wave is biggest at, on any given day because the mechanism reveal pattern transferred to food science, neuroscience, physics. [chuckles] Uh, okay. That's the-- His top 10 has two tech posts and eight random curiosity posts. That's the opposite of Pavlo Lane strategy. The question isn't whether to copy some hypothesis. You already have it documented. The question is whether the mechanism reveal voice can work within your lenses. Uh, AI architecture, agent design, PM tooling at the same scale. Like we have many of those conversations with [chuckles] Claude Code and Cowork. Uh, yeah, ju-just to demonstrate, and similarly, I feed it with infographics, and it tries to, uh, extract some patterns that, uh, catch the attention or work in other creators. We select infographics that are easy to codify as HTML. [chuckles] And when they are easy to codify, then we design, we extract components that we can reuse, and then it uses the growing library of components to design infographics for me. Does it make sense? [chuckles]

    18. AGAakash Gupta53:33

        Yeah, it does. The last component is how do you personally-

    19. PHPawel Huryn53:36

        Yeah.

    20. AGAakash Gupta53:36

        Did it pass fetch the profile pictures of Boris and Tariq and-

    21. PHPawel Huryn53:41

        Uh, ah, this one. Yeah. Uh, but it was like a custom qu- custom query, like to... I'm not sure where it was. Um, no, it was ad hoc. Ad hoc, uh, maybe article, maybe something else. Let me see. Uh, calendar, no. Assets, no. Uh, maybe drafts. Hmm. No. It was like a temporary artifact. I asked it to create a script and go through, uh, first, uh, find. So I knew the free Anthropic accounts, so then I asked it to analyze their recent posts and reposts, and I assumed that they will eventually retweet every other, uh-

    22. AGAakash Gupta54:32

        [chuckles]

    23. PHPawel Huryn54:33

        ... Claude, Claude Code team member. Then it analyzed, uh, which team members used we or we just released or my team just released. Uh-If it was unsure, it verified the comments, so that, uh, yeah, just to make sure that the person was, was a, a Anthropic employee. [chuckles] And, and thennn, yeah, it went through like 15 people from Anthropic, uh, mapping for every feature, mapping who first wrote about it. Um, for, for those people, we got pictures from, also from Twitter, from X, and then, uh, we iterated several times on how to, how to visualize it. So that was the final result, but this is just, like, done by Claude Code. This, this particularly, this one was done by Claude Code, and this is, uh, HTML, uh, based on the research, and the, the, the most difficult part was the research. [chuckles] So just getting all this information from Twitter.

    24. AGAakash Gupta55:38

        Yes. It was difficult, but in the end, you were just, you were talking-

    25. PHPawel Huryn55:42

        Yeah

    26. AGAakash Gupta55:42

        ... to Claude Code over hours in natural language. So it's not impossible for other people to do.

    27. PHPawel Huryn55:49

        Yeah. Um, I do not code at all. Like, I don't write any code. I, I don't even review the code. Uh, if I want to know how something works, I just

13. 56:00 – 1:10:00

    Self-Improving Knowledge

    1. PHPawel Huryn56:00

       ask questions in the chat window.

    2. AGAakash Gupta56:03

       So how do you make this system self-improving, and what are the tips and tricks that people need to know around Claude MD files and folder structure to make their system really sing?

    3. PHPawel Huryn56:16

       Okay. So first, s- a lot of people, uh, so people are, uh, discussing Claude's .md file, that you can put your instructions there. But, but the problem with Claude MD is that, uh, if you put all the instructions inside, it, it will grow, maybe not exponentially, but it will keep growing and growing and growing, and e-eventually it will consume a lot of your, of your context window. And every time you ask a simple prompt in your project, all this Claude MD context will be included, right? It is, it is part of your prompt. Uh, so a smarter approach is to organize your knowledge in, uh, files dedicated to specific domains, and let me demonstrate mine. Uh, so I have Claude MD. So I have this main Claude MD. So the only goal of Claude MD is to explain what this project is about. It doesn't have detailed instructions or detailed information about, uh, like pr- good practices or bad practices of what to avoid, what to do more often. Uh, all this information is inside, uh, in other files, and the only way, uh, goal of Claude MD is to, to give those instructions, how to find the knowledge and what to do with the new knowledge. Uh, so project structure, this is just what we can see on the left side, so I can s- skip that. Like, there are tools, there are scripts, um, just so that the agent knows without scanning the, the repo. Uh, another one, uh, is where things live, so this is similar. Uh, and by the way, this was also created by Claude. I, I didn't wrote this, so I just, [chuckles] I just discuss in the chat how we organize things, and Claude writes instructions for itself. Uh, who I am, like basic information. Uh, and, uh, the knowledge system, this is the most important part. This is the self-organizing system where, uh, I have index of all files, like X files, LinkedIn files, Substack files, Substack notes files, uh, voice archetypes, craft. So those are-- this is gen- general knowledge across platforms. And then for every platform, where different, uh, elements live. And now, uh, there are also workflows that the system should follow. So how to fetch data for Twitter, how to f-

