Loading summary
A
The head of product at Entropic just said seven words that should make every knowledge worker sit up and pay attention. We are now transitioning into Vibe working, not Vibe coding. Vibe working. And if you are sitting there thinking, okay, another AI buzzword, I can just forget this. It's going to disappear in the next couple of weeks. I need you to stay with me for the next 12 minutes. Because what just shipped last week from Claude and Antropic wasn't an incremental update. It was really a moment in time. It was when I stopped being the tool you talk to and became a team you manage. By the end of this, you're going to understand exactly what Vibe working means, why it changes the game for anyone who works with information for a living and the specific shift you need to make right now before everyone around you figures it out. First, let's get into today's show.
B
Before we continue, let me tell you about a podcast I love. Marketing School. Marketing legends Neil Patel and Eric Su bring you daily actionable digital marketing lessons learned from years in the trenches. Marketing School delivers bite sized marketing wisdom you can implement immediately, whether you have a new website or you're an established business. You'll learn the latest in SEO, content marketing, social media, email marketing, conversion optimization, and general online marketing strategies that work. Today they just did a great episode called 10 SEO Lessons that Still Work in the AI Era. Listen to Marketing School wherever you get your podcasts.
A
Okay, let's rewind 12 months. The big thing everyone was talking about was Vibe coding. Andre Kapathe, one of the ex co founders of OpenAI, coined the term here in this tweet. The idea that you could describe what you wanted built and AI would just write the code. And that was based upon his early experience with cursor and cloud sonnet. And it worked and it's gotten much, much better. People with zero code and experience started to chip in apps and developers started building 10 times faster. It was a huge shift. I'm an ex developer, not a very good coder and I was able to launch my own product through a lot of this Vibe coding. But here's the thing no one really talked about in relation to Vibe coding. It only mattered to people who were building software, right? What about the rest of us? Us marketers, the analysts, the consultants, all of the different founders who are drowning in decks and spreadsheets and research. The millions of knowledge workers who don't write code but still spend their entire day producing work, producing output. That's what's just changed this week. Entropic the company behind Claude, shipped a whole set of updates that take the same principles behind Vibe Coden and apply it to every single knowledge worker. Their head of products, Scott White, who came on the show, really said it perfectly when he said Claude went from a model that you can kind of sort of talk to to accomplish a very small task and get an answer to something that you can actually hand real significant work to. And it can do it really autonomously so you can get an outcome and it can do that work. That is not buzzword speak, that is not marketing speak, that is a description of what the head of product at Antropic actually said. And I want to break this down into. Into three specific launches from last week. And I want to make them real for you, to tell you the core shifts that are happening and why last week was just so historic. Okay, Shift one is really agentic teams. So here's how Most people use AI today, right? You open ChatGPT, you open Claw, you type a prompt, you get a response. You are saying, this is the task I want. This is an example of the output I need. And you probably go back and forth and you refine it, you make it better. That's like hiring a person and kind of giving them one job at a time, right? I give them the task, I kind of describe my outcome, I work with it, and it's able to come back with the output. What Anthropic just launched is called agent teams that can orchestrate in cloud code sessions. And it really starts to flip this model. Instead of giving AI a single task, I describe an outcome. I say, hey, I want competitive analysis of these five companies. I want a summary deck and then I want a brief that I can send to my CEO. And in the way we work with AI today, I would do all those things separately, right? I would probably have to do some competitive analysis, give it a ton of content. To be able to do that, I would have to describe the five companies. I would have to, like, give it a template to create the deck. I would have to give it a template and say, here's how you write the brief. That's how we would do it in the past. But now with these agentic teams, Claude just starts doing the work. It spins up a team of agents and one is going to do the research and one is going to write, one is building the deck. So it builds up this team of agents and is able to orchestrate the work across them. So it's really creating digital workers who are specialized in individual tasks and Then it orchestrates that work across them. These are really like actual colleagues.
C
Look, if you want to get ahead of the vibe working era, we just dropped a guide on the $100,000 AI jobs that no one is talking about. These roles are paying anywhere between $50,000 to well over $100,000 salaries. We put together a guide that shows you exactly how to position yourself for these roles before everyone else catches on. Get it right now. Click the link in the description. Now, let's get back to the show.
