핵심 요약
Claude의 웹 브라우저 채팅 방식은 파일 복사·붙여넣기를 직접 진행해야 하며 대화량이 늘어날수록 성능이 떨어진다. 반면 Claude Desktop 앱의 Co-work 기능을 활용하면 사용자가 자리를 비운 사이에도 PC 내 수백 개 파일 정리를 자율적으로 완수할 수 있다. 웹 채팅을 넘어 데스크톱 에이전트를 활용해야 긴 호흡의 복잡한 작업(Long horizon task)을 효율적으로 자동화할 수 있다.
주요 포인트
- 대부분의 유료 사용자는 웹 채팅에서 질문하고 결과를 직접 복붙하는 단발성 작업 방식에 머물러 있음
- 브라우저 채팅은 많은 데이터가 채워질수록 AI의 맥락 처리 능력과 지능이 급격히 저하되는 한계가 있음
- Claude Co-work(데스크톱 앱)은 여러 단계와 다수의 파일을 다루는 복잡한 장기 작업을 자율적으로 처리함
- Windows, Mac, Linux 등 전용 데스크톱 앱을 설치하는 것이 Co-work 에이전트 활용의 핵심 출발점임
- 단발성 작업을 반복 실행 가능한 에이전틱 자동화 시스템으로 전환하여 생산성을 극대화할 수 있음
This morning, Claude organized 400 files on my computer. It renamed every one of them, and it left me a summary of what it found, and I wasn't there. I was making coffee. Meanwhile, most people who pay for Claude spent this morning the same way they always do, asking it a question, then copying the answer out, and then doing the rest of the work themselves. They have the same subscription I do, but we had two completely different mornings. And the difference isn't skill or price. It's one half of the product that they've never opened. I coach people on AI every single week and I see both sides of this all the time. So, I'll show you the exact setup for Claude Co-work, the best first task to give it, and how to turn that task into a system that runs again and again without you. So, let's get into it. Now, most of you with a Claude subscription, you're stuck in one side of the world of Claude, and that's in the chat. But very rarely do you ever expose yourself to the co-work side where most of the work actually gets done. If you ever see anybody talking about how they can autonomously do different skills and get serious leverage from tools, they're almost always talking about using desktop agents, specifically cloud co-work from OpenAI. They're almost never talking about using the chat, which is what most people still use. But that's slowly changing because both Anthropic and OpenAI are pushing people towards these desktop AIs that can work more autonomously on longer horizon tasks. And we can see that if we go to Claude. So if I open up Claude in the browser specifically, I can see that we have a new section here that says chat and co-work. So I can switch between these two. And what's going to happen is this black box pops up down here. And I'll talk more about what this means, but more importantly, the most important thing you need to do right now to start things off is to go to this page here, the download the Cloud Desktop app. So you need to get it on Windows, Mac or Linux. Doesn't matter what you're using. Download the app. Reason being is that there's a huge fundamental difference between using Claude in the browser and Claude on your computer. And it really just comes down to how effective an AI can work on a long horizon task, meaning a complex task that has many files or many steps inside of it. Now, the question is, why is that the case? Well, if you've watched any of my videos in the past, you know that I talk a lot about how intelligence drops quickly when you fill up an AI's head. Is the AI only has so much space in its head. the fuller it gets with data, the dumber it gets. When you're using chat inside the browser, this is what your experience looks like. Where we have an AI that starts out smart, but the more files you give it, the dumber it gets. Now, the painful part here isn't just that the AI's intelligence drops quickly. It's that you have to copy and paste files back and forth into the chat as it's processing them for you. Because let's say you want to process 25 files. Well, if you've probably experienced this, Claude has a limit on how many files you can give it at any given point. It's usually between five and eight files, depending on how big the files are. Quick pause. If you're enjoying this, you're going to enjoy two other things. First off, blow is a 30-day AI insight series, completely free. You'll get 30 insights in your inbox of how you can apply AI to your business and your work. The second thing is if you'd like to work with me, Blow series of offerings to see if there's a good fit between the two of us. Now, let's get back to the video. So, when you're copying and pasting these files back and forth, you have to copy in five and then wait for it to process. Copy in another five, wait for it to process. And that's a thing of the past because if you're using the right tools in the right places, you no longer have to do that. And that's where these desktop apps come into play. So first off, the intelligence is sustained for longer periods of time. So if the task is more complex, so there are more files, the intelligence can stay high on that task longer than it can in the chat in the browser. And that's because instead of you giving the AI the files, the AI comes to your files. And that's what this title helps you remember. You should be pointing, not pasting. pointing in the sense you're pointing your AI towards a folder with a bunch of files instead of you copying and pasting those files to the AI. And that's where co-work truly shines as long as you're using the desktop app. Now, I'll talk a bit more about the benefits of this. But first, let's look at some terminology. So, if you're going to open up cloud co-work or really cloud anywhere, you're going to come across a series of terms and you can get overwhelmed. So, I want to give you an 8020. So, what are the few features that matter most based on the tasks that most people do? Instead of me reading off 35 features and a