    4. AGAakash Gupta59:00

       Mm-hmm

    5. PHPawel Huryn59:00

       ... how to, uh, work with LinkedIn posts, how to link with, work with Substack. Uh, the most important part is when asked to study and analyze. So every time I give it some posts, like, "Let's study the last 100 posts," or, "I like this one. Uh, let's analyze what made it work." Uh, it knows what tools to use, so it will extract hook pattern structure, sound, soundbites, and engagement metrics. Uh, only if I request visual analysis, it will also analyze the attached graphic. This is to save tokens, so I, I don't analyze the graphics every time I analyze a post. Uh, check against existing patterns and false beliefs. And then if, if there is an existing hypothesis, it will update the existing hypothesis with, with the new ev- evidence. If there is an existing hypothesis, but we see that the specific post didn't work, we can demote the hypothesis, that the hypothesis is less likely, or it will, it will become rejected. Um, yeah, and it will append the post that was analyzed to the, to the database. So, uh, I think we may need to simplify this, [chuckles] uh, but basically it gets information about posts, and it tries to decompose this knowledge and organize this knowledge itself into rules, into hypothesis, and yeah, like other elements like structure, hooks, soundbites, uh, engagement metrics, and so on. So the system learns itself without me telling it, uhwhy the specific piece of content worked. Uh, the Claude, Claude does it itself. And the next time I have an idea and, or I express opinion about s- something that the other person tweeted, then I can say, "I like this, but I think this, this is connected to another idea." Or for example, I like Carpaty Post, but, uh, it will degrade over time. And [chuckles] by the way, uh, you don't need Obsidian where if the person... I- if the user is not a person, mm, because we are building a knowledge base for agents. And yeah, how can we re-retweet that? Or maybe I can add some additional details, and then Claude, it will suggest. It will use my ideas. It will, [chuckles] it will use my, my take, but it will format it in a way that resonates.

    6. AGAakash Gupta1:01:36

       So this is very content-focused. It-

    7. PHPawel Huryn1:01:39

       Yeah.

    8. AGAakash Gupta1:01:40

       A PM watching this might say, "This is kind of content-focused. What's in this for me? And I don't write code." So what is the minimum viable setup for a product manager and how they should be setting this up?

    9. PHPawel Huryn1:01:52

       Uh, self-improving system?

    10. AGAakash Gupta1:01:55

        Yeah.

    11. PHPawel Huryn1:01:55

        Mm, that's, uh, yeah, I have recently shared that, and let me demonstrate it. So I have actually, uh, I have a poster for this. So how to create a knowledge system that learns itself, and the prompt is very simple, so you don't have to build everything I did. You can start with just this prompt. [chuckles] So before starting a new task, re- uh, review, um, existing rules and hypotheses for this domain. Then apply rules by default. So for example, if, if we know that, uh... So basically, this is the most important part that you need to paste to your Claude MD that, uh, and this is not content-specific, so that whatever Claude does something in, in a specific domain, like testing software, like, uh, writing, uh, marketing materials, maybe writing release notes, it should review, uh, the rules and hypotheses from this domain, and it should apply the rules that were confirmed to its work. So for example, how a good test case, how good, uh, user story is formatted, or how, how the, uh, acceptance criteria should be written, or, uh, what are the good examples of, of release notes, uh, or customer offer, whatever you do. [chuckles] What are the good examples of something, and what are the bad examples of something? And it, it is, it will try to extract the rules. And then when you ask it to perform a task, like, "Hey, you saw the 10 good examples to, to bad examples, then let's try to, to create, uh, another offer for, for this new customer," it will review the existing rules and hypotheses because you, you didn't write it. Uh, it reasoned what are the good rules and bad rules, or what are the hypotheses from the data that it, it has seen, and it will also keep learning. So every time you give it new information, it will update its knowledge. It will update the hypothesis. It can also ask you questions if this is something that where you can give agent the feedback. Um, and it basically, it keeps learning, keeps, keeps, uh, adding knowledge, and the knowledge is organized by the domain, so pricing, so marketing, um, testing, quality, strategy.