A
So let me tell you a quick story to kind of bring this to life. A friend of mine, they run a small agency. There are six people in that agency, so it's a pretty small agency. And when they go and try to get work, they have to pitch for that work, right? So they were trying to pitch for a big potential client last month and the entire pitch took them a week, right? They did full market analysis, they did strategic dog, they did a presentation. You know, he tells me, hey, we're spending like a week, we have to do this. All of the other agencies are doing that. We turn up, we pitch, we try to win those dollars. Well, with Claude code and the skills and the orchestration of this work, you can basically just say the outcome is an incredible pitch for this customer that I want to win on these reasons, right? Like go pull the information, talk about our key selling points and show some examples of how we can make their lives better. And it will actually go and orchestrate that across your team. So now that agency, he can spend much more time doing the work of clients, making that work really great because he can have a team of agents orchestrated by cloud code to do all of the pitching. So for him, he's gone from orchestrating that work across a small team and doing some of that himself to being a agent orchestrator, right? He's actually using agents to go and do that work. He's a manager of agents, not manager of people. And the people are busy crafting great assets and marketing and things like that for clients. This is available right now in cloud code. And honestly, this is the reason in a recent post I talked about what I'm spending my time on this year, cloud code was number one. I got a lot of feedback to say, hey, you're a marketeer, ex developer, by the way, but you're a marketeer. Why are you learning the cloud code? Because Claude code is. Can do way more than actually build code. I use it for tons of marketing use cases. I've built a ton of Claude code skills that help me do my work. And now I can build teams of agents that can orchestrate work across them to do that work. Stay tuned. Of course, we have you covered with a cloud code for marketers coming very soon. Okay, shift number two, and this one is really going to hit home for a lot of you knowledge workers. Right. If you're watching this video, you raise your hand if you've ever had to use Excel or PowerPoint. I'm sure most people are using those tools. And what you really do is you open chatgpt, you kind of copy and paste some content into them that gives you some output and you copy and paste it back in. Right. If you're using PowerPoint, you're probably going to work in Claude. You're going to take some stuff, put it into PowerPoint and back and forth. The same with Excel. Well, that entire back and forth has really just disappeared because CLAUDE is bringing agents into the actual tools that you use. So CLAUDE can now sit inside Excel. I'm a really good data person. I was never really good at Excel. I've always been very fortunate to have great analysts who work with me and I know how to do data analysis. Well, now I don't have to worry about that because I have the ability to understand what data matters, turn it into action and cloud code and do all of the data work for me in Excel and build all the stuff for me. So that's an example. And not just that you can do the same sort of thing in PowerPoint. So again, it's really becoming an assistant that's available in all of the tools that you use. You can start to just build all of your slides in PowerPoint. Also a shout out to Claude. In general, building slides in Claude has gotten so much better. You can really build great decks right within claude. But the key thing here is CLAUDE is really bringing agents into your workspace which where you can use them with the tools that you use each and every day. So that's shift number two. I thought that was pretty incredible. And then the shift number three is really Opus4.6 with its 1 million token context window. So Opus4.6 was released on Thursday. I'm recording this Friday, and it is pretty amazing. I think this is going to be another step up. One of the biggest limitations of AI up until now has been context. And context is you can upload some documentation and then you can work with that documentation and that data right within that context window. You've had this incredibly powerful tool, but you can only feed it so much data at once. And you'd have to chop up the work into little pieces to be able to work with it on the tasks you want to work with it on. Claude now has a 1 million token context window. To put that in plain terms, you can feed it your entire code base, right? If you're a developer, if you have a piece of software and you have a large code base, you can give it all your financial filings, you can give it every single doc you want to give it all at once. And it has that in its context window. And, oh, Opus 4.6 can now work with that context window. And Opus 4.6 is actually incredible. But here's why this really matters. When you have a 1 million token context Windows and you can provide it much more data, AI is able to make connections across all of this data versus you trying to piecemeal it together. It can look at thousands of pages of documentation and it can figure out relationships between something that's on page a thousand and something that's on page one. Right? It can really maintain consistency across the task that you're trying to do with it. It doesn't lose its thread like it does when you're trying to piecemeal the data in. And that, paired with the fact that opus 0.6 is another step up in terms of reasoning models and thinking models, means that you now have a powerful AI assistant that can autonomously do tasks in a much, much greater way. And so they're really the three trends that have happened over the past