[00:02] for coding. You skipped anti-gravity because you don't code? Big mistake. Google just flipped one switch. The terminal is gone. The code is gone. The In this video, I'll show you what [00:17] actually changed, the one setting you need to flip and the catch most people are going to miss. Google shipped version 2.3.0 of the anti-gravity app and it comes with a brand new layout called knowledge work. It sits right [00:30] next to the old layout, which is now called software engineering. And here's the part that matters most. The agent is the same. Same frontier models, same access to your local files, same plugins, skills, rules, custom agents, [00:45] and MCP servers. MCP just means a standard way to plug the agent into other tools and data. So nothing got watered down. Google only changed the screen in front of you. Turning it on takes about 5 seconds. Open [00:58] anti-gravity, go to settings. That's command and comma on a Mac or control and comma on Windows and Linux. Click general, find interface layout, pick knowledge work. That's it. It switches right away, no restart. If you already [01:13] use anti-gravity, you'll still be on software engineering by default, so you have to flip it yourself. And if you're brand new, anti-gravity now asks you a question the first time you sign in. What best describes your work? You can [01:25] pick software engineering, knowledge work, data and analytics, science, or design. Your answer sets your starting layout, your starter prompts, and the plugins it recommends. Google Docs say knowledge work, science, and design [01:38] start you in the knowledge work layout. Software engineering and data and analytics start you in the coding layout. Now, before I show you exactly quick pause. If you want help going [01:50] anti-gravity, check out the AI profit boardroom. We've got a 30-day AI agents roadmap that takes you from setup to using agents in your work one step at a time. And there are four coaching calls every week. So, if you flip this setting [02:05] and get stuck on a plugin, a permission prompt, or what to even ask the agent, you can bring it to a call and get unstuck. It's a place to learn this stuff properly with people who are learning it right alongside you. Link in [02:18] aiprofitboardroom.com. Okay, so what does the Knowledgework screen actually look like? There are four big changes. Number one, projects. In the coding layout, every project is tied to Git. Git is the tool coders use [02:34] to track changes to their code. So, you get prompts about work trees, branch pickers, and buttons to open things in a code editor. In Knowledgework, all of that is hidden. Your projects are just folders, normal folders on your [02:46] computer, and you can also start a quick chat with no project at all. Number two, the conversation. In the coding layout, you see every tool the agent uses, every step, every line of code it changed, little cards for each [03:00] helper agent running in the background. It's a lot. In Knowledgework, all of that turns into one live progress line. It shows what the agent is doing right now, and when it's done, that line clears. You just get the answer. Number [03:13] three, permissions. This one is huge for non-coders. When the agent wants to do something, it asks you first. In the coding layout, those asks can look like raw commands. In Knowledgework, you get a plain language summary of what the [03:26] agent wants to do, so you actually understand what you're saying yes to. Number four, the side panel. In Knowledgework, it stays hidden until the agent actually makes something. Then it shows up with documents, artifacts, [03:39] links, uploads, and background tasks. The review tab is gone, the terminal is gone, and the files tab is now called folders. Artifacts, by the way, are the things the agent makes while it works, plans, task lists, documents, that kind [03:53] of stuff. Google built anti-gravity around them, so you can check the single step. Settings get cleaned up, too. Plugins front and center. Skills, rules, custom agents, and MCP servers still run in the background. And the [04:09] techy stuff like the terminal and detailed permission rules moves under advanced settings. So the agent is just as strong. You just don't have to stare at the engine anymore. Now I promised you a catch. There's actually three. [04:24] Then I'll show you one more update from the same week that's just as important. Catch number one. Knowledge work is for personal Google accounts. Google's own docs say business and Gemini Enterprise accounts always use the software [04:38] engineering layout. They don't even see the interface layout option. So if you're signed in with a work account and you can't find it, that's why. Catch number two. Google says the rollout is gradual. So if you update and you don't [04:52] see it yet, give it a few days. Catch number three. Be clear on what this is. It's a new screen, not a new agent. It's still a chat with an agent plus a panel for what it makes. It's not a full document editor like Google Docs. [05:08] There's no live editing together with your team and there's no publishing flow built in. So think of it as your agent's desk, not your final office. Okay, now that other update. The same week Google also rolled out remote agents in [05:22] anti-gravity. Here's what that means. Normally the agent runs on your computer. Close your laptop and the work stops. With remote agents the work runs in an isolated cloud environment from Google. So it keeps going even after you [05:37] You set up a remote project with a GitHub repo or a Google Drive folder and you can check on it from your phone. Right now remote agents are for Google AI Pro and Ultra subscribers signed in with a personal Google account. They're [05:52] not on the free plan yet and not on Workspace or Google Cloud accounts. So why does this matter for knowledge work? Put the two together. A simple screen anyone can use and an agent that can keep working in the background. That's [06:07] Google pushing anti-gravity way past just code. One more small thing in version 2.30. If you already use other agent tools like Claude, Code, Codex, or Cursor, anti-gravity can now import your MCP servers, skills, and instruction [06:24] files from those tools into a plugin, so you don't have to set everything up again from scratch. So here's what I'd do. If you've got a personal Google account, update anti-gravity, flip interface layout to knowledge work, then [06:37] give the agent a real task from your normal day, like researching a topic or turning a pile of notes into a clean document. Watch the permission summaries, check what shows up in the side panel, and if you ever want the [06:50] full developer view back, it's one setting away. The big picture is simple. AI agents used to look like they were only for coders. Google just said that's not true anymore, and the people who learn these tools now are going to be [07:04] way ahead of everyone who waits. There's a window right now where most business owners haven't figured this out yet, and that window won't stay open forever. If you want to get ahead on learning tools like anti-gravity, join the AI Profit [07:17] Boardroom. Inside, you get daily tutorials so you can keep up as tools like this change. You get four coaching calls every week where you bring your questions and get help live. You get a prompt library so you're not starting [07:30] from a blank box every time, and you get the member map so you can see who else description or go to aiprofitboardroom.com. If you want the full process, the SOPs, and over 100 AI use cases like this one, [07:46] join the AI Success Lab. It's our free community. Links in the comments and description. You'll get all the video notes from this episode plus access to notes from this episode plus access to over 94,000 members. New AI trainings [07:58] get added in there every day, and it's free to join. The video notes give you every step from today so you can switch to knowledge work without rewatching this whole video. So come join us and I'll see you inside.