    12. AGAakash Gupta1:04:28

        Yeah. So you wanna create that knowledge with an index.md that has a router, and in your Claude MD, you wanna give it this prompt-

    13. PHPawel Huryn1:04:35

        Yeah

    14. AGAakash Gupta1:04:35

        ... so that it's self-improving.

    15. PHPawel Huryn1:04:37

        It will figure out what to do, and every time it, it's-- every time you, you ask it to do something, it will use the existing knowledge that it, it is growing. And every time it s- it sees new information with the context, like when I, um, quote a tweet, the tweet has some metrics, like this tweet worked or didn't work, or how many people liked it. Uh, but, uh, maybe you can feed it with, with the offers that worked or with, um, yeah, su- c- uh, resumes of successful candidates, and it will start generating those rules. Uh, and, and the next time you will get a candidate, you can ask Claude, "Hey, uh, is this a good candidate?"

    16. AGAakash Gupta1:05:22

        Love it. So one of the things that you've written about that I haven't seen a lot of people write about, and maybe you can show us, is the Chrome MCP. When and how should we be using that?

    17. PHPawel Huryn1:05:31

        Uh, I don't use Chrome MCP anymore. Uh, [chuckles] uh, and the reason is that, um, I've been testing different approaches, and Chrome MCP, it's basically MCP that controls your browser. Uh, it works well, probably similarly to Claude in Chrome, which is, uh, Anthropic extension that you can use, like this guy here. You can ask to do something, or you can also call it from... You can schedule tasks here, or, uh, you can call to, to Cowork, and Cowork will, will call this, um, Claude in Chrome. Uh, the problem with those extensions is that they rely heavily on taking screenshots, and screenshots mean a lot of tokens. And if you have some tasks that you want to repeat regularly or complex processes, the, yeah, you can easily consume like $100 in an hour, especially with, uh, with Opus. Uh, so what I do instead is I use, let me check, but I think my agents, uh, right now use Agent DevSo what Agent Browser does by Vercel Labs, this is the most reliable according, like in my tests. Uh, it also uses the real browser, but it can do that in a headless mode, and it explains the structure of the page to the agent without presenting the entire HTML. So this is-- it is token-efficient. And, uh, so the agent doesn't have to see HTML. It doesn't have to interpret HTML, but can take actions. So it, it will see buttons with specific IDs, even if the original button didn't have ID. Uh, [chuckles] so, uh, so this is real browser, uh, real rendering. Uh, it can execute JavaScript if needed. Uh, it waits for rendering, which is important. So if this is a single page application or there are some additional resources, uh, that load a few seconds later, it will wait for it. Uh, and yeah, it presents this page to an agent without taking screenshots. So the agent can see the text, it can see the layout, it can see different components, the content. It can say, uh, "Hey, agent browser, click this button," but it doesn't have to interact with parse or interact with HTML itself. Uh, it is very simple, simple, uh, protocol. It, it's not an MCP, it's a CLI tool.

    18. AGAakash Gupta1:08:24

        Very cool.

    19. PHPawel Huryn1:08:25

        Uh, and-

    20. AGAakash Gupta1:08:26

        Use Agent Browser from Vercel instead of Chrome MCP. And what do you use it for exactly? Like, what is the use case that a PM should think about? This is when I'm gonna use Agent Browser.

    21. PHPawel Huryn1:08:34

        Like e-everything where you do not have an API to get data from external systems. So for example, if I want to get data from LinkedIn, of course, I can fire Chrome MCP or, uh, trigger a cloud in Chrome, but it will start taking screenshots every half a second or every second. And-

    22. AGAakash Gupta1:08:55

        [chuckles]

    23. PHPawel Huryn1:08:56

        ... uh, so it's like crazy token consumption, and, uh, it, it is also visible in most cases o-on your screen. So it's, uh, not something that, uh, [chuckles] you would like to see. And, uh, with this, I can just ask it to, "Hey, go to LinkedIn, check my, uh, inbox, or analyze the top last posts by Akash and see the comments." Or yeah, something like that. So, uh, maybe more enterprise use case will be accessing some legacy software without API or without MCP, uh, servers. Uh, maybe SAP, maybe some old CRM system. Uh, yeah, that, that the agent can access just, just by using a web browser.