number of days. We've seen agent teams become a real thing. We've seen AI being brought into the tools you use. We now have AI that has a huge context window and, and is a step up again in terms of thinking and reasoning. And we should just step back and think about what is the bigger picture, right? If we really kind of zoom out here, because the real insight isn't about a single feature, and that's what I'm trying to show here. It's really about a pattern. Vibe coding was a big deal because it allowed you to describe what you wanted built and AI was able to write the code and build that thing. And now over the course of the last year, the ability for AI to code that thing without you having to step in and provide feedback and iteration and debugging has got better and better and better. Vibe working is going to go down the same path. It's describing the outcome you need and AI is going to do the work. And that means AI is going from, hey, AI, you can do this like one narrow task to hey, AI, you can do the entire piece of this work yourself because you are orchestrating that work across different agents. Now, I think there's a version of this online that is all hype and no substance. That is not what we do here. So here's what I really believe is happening. Part of every knowledge worker's job is going to become building and training and deploying teams of AI agents. Not using AI, not prompting, but actually managing AI teams. And if you think about what a manager does, a manager doesn't do every task themselves. Right? A manager defines the outcome, they assemble the right people, they give them context, reviews their work and they iterate with that team. And that's the skill that is really going to become relevant for every single knowledge worker, not just people with direct reports. And here's the thing to bring this into reality of how important is most people aren't prepping for this. They're still in the I'm going to go to ChatGPT and let me ask you a question and be amazed at ChatGPT can come back with the answer. That's kind of like being in 2008 and still thinking the iPhone is just a phone with a touchless screen. The way you think about AI and what's happening is way too confined, it's way too small. It's not an answer engine. It is a platform that allows you to build an entire digital team. So what do you do about it? Let me give you three concrete steps. First, you're going to need to start thinking in outcomes and stop thinking and tasks. So when you go to AI, create me a blog post on the 10 tips to do this thing is a task. When you go to AI in the future and you're able to have an agentic team, you may ask it to say, hey, I really need a content strategy that can actually help grow my ex volume by 30% for technology founders. And I want to post three times a week and I want to be able to figure out how I can measure the success of that month on month. Do you want to think of an outcome? Right. I want to grow my audience by 30% for this audience. And so that is one way to start thinking about how you're going to work with AI outcome based versus task based. The second is Claude code and I will have a lot of things coming around Claude Code, both on this channel and everywhere else that I am. I know that sounds like I'm telling you to go learn coding. It's not cloud code is Really a command line tool where a lot of these agentic teams live. And I use it for lots of things, not just building software. I use it to do a lot of research, I use it to do a lot of marketing content tasks. It is an incredible tool that allows you to build skills and have agentic teams to do everyday work. Cloud code is much more than building software. And the third one is you are going to have to start thinking in terms of workflows, right? Not using AI for these kind of one off tasks, but trying to think about how I can build workflows that are repeatable and scalable and can be handed off to AI teams. People who win in the Vibe working era aren't going to be the ones who use AI the most. You know, AI usage doesn't necessarily correlate to you being the most productive person or being the most impactful person. The people who are really going to succeed in the Vibe work in an era are the ones who systemize AI and come up with repeatable workflows and know how to delegate that to their agentic teams to do. Here's what I want to leave you with, right? Every few years there's a shift that splits the timeline into the before and after. In HubSpot, we love a today and tomorrow slide. Here's how things work today, here's how things will work tomorrow. The Internet was one, mobile was one, Vibe coding was one. And now Vibe working. Is death going to be one? Not because technology is magic, but because it changes who does the work. For the first time in history, a single knowledge worker can describe an outcome and have a team of agents execute on that work. They're going to be doing research, write in analysis, presentation, code generation all at once. Again, they do it in parallel. They have a team of agents doing this in parallel inside tools that you already use. The question isn't whether this is happening. I want to prepare you. This is happening and it's going to get faster because all of these companies are using their own tools, agentic teams to ship these products. It's whether you are going to be the person who manages the agents or you are going to be the person who gets replaced by them. If this video was useful for you, make sure you subscribe. I know everyone's going to watch this and they want to go, well, how do you do I get into the viworking era, I need to learn Claude cowork and Claude cowork skills. I need to learn Claude code for marketeers and Claude code for growth professionals. We have got you covered. More tutorials coming your way. Subscribe to the channel and I'll see you in the next video.
