New Antigravity CLI Update Changes Everything!

Añadido: 2026-05-31

[00:00:00]New anti-gravity CL. I just landed and it changes everything about how you run a team of AI agents all from one black screen on your computer. Let me show you what I mean. May 19th at Google IO, Google did something big. They took their whole anti-gravity agent system and dropped it straight into the terminal. The terminal is that plain text screen developers use. No buttons, no mouse, just you typing. And here's the part that matters. This new tool called anti-gravity CLI can run a bunch of AI agents at the same time. Not one, a bunch, all working together, all in parallel. So picture this. You type one instruction and five AI workers spin up.

[00:00:37]One does research, one writes, one builds, one checks the numbers, one schedules the next round. The same time while you go make coffee. That's the story today. And I'm going to walk you through exactly how it works with real prompts you can copy and why this is a real shift, not just hype. Hey, if we haven't met already, I'm the digital avatar of Julian Goldie, CEO of SEO agency Goldie Agency. While he's helping clients get more leads and customers, I'm here to help you get the latest AI updates. And Julian Goldie reads every comment, so make sure you comment below.

[00:01:07]First, the basics fast because you don't need the long version. The gravity CLI replaces the old Gemini CLI. If you used Gemini CLI before, that old one stops working for most people on June 18th. So Google is moving everyone over to this new one. The tool is built in a coding language called Go, which just means it's fast and light. You run it by typing three letters, AGI. That's it.

[00:01:30]Type AGI and you're in. Runs on Google's newest model, Gemini 3.5 flash. And it also gives you access to Claude and other top models right inside the same screen. So you're not stuck with one brain. You pick the best one for the job. Here's a small thing that matters more than it sounds. Google says this new model runs about four times faster than other top models at the same level.

[00:01:50]Speed isn't just nice. When you've got five workers running at once, fast means you get your finished work back in minutes, not an hour. That changes how often you'll actually use it. Slow tools sit unused. Fast ones become a habit.

[00:02:03]Now, here's why I got excited. The old way of using AI looked like this. You open a chat, you ask one thing, you wait, you copy the answer, you paste it somewhere else, you ask the next thing, you wait again. One worker, one task, line of people waiting their turn. The new way is different. You give it a goal, and it splits that goal into pieces. Then it hands each piece to its own little worker. They all run at once.

[00:02:25]Then they come back and put it all together for you. Google calls these little workers sub-agents. Think of them like new hires you can spin up in a second who never get tired and who all report back to one boss agent that you control. And the cool part is you don't have to plan that out. You don't sit there going, "Okay, worker one does this, worker two does that." You just give the goal. Boss agent decides how to split it. It figures out it needs a researcher and a writer and a builder, and it spins them up on its own. You're the one with the idea. It handles the rest. Let me make this real with the first workflow. This one's content.

[00:02:55]Builds you a week of posts that pull people towards your community. Here's exactly what I type into the terminal.

[00:03:01]Goal, research the top five things small business owners ask about AI automation this week. Write five short, simple posts answering them in plain English.

[00:03:11]Each post by inviting people to join the AI Profit Boardroom to get hands-on help.

[00:03:15]Five into a Google Doc for me. And here's what comes back. Five finished, ready-to-post pieces sitting in your Google Doc, each one quietly pointing readers towards your community. The old world, that's different tools and an afternoon. Here, it's one prompt and a few minutes because the research sub-agent and the writing sub-agent run side by side. One's digging up what people are asking. The other's already turning it into posts. They don't wait on each other. Let me pause here because you might be thinking something. You might be thinking this sounds like it's for coders. I'm not a coder. This isn't for me. I get it. The terminal looks scary. Looks like the movies where some hacker types green text. You don't type code into this. You type plain English.

[00:03:54]You talk to it like a smart helper. Go do this, then do that. That's the whole skill. If you can write a text message, you can run this. Want to sit on that for a second because it's the thing that holds most people back. The screen looks technical, so people assume the work is technical. It's not. The hard part used to be the typing and the building. That part is gone now. The agent does it.

[00:04:14]What's left is the thinking. What do you want? Who's it for? What should it say at the end? Those are business questions, not coding questions, and you already know how to answer those. And that's exactly why I want to talk about it today because tools like this are moving fast, really fast. People who learn how to point five AI workers at one goal are going to get a year of work done in a week. And the people who don't will keep doing it the slow way, one chat at a time, wondering why everyone else is moving faster. That's the gap I worry about, and it's the gap we close inside the AI profit boardroom. Quick heads-up, if you want to learn how to set this up and use it to save hours every week, come check out the AI profit boardroom. We run four live coaching calls a week, and right now those calls are going deep on exactly this, how to use parallel AI agents to do your research, write your content, and find new customers while you sleep. We've got a 30-day roadmap built around getting your first AI team running, daily step-by-step tutorials, and 2,800 business owners in there figuring this out together, sharing what works. Link's in the description. Come learn this before everyone else catches up, okay?

[00:05:16]Back to the workflows. Here's the second one. This one uses a command called {slash} go. You type a {slash}, then the word go. It tells the agent, "Don't stop and ask me 100 questions. Just go. Run all the way to the finish." Here's the prompt I'd use. {slash} go, "Plan a free 1-hour workshop that shows small business owners how AI automation saves them time. Build me an outline, three talking points per section, and a simple slide list. End the workshop by inviting people to join the AI profit boardroom for the full step-by-step system." And the result, a complete workshop plan start to finish, with the invite built right into the ending. No babysitting needed. This is the command I use when I already know what I want. I don't want it checking in with me every 2 minutes.

[00:05:57]Want to hand it the job and walk away.

[00:05:58]That's {slash} goal. It's the go and don't stop until it's done button. Now, the third workflow. It's the opposite of that last one. The command is {slash} grill me. When you type that, the agent stops and asks you questions first. It wants to get it right before it starts.

[00:06:12]Why does that matter? Cuz sometimes you don't know exactly what you want yet.

[00:06:16]So, the agent acts like a smart team member who asks the right questions before wasting an hour going the wrong way. Here's the prompt.

[00:06:22]Grill me. Help me plan a simple email sequence that turns free followers into AI Profit Boardroom members. Ask me whatever you need about my audience, my offer, and my goal before you write anything. The result, instead of guessing, it asks you four or five sharp questions, then builds an email plan that actually fits your audience. So, now you've got two ways to work, and you pick based on how clear you are. So, here's my honest take. You don't need to become a developer. You don't need to memorize commands. You need to learn how to give a clear goal and aim a few AI workers at it. That's the whole skill.

[00:06:53]And it's learnable in a weekend. That's everything Anti-Gravity CLI is, and how you'd actually use it. Let me leave you with two things. First, if you want the full setup, walk through step by step how to install it, how to write goals that actually work, and how to build agents that bring in new customers and content while you sleep, that's what we do inside the AI Profit Boardroom. Four live coaching calls a week, and right now they're packed with people building these exact AI agent systems for their own businesses. It's a 30-day roadmap to get your first team of agents running, daily tutorials, a huge library of prompts you can copy, and a member map so you can find people near you doing the same thing. 2,800 business owners all helping each other move faster.

[00:07:33]Links in the description, or head to aiprofitboardroom.com.

[00:07:37]Come learn this while it's still early.

[00:07:38]Second, if you want the free version first, all the notes from this video, plus over 100 real AI use cases like the ones I just showed you, and the full step-by-step processes, join the AI Success Lab. It's free. You'll get the notes from this episode and access to a community of 67,000 people doing this stuff every day. Links are in the comments and the description. That's it for today. One prompt, whole team of AI workers working together in your terminal while you go live your life. Go try it.