    24. AGAakash Gupta1:09:47

        This is pretty epic. So as a PM, you need to hook your Claude code into absolutely everything, and you should use Agent Browser for the stuff where you don't have an MCP, CLI, API to hook into. Now I want to move into

14. 1:10:00 – 1:21:07

    Dispatch and Remote Work

    1. AGAakash Gupta1:10:00

       remote work. Anthroshop-- Anthropic shipped four remote surfaces in recent months: Web Sessions, Remote Control, Dispatch, and Channels. You use all four of these remote methods. Walk us through how you actually combine them for the same project and when a PM should use which.

    2. PHPawel Huryn1:10:19

       Okay. [chuckles] So the, hmm, one issue with Anthropic is that those surfaces overlap, and, uh, I, I don't use them in the same proportions. Uh, so I have tested Channels. Uh, I have abandoned Channels because I, I didn't see value like having this, uh, Telegram interface because I have clouds. Um, I have a dedicated clouds app where I can do everything. Okay. But starting with, maybe let's start with Dispatch. So, uh, Dispatch is, uh, a new tab that appeared in the desktop app and also on your phone, and this is exactly the same. So here, um, uh, okay. I'm not sure what it is. Uh, so Dispatch is like this single, single interface in which you can interact with, uh, cloud code and Cowork. It is displayed under, under Cowork. It, it is a bit confusing. It is a completely different product. It's like walkie-talkie. So, uh, it can start multiple background tasks, and, uh, every time the task is completed, it will report back what is the status. So for example, uh, create, create an infographic in Anthropic style for the following text. And let me find something, um, following text, and the text will be about bolt and windsurf and some, some... It, it's probably your work, so. Uh, in Anthropic style for the following text, and this will be, um, LinkedIn resolution. Uh, then I can ask it to, "Hey, how many emails, hmm, in the last two hours did I receive? No names, no personal data."So like, a-and it doesn't have to start the previous task. It will just start another one. And if we open the recents, I will show them, we should see that, uh, it delegates them to-- it should delegate them to, to other threads.

    3. AGAakash Gupta1:13:19

       Hmm.

    4. PHPawel Huryn1:13:19

       Uh, three emails in the last couple of hours, so maybe it did it directly. And let's go to Code. Ah, analyze Akash po- Akash posts. No-normally, it, uh, it, it would delegate those tasks to sub-agents, and you will see them as dispatched tasks in the left panel. So you can start multiple agents and use a single interface to communicate with all of them. Uh, and I can take my phone, open the-- Let me open the cloud app, and I will try to demonstrate that. Uh, but first, I will ask it, "Are you there?" I'm not sure this will be visible, but like I have the same interface here.

    5. AGAakash Gupta1:14:17

       Yep, I can see it. And so you can basically talk to it on web and mobile now.

    6. PHPawel Huryn1:14:23

       Yeah.

    7. AGAakash Gupta1:14:23

       And so there's really no excuse. You're out on a walk, you're at a meeting, whatever it might be, you're at a conference, you can have your agents running for you.

    8. PHPawel Huryn1:14:30

       Yeah.

    9. AGAakash Gupta1:14:31

       And what about Code Web Sessions? You said you use those roughly 60% of the time and Dispatch 40%?

    10. PHPawel Huryn1:14:37

        Yeah, I don't remember the exact, the exact stats. Like I, I really use all, all three surfaces, uh, and this proportions differ from day to day. Um, but most of the time I use Dispatch and Web Sessions, um, maybe like 70% together, uh, then 5% chat sometimes, and the rest is Claude Code. And the reason I, I use Dispatch so much is that, uh, I just don't work with my laptop. I, I go for a shopping, I go somewhere with my kid, and I just dispatch tasks. Uh, I already explained that I don't code. [chuckles] I don't. Uh, I just provide feedback, uh, text feedback in the chat, and then Cowork Dispatch presents me the results. Uh, I look at the results, I, I dispatch another task, and then I can continue what, what I was doing. [chuckles] So yeah, it, it really transforms my, transformed my, my ways and my days.

    11. AGAakash Gupta1:15:42

        Amazing.