D
Hey everyone. You know Kieran and I have been doing the podcast for a while now. We've been at this for a couple years. We love it. We could not be happier to be doing this, but we wanted to take things to the next level. We want to level up the impact we're having with marketing Instagram. So the next step of our journey is something we're really, really excited about. We're going to launch the Marketing against the Grain newsletter. And Marketing against the Grain newsletter is going to be amazing. If you are a marketing leader, practitioner, you're in the trenches doing marketing every day. This is for you. We're going to deliver right to your email inbox and you're going to get all the behind the scenes frameworks, practices, tutorials from us, from guests we have on the show, and from people even beyond the podcast that we think are going to be helpful and and really have an impact on your day to day, week to week doing marketing. You're going to love it. It is something we've been talking about for a while. We're really excited to have it out in the world. We've already got a hundred thousand marketers who are on this newsletter. Please join. It's completely free. We'd love to have you as part of the Marketing against the Grain community. And it's easy. You can click the link in the description below or you can head to marketingagainsthegrain.com subscribe.
Episode: The Claude Update That Just Changed Marketing Forever
Hosts: Kipp Bodnar (HubSpot’s CMO), Kieran Flanagan (SVP of Marketing, HubSpot)
Date: February 10, 2026
This episode dives into the monumental February 2026 update for Claude (by Anthropic), introducing the concept of "Vibe working." The hosts passionately explain why this isn't just another AI buzzword, but a true paradigm shift for knowledge work. They meticulously break down how last week's Claude update brings agentic teams, tool integration, and massive context capacity, marking the dawn of an era where marketers (and all knowledge workers) orchestrate teams of digital agents—reshaping productivity, roles, and competitive advantage.
"By the end of this, you're going to understand exactly what Vibe working means, why it changes the game for anyone who works with information for a living..."
– Kipp Bodnar [00:01]
"Now with these agentic teams, Claude just starts doing the work. It spins up a team of agents and one is going to do the research and one is going to write, one is building the deck... These are really like actual colleagues."
– Kipp Bodnar [04:15]
"He's a manager of agents, not manager of people... The people are busy crafting great assets and marketing and things like that for clients."
– Kipp Bodnar [07:10]
"The entire back and forth has really just disappeared because Claude is bringing agents into the actual tools that you use. So Claude can now sit inside Excel... do all of the data work for me... you can do the same sort of thing in PowerPoint."
– Kipp Bodnar [08:08]
"Opus4.6 was released on Thursday... And it is pretty amazing... you can feed it your entire code base... every single doc you want to give it all at once... It can look at thousands of pages of documentation and figure out relationships..."
– Kipp Bodnar [10:29]
"It's really about a pattern. Vibe coding was a big deal... Vibe working is going to go down the same path. It's describing the outcome you need and AI is going to do the work."
– Kipp Bodnar [12:09]
"Part of every knowledge worker's job is going to become building and training and deploying teams of AI agents. Not using AI, not prompting, but actually managing AI teams."
– Kipp Bodnar [13:03]
Three Steps for Marketers & Knowledge Workers:
"The people who are really going to succeed in the Vibe work in an era are the ones who systemize AI and come up with repeatable workflows and know how to delegate that to their agentic teams to do."
– Kipp Bodnar [15:36]
“Every few years there's a shift that splits the timeline into the before and after... The Internet was one, mobile was one, Vibe coding was one. And now Vibe working is definitely going to be one. Not because technology is magic, but because it changes who does the work.”
– Kipp Bodnar [16:30]
“It's whether you are going to be the person who manages the agents or you are going to be the person who gets replaced by them.”
– Kipp Bodnar [17:29]
Kipp Bodnar and Kieran Flanagan argue persuasively that this Claude update is more than a technical leap—it’s the beginning of a fundamental shift in who does knowledge work and how. Marketers who master outcome-based management of agentic AI teams will have an edge; those who stick to old models risk falling behind. The episode is a call to action for listeners: get hands-on with Claude, master Claude Code, and start building workflows and teams—or risk being replaced by those who do.
For further learning: The hosts tease upcoming guides ("Claude Code for Marketers") and invite listeners to their forthcoming newsletter for in-depth tutorials and frameworks.
Summary by Marketing Against The Grain Podcast Summarizer.