For those that may not have been paying attention to all of the Anthropic updates recently, they have eliminated Cowork and moved all conversations from the Desktop app to the cloud. Code from desktop still reads your computer folders. They have been rolling this out slowly to users and it started on October 6th. I noticed it the first time last week when I got a message all conversations have been moved to the cloud. 1. Claude Cowork and chat are one Claude (updated this week) This is the main article. Chat and Cowork are merged into a single conversation, so you no longer pick a mode first. Claude decides which tool to use, and anything that used to be Cowork-only now works from any conversation. It is rolling out in stages to Pro and Max on web, desktop, and mobile. If your message box still shows “Chat” and “Cowork,” you don’t have it yet. Once your account switches, you can’t go back to separate options. 2. Use Claude Cowork on web, desktop, and mobile See the section “What changed for Pro and Max plans on October 6.” As of October 6, 2026, new Cowork tasks on Pro and Max run in the cloud, and the “Only on your computer” option in Settings > General is gone. If you were still on local-only, you get moved the next time you open the desktop app, with nothing to set up and no way to switch back. Timeline from the official release notes - July 7, 2026: Cowork came to web and mobile, with sessions running remotely so work continues after you close your laptop. Chat and Cowork also began sharing one home. - August 25, 2026: Memory started working across chat and Cowork in the cloud. - September 16, 2026: The combined experience began rolling out, with chats, Cowork tasks, projects, connectors, and skills carrying over. - October 6, 2026: Cloud became the only option for new Pro and Max tasks (above).

I have been using Claude Code, Cowrk, and Chat for more than 2 years. I built my existing folder structure before Cowork was even a thing, so all of my AI project folders are on my root folder where Documents, Desktop, Downloads, and the other system folders are located. By the time Cowork was released I had already built several apps with repos, databases, and deployments that relied on my current folders staying right they were or I would have to rebuild everything. Now I'm trying to modify this mess of a structure, which as a global ~.claude/CLAUDE.md file that is now over 1MB, which by markdown file standards is the bloat of all bloats. I just updated the icm-architect skill to use AGENTS.md files instead of CLAUDE.md, but in the global hidden folder, this does not work, because it is ONLY Claude specific. So, as I'm making these changes I'm getting pushback from Claude. The biggest is routing. Here's my question. Claude code uses the .claude folder, so I can't get around this. Should the CLAUDE.md file be a duplicate of the AGENTS.md at the ~root folder, in my case? Right now my ~/AGENTS.md file has 11 global rules, mapping to all the project folders, and instructions to use the icm-architect skill for creating new projects.

I just joined this community this morning, went through the Foundation class, I already had a good understanding of folder structure, been using Claude and Claude Code for 2 years, then went through the AI Animations & Workflows up to 1.7, the mp4 download. Here's my new video test from the project; https://vimeo.com/1237753757?share=copy&fl=sv&fe=ci Now, if you will excuse me, I must go create my second test project for task 1.8 building on this first video. Interesting process. I didn't really need to give Claude Code the prompts in the video lessons after the first prompt, it knew what to do already. This tells me that the instructions in the .md files are working and I could see as the project was built that it stuck to only the tools it needed. I used only 7% of my 1M context window to get to the mp4 rendering, which is incredible considering the amount of token spend I'm used to using. For transparency I have the Claude Max 200 plan. I'm in the process of adding every parcel in the US to a database, I started in April 2026, and I'm using Claude Code to accomplish it. We eat through context tokens like wildfire and even the starter prompt for each session is at 20% context window before we even get moving. I've already asked Claude if the ICM framework for continuing the ingest of all the parcels, there are ~119M in the database as of today, is feasible and Claude said this will solve our memory and history issue, as I have hit their limits and it's been dropping the older sessions. I look forward to learning more, and starting my first real build from scratch, before I I add the framework to my property-registry project, and I will continue to report my progress. David Monroe, CCIM Founder/Investor/Broker/AI Consultant/CCIM Instructor

I'm bringing this subject to the entire community that started from a different post in the comments. I'm building a property database that will include every parcel in the US. There are currently 119,746,668 parcels in 1855 counties and 48 states in the db. There are 3143 counties in the US as of the 2020 Census. The database is currently 567.21GB. It has taken my more than 6 months to get to this point. I would like to create an ICM folder structure that can ingest the remaining 1288 counties to reduce time and token cost. It won't change the growth of bloat to the db, but that's OK, I'm no where near the db limit. I already have a good idea of how to update the database after all 3143 counties have been ingested by having a controlling markdown that send the request to a state markdown. The state markdown will then take the request for the county and send that request to the county markdown where the instructions for the update will be held. The largest number of counties in a single state is in TX at 253 counties. I have already created a spreadsheet that lists all 3143 counties and gets updated with successful ingests to the db with the correct url endpoint, script, and instructions on how the the ingest for that county works. My challenge remaining, because every remaining county has a different way to find the endpoint and the data they allow exported, is unique to each county,. How do I create an ICM folder structure to complete the remaining ingests? Thank you for your help in advance! David Monroe, CCIM Founder/Investor/Broker/AI Consultant/CCIM Instructor

Welcome to Clief Notes. Here’s where to start. 👋 You don’t need to watch the entire classroom before trying something. Pick a starting point, make something small, and bring your questions back here. 🧭 Find your way around Navigating the Course shows you what’s here and how to get around. If you’re new to working with AI, start with The Foundation. We work through prompting, context, and organizing your files so you can give an AI a job you can actually check. You’ll find guidance for Claude Code and Codex, plus a regular chat route where you save the files yourself. 💬 Use the check-ins At the bottom of the lessons, you’ll find polls and discussion posts. Share what you tried, what worked, and what needed correcting. If something doesn’t make sense, post the lesson and timestamp. “I got stuck right here” is a perfectly useful question. 🏆 Put something into practice Visit the Competitions Mega Thread for the current challenges, rules, and prizes. It’s a good excuse to finish something and see how other people approached the same problem. 📦 Looking for more? Premium: Explore The Vault for practical packages, templates, meeting takeaways, and worked examples. VIP: Head into The Drawing Room for VIP sessions and resources. Post your work. Ask questions. Help someone else when you can. What are you here to build—or what’s one task you’d love to make easier? Tell us below.