    12. PHPawel Huryn1:15:43

        Yeah, we have not covered Web Sessions. So Web Sessions are something different. Uh, so here in Cowork, in Dispatch, depending on the task that you give it, it can either dispatch task to Cowork, like I asked about last three tweets, and you see there is a dispatched thread here. So this is something where I, I do not have to interact with it. It's dispatched talking to this agent, but it is visible for humans. And similarly, I can also dispatch some coding tasks from Dispatch, and it will, it will implement that and report back. Uh, oh, it even created a graphic for your post. Like, like let's say, "AI prototyping graphic."

    13. AGAakash Gupta1:16:32

        Ooh, I wanna see it.

    14. PHPawel Huryn1:16:34

        Uh, [chuckles] open like-

    15. AGAakash Gupta1:16:39

        Whoa.

    16. PHPawel Huryn1:16:39

        Yeah.

    17. AGAakash Gupta1:16:39

        Pretty good.

    18. PHPawel Huryn1:16:41

        Yeah, sometimes you need a few iterations, but it is not the, the, the worst point to start. [chuckles]

    19. AGAakash Gupta1:16:48

        Very cool.

    20. PHPawel Huryn1:16:50

        Yeah. Uh, okay, so this is this Dispatch. So we can think as Dispatch, normally you don't use it on, on your desktop. You, you have the single chat, single interface that you use on mobile. And, uh, but sometimes, mm, but co- for Dispatch to work, your computer must be online. And sometimes this is not the case, and sometimes there is a problem, and it stopped w- stops working. Uh, or sometimes you have so many parallel threads that it is difficult to manage them from a single chat interface. There are no tabs here. Uh, on your phone, you will see just one big stream of messages. You cannot organize them. Uh, so when I do a more complex work, I switch to this Code task, and this is-- You can think of it as, uh, Visual Studio Code and Claude Code. It's just in the cloud, so it is hosted by Anthropic. Um, so it's, yeah, it's just, just a lit of se- a list of sessions where I can select a specific folder, either my local folder or folder in GitHub, like my editor project, it is synced with GitHub. So all those files, so all knowledge files, uh, hypotheses, uh, sound bites, hooks, this is all synced with my private GitHub rep-repository. And, and even when my laptop is offline, I can go here from my mobile phone and ask it some question, and it, it will, it will work in the cloud wi-without, uh, any device.

    21. AGAakash Gupta1:18:43

        Yep, and this is actually a more secure way to run stuff, guys, [chuckles] 'cause it's in Anthropic servers. So I highly, highly, highly recommend everything you build, all your operating systems, you put them into GitHub, you point them via Code Web Sessions, and you work with stuff here. You can also-- The coolest thing, as we just showed with Dispatch, is you can start a chat here, start something here on your desktop, and then you can go to your phone. And so it's like you can be working twenty-four/seven with this.

    22. PHPawel Huryn1:19:12

        Yeah, that's correct. And, uh, I, I don't work [chuckles] twenty-four/seven, but, uh, like I feel that my life is now much more-- work's much, much better integrated with life, so I don't have to have these blocks dedicated to work. I can go on a shopping and yeah, maybe when, uh, when shopping, dis-dispatch some task and then check 10 minutes later when I have a graphic, provide some feedback, and then continue what, what I was doing. Uh, so this is really transforming. And when you need a better organization or when you know that you want to focus on coding, like my accridia.io platform, mm, then I switch to code, and I also do it primarily on mobile. Uh, and yeah, like in some cases I, I also use Cowork and, and Visual Studio Code, but [chuckles] primarily this is remote. Yeah, mo-most of the work is remote work.

    23. AGAakash Gupta1:20:12

        This is epic, guys. So hopefully you can understand how your life will change once you set all this up. I want to draw out some key lessons, Pawel, from all of the hundreds of hours you've now put into these things. Starting with, what's the biggest mistake PMs make when setting up Claude?

    24. PHPawel Huryn1:20:27

        Yeah. I think that the biggest mistake would be to prompt it every time from scratch instead of using Claude. I force myself to organize knowledge, and this is very difficult to me when writing articles and, uh, but Claude can do it without effort [chuckles] . So, uh, instead of collecting prompts and, uh, figuring out how to do something, it's much easier to just to build a system where Claude can learn from its mistakes and either from, from your feedback or from data, and it can figure out, hypothesize, and figure out the better ways to, to do work. Um, yeah, not doing

15. 1:21:07 – 1:32:31

    PM Mistakes and Future

    1. PHPawel Huryn1:21:07

       that, not organizing your knowledge, not learning from, from mistakes, and just having everything in your head, uh, and yeah, this is the, the biggest mistake that PM can make. Just co-- hoping that you can learn the better prompts. Uh, y-you can't. There is too much data to analyze.

