Claude Code Update is INSANE

Key Takeaways

Claude Code update is insane. You're sitting at your desk, stuck, waiting for your dev environment to load. Again, you open your terminal, you type, you wait, you lose your train of thought, you switch windows five times just to get one thing done. Sound familiar? This is what most people's coding setup looks like in 2026, and it's costing you hours every single week. But what if you could kick off a coding task from your phone while you're getting coffee, and come back to a finished pull request? That's exactly what Claude Code on the web and mobile lets you do right now. And most people have no idea it exists. Stay with me, because by the end of this video, you'll know exactly how to use it. Hey, I'm the digital avatar of Julian Goldie, and I help people learn and actually implement AI tools in the real world, not just talk about them, actually use them. And today we're going deep on Claude Code on the web and mobile because Anthropic just refreshed the whole experience, and there are some genuinely useful things you can now do that weren't possible before. Let me start at the beginning for anyone who hasn't used Claude Code yet. Claude Code is Anthropic's AI coding agent. It reads your code base, understands your files, your tests, your Git history, all of it. You tell it what you want in plain English, and it does the work. Write code, fix bugs, run tests, create pull requests. It's not just auto complete, it actually understands the structure of your project and makes decisions based on that context. Before this update, Claude Code was a terminal only tool. You had to be at your computer in your dev environment watching it run. That was fine for a lot of tasks, but it meant you were always the bottleneck. Had to be present for anything to happen. That constraint is now gone. So let's get into what actually changed. Claude Code on the web lives at claude.ai/code. You connect your GitHub account, write a prompt, and kick off full coding tasks straight from your browser. No terminal, no local setup. Anthropic runs everything on their cloud infrastructure in an isolated virtual machine. That VM comes with Python, Node.js, Ruby, Go, Rust, Java, PostgreSQL, Redis, Docker, Git, all pre-installed. Claude clones your repo, does the work, and you review the diff right there in the browser. Here's the part that makes this genuinely different. Sessions persist even if you close your browser tab. Claude keeps working in the cloud. You come back later, it's still going. Or you check in from the Claude mobile app. That's the mobile piece. You can monitor running sessions, steer Claude mid-task, answer its questions, and approve things, all without touching your laptop. Now let me walk through how you actually use this. The simplest way is straight from the browser. Go to claude.ai/code, connect GitHub, pick a repo, write your prompt, hit go. Claude spins up a fresh VM, clones your repo, and starts working. You can watch in real time, leave inline comments on diffs, and when it's done, you create a pull request right there. The second way is from your terminal with the remote flag. You type Claude remote followed by your task in quotes. Claude creates a cloud session based on your current repo and branch, runs in the cloud, and you keep working locally. Check progress with tasks in the CLI, or flip over to the browser or mobile app to watch it. There's also a smarter workflow for complex tasks. Start Claude in plan mode locally using permission mode plan. Claude explores your code base and proposes a plan without touching your code yet. You review and agree on the approach, then you send the actual execution to the cloud with remote and let it run autonomously. This one step saves a lot of wasted back and forth. Think of it like a contractor reviewing blueprints before breaking ground. You get visibility and control at the planning stage, then Claude executes without needing you to hover over it. And then there's auto fix for pull requests, which might be the most underrated feature here. Once Claude creates a PR, you turn on auto fix, and Claude watches that PR on GitHub. CI check fails, Claude investigates and pushes a fix. A reviewer leaves a comment, Claude reads it and addresses it. You come back to a pull request that already responded to feedback, and Claude notes in the thread that its replies were written by Claude Code, so reviewers always know what came from the agent versus what came from you. That transparency matters on a team. That's a real shift in how async work can flow. I want to pause here because when I first got into these kinds of AI workflows, it was genuinely overwhelming to figure out what actually matters versus what's just noise. And that's why I created a community called AI Profit Boardroom. Over 2,000 members all learning and implementing AI together, sharing what actually works in real workflows, not theory, real use cases. If you're serious about building your AI skills, check it out. Link is in the description. Let me cover how the cloud environment actually works because this matters. Every session runs on fresh Anthropic managed infrastructure. Your local machine doesn't need anything installed. There's also environment caching. Claude runs your setup script once, snapshots the file system, and future sessions start from that snapshot. Fast boot times even with heavy dependencies. Security-wise, your Git credentials and sensitive keys are never inside the sandbox. Authentication goes through a secure proxy with scoped credentials. Claude gets just enough access to clone and push, nothing more. You can also run multiple tasks in parallel. Run remote multiple times, and each creates its own independent cloud session running simultaneously. So while Claude is fixing a bug on one branch, it's also writing test coverage on another, and refactoring a module on a third, all at the same time. Monitor them all from tasks. When one finishes, you can teleport it back to your local terminal using teleport. It fetches the branch, loads the full conversation history, and continues right where it left off. Now some practical tips. Use the claude.md file in your repo to give Claude persistent context about your code base, your standards, your patterns. That file travels with the repo into every cloud session automatically. It's the single best thing you can do to get consistently good output from Claude Code on the web or in the terminal. Set up your cloud environment with a setup script for dependencies. The first session installs everything. After that, the cache kicks in, and sessions start fast. You can also configure network access for your cloud environment. Default trusted mode already allows connections to all major package registries, npm, PyPI, RubyGems, crates.io, DockerHub, GitHub, and more. If you need access to internal services or custom domains, you can add those in the environment settings. If you need to lock things down completely, there's a no network option, too. Use inline comments in the diff view when reviewing. You leave a comment on a specific line, Claude picks it up and addresses it. Much cleaner than going back to a chat box. For long sessions, use at compact to free up context window space, and context to check how full it is. If you're on a team or enterprise account, sessions have built-in sharing. Toggle visibility, share the link, and teammates can follow exactly what Claude did. Here's the actual shift this represents. Instead of Claude being a tool you sit next to while it runs, it becomes an async collaborator. You hand off work, you go do other things, you come back to results. The mobile piece makes that loop feel natural. You're not chained to your desk to give it direction. And because sessions live in the cloud, you don't lose work when you close your laptop or switch locations. The context, the conversation history, the branch, it's all still there waiting for you. That's a genuinely different relationship with a coding tool than most people are used to. Availability right now, Claude Code on the web is in research preview for pro, max, and team users, and for enterprise users with premium seats. Go to claude.ai/code to get started. If you want to go deeper on AI tools like this and actually implement them in your work, come join AI Profit Boardroom. Over 2,000 people learning how to use AI effectively. Everyone shares what's working, what's not, which tools are worth your time, and which to skip. No hype, just practical guidance from people doing the work. It's helped me stay on top of updates like this and figure out how to actually apply them. Link is in the description. And if you want the full process SOPs and over a hundred AI use cases like this one all in one place, join the AI Success Lab. Link's in the comments and description. You'll get all the video notes plus access to our community of 58,000 members who are putting AI to work every day. Thanks for watching. Go check out claude.ai/code and see what you can hand off today.