Claude Code Channels + Google AI Studio + Perplexity Health + XiaoMi (AI NEWS)

Claude Channels, plus Google Studio Update, plus Ploplexi Health, plus OpenAI and Astral, and the new Xiaomi AI agent. This is the AI news today, and something happened today that I think a lot of people are going to miss, right? Not one thing, five things. Five separate announces dropped in the last 24 hours that all point in the exact same direction.

AI is not just getting smarter, it is getting closer, it is getting into your phone, into your health records, into your coding tools, into your body data. If you're a business owner, a freelancer, a creator, or anyone who works for a living, today's news matters to you more than anything else you will read this week. So let's go through all five. Let's start with the one that most people are completely missing this morning.

You've probably heard of OpenClaude by now. If not, it's just an AI agent that runs 24-7 on your computer. It exploded this year. Developers everywhere started dealing with it.

And the company that built the most popular AI coding tool in the world, Anthropic, the people behind Claude, watched that happen. And today, they responded. Anthropic just launched Claude Channels. His what that means in plain English.

The war over AI coding is the biggest war in tech right now. OpenClaude, sorry, OpenAI has Codex, 2 million weekly users, usage up five times since January. Anthropic has Claude Code. Developers love it.

Reasons better than almost anything out there. But both of them had the same problem. They lived on your computer. They needed you to be sitting at your desk, staring at your screen, actively managing sessions to actually use them.

And that was the ceiling. And OpenClaude, because it was built from the ground up as an AI agent framework, was already pushing past it. OpenClaude was built to run tasks on its own in the background without needing you to babysit it every single second. And Anthropic just answered that directly with Claude Code Channels, which lets you push Telegram and Discord messages directly into a live Claude Code session.

So you can keep sessions running on your machine on the web and react to external events whilst you're away. And it's built on MCP with channel support. Right? So you can start a coding job.

You can tell Claude to build something. You can walk away, go to lunch, go to the gym, you know. And whilst you're doing all of that, your AI is still working, still building, still coding on your phone, on Telegram, on Discord, wherever you are. And here's why this matters in the context of OpenClaude specifically, right?

OpenClaude became so popular because it gave developers a way to run agents that did not need constant babysitting. That is a