    2. AGAakash Gupta1:21:27

       Cowork versus Claude Code. If a PM only had a little bit of time to learn one, which one should they choose and why?

    3. PHPawel Huryn1:21:33

       This is not an alternative. I would start with Cowork because everything you, you will learn in Cowork will help you better understand code. Uh, I use the same repo from Cowork and from, from Claude Code. So for example, I already presented that, but once again, I can select this editor project. Uh, maybe let's remove this one. Um, and it has Claude MD. It will see the same files. It will-- So this, this is Claude MD that I presented in Visual Studio [chuckles] . It will see the same files, the same structure. So I, I can ask it to analyze tweets here in this interface. I can do it on my mob- mm, in the, uh, using Dispatch. Uh, I can do it in, uh, [chuckles] web session, web Claude session, and I can do it in Visual Studio that is so connected to the same repo. Uh, I would start with Cowork because this interface is more simple. You don't have to, uh, get used to, uh, Explorer and terminal, and it's more user-friendly. You can see files. You can open those files directly here, uh, without plugins, without how to display markdown or how, how to display HTML. If you see HTML here, you click HTML, and it will show you HTML. And, uh, and in, in Visual Studio, there are cer-certain tricks that you, you must learn. So, uh, I would start with Cowork and just understanding how to work with agents, how to aggregate this knowledge, uh, how to define your workflows. And then once you feel comfortable with Cowork, then add this terminal aspect. That, in my opinion, that this is more effective. Should be more effective for many people than trying to, uh, learn Claude without experiencing Cowork before.

    4. AGAakash Gupta1:23:32

       What's your hot take on where AI PMs are headed? What does a PM's daily workflow look like in 12 months?

    5. PHPawel Huryn1:23:38

       I doubt, uh, that in 12 months, uh, the role will disappear [chuckles] or something. Uh, but yeah, we are heading into like s-super individual contributor PM and then CPO, CEO at the top. Uh, so I imagine that most of the time you, you will be working with agents, um, orchestrating multiple, like multiple agents at the same time, uh, switching, switching the context. It will not be easier. It, it, the work might be even more demanding. Uh, but at the same time, you will focus on-- There will be less trivial things because those can be automated, like writing tickets or debugging something or preparing a presentation. Those are the things that should be automated. So, mm, and eventually, uh, people who develop skills from multiple areas, so like P-shaped or even broader shaped that understand marketing, understand strategy, understand technology, understand products, can talk, talk to customers, and understand enough to delegate the work and, uh, assess the, the results.

    6. AGAakash Gupta1:24:53

       What's overhyped versus underhyped in the Claude ecosystem right now?

    7. PHPawel Huryn1:24:57

       I get the impression that people still have not realized [chuckles] uh, what agents can do, especially with the right harness and with the right systems around them. Uh, I just started ex-experimenting with that. Like I've been working with Claude and with agentsfor two years or more, but I just started to, uh, like investing heavily in automation and hardness myself, and I know that I can be much more effective. Uh, I doubt there is, there is something that is like there's no hype, there is not enough hype.

    8. AGAakash Gupta1:25:32

       So everything is under-hyped right now, guys, in the cloud ecosystem. Go learn what we've showed you right now. Final question for you, Pawel. Your n8n episode, it did really well. Is n8n over? I mean, should everything just be done in Claude Code now?

    9. PHPawel Huryn1:25:48

       No. [laughs] n8n is still relevant. There are two types of, uh, there are two types of automation. One automation is when you automate things for yourself, like, uh, I want to analyze 100 tweets, or I want to draft a, a response to customer email, then you have this personal automation, and you can use Claude Code for it. Also, we can automate specific processes inside your code base, like code review or release notes or front-end design, uh, and have those sub-agents. That's fine. Uh, but when you want to automate production processes, uh, you-- Like the logic that I presented, this is part of the prompt, and the, [laughs] the agent can respect it. It may not respect it. Uh, all we do is edit text files within hardness that was defined by Anthropic. We, we, we cannot, we, we cannot tell the agent that, "Hey, if Anthropic AP- API fails, try three times." Or, uh, for this, you should always use this tool before, uh, you should always verify that customer email exists before doing something, or that customer has access to this data. Uh, we just create text files, and we hope that agents will follow our instructions more or less like, uh, you can add hooks, you can add some... There are some exceptions from this rule, but overall, we rely o- on Anthropic hardness interpreting our, our text files. Uh, and in, uh, this doesn't scale. This is not, uh, secure enough, and this is not effective enough, uh, for production processes. So if I want to design a, a system that replies to customer tickets or maybe chats with the customer, I want to ha- have a hard guide- guidelines, n-not prompts that the customer cannot access data of another customer, for example, or that cannot delete data [laughs] or, um, uh, if you want to send an email, then yeah, like copy a file that has one gigabyte from one place to another. We don't want the, the agent to do that, or maybe even we don't want agent to look at this data. It should be handled by the code. So Aakash, I, I'm not sure you remember that we built an agent in three versions and why one was fully autonomous and, uh, we started with the least autonomous one where most of the process was code and there was one LLM call. Uh, then we built a hybrid scenario, and then the last version was fully autonomous. And in production, everything that doesn't have to be autonomous shouldn't be autonomous. So we should have code, we should have conditions, we should have guidelines. Like if there is a process that should be followed, it should be code. And maybe in this process there are some LLMs co- LLM calls or agents inside. So this is, this is more difficult to implement, but it is much more cost-effective. It's much more safe than relying on agents respecting instructions.

    10. AGAakash Gupta1:29:11

        So when it comes to the takeaway from our n8n episode, as we showed you guys, you want to actually be like less vague and with these Anthropic-based Claude Code systems right now, there's a lot of room for interpretation, less controls. If you're going to build a true production-grade automation for your company, you're still going to be using n8n. You're going to be defining as much as you can, some of those hard rules Pawel talked about. Did I summarize it correctly?

    11. PHPawel Huryn1:29:36

        Yes, and, uh, you can also use Anthropic API to, to define agents, uh, in code. Uh, but that's a separate story. And, uh, yeah, you can use Anthropic API to define your workflows in code, but this is not the same as organizing text files so that the agent follows the instructions. Um, we can also code, uh, code our workflows with, with Anthropic API or, uh, yeah, OpenAI agentic API or other APIs. Uh, the simplest way for a person that doesn't code is using n8n.

    12. AGAakash Gupta1:30:18

        Makes sense. All right, guys, we have walked you through the AI PM tool universe in today's episode. If you haven't yet, be sure to subscribe to Pawel's newsletter. He has an upcoming Quadathon starting May 9th, which you may want to participate in. Check out our other episodes if you want to learn more about AI product management, n8n, or customer discovery.

    13. PHPawel Huryn1:30:39

        On May 9th, so the ne- in a month, we are starting Buildathon, Cloudathon with Claude, and the previous edition it was 250 students. We were building, uh, real products with n8n and with Lovable, and this time we will focus on Claude Code and n8n at Trigger.dev to build real agentic workflows. Uh, so yeah, I encourage you to check the program. There are only 60 places, uh, in total, and you can also see the gallery of the real things that our builders have shipped. So this is not the theory. There are a few dozens of solutions that are available to browse in the gallery, and you can vote for your favorites and also understand how they were built, what is the architecture, and yeah, um, even download some documents. So all this is public. And the next, uh, class starts next month.

    14. AGAakash Gupta1:31:41

        All righty, Pawel, thank you so much for being on the pod.

    15. PHPawel Huryn1:31:43

        Thank you, Aakash. It was a pleasure.

    16. AGAakash Gupta1:31:46

        I hope you enjoyed that episode. If you could take a moment to double-check that you have followed on Apple and Spotify podcasts, subscribed on YouTube, left a rating or review on Apple or Spotify, and commented on YouTube, all these things will help the algorithm distribute the show to more and more people. As we distribute the show to more people, we can grow the show, improve the quality of the content and the production to get you better insights to stay ahead in your career. Finally, do check out my bundle at bundle.aakashg.com to get access to nine AI products for an entire year for free. This includes Dovetail, Mobbin, Linear, Reforge, Build, Descript, and many other amazing tools that will help you as an AI product manager or builder succeed. I'll see you in the next episode.

Episode duration: 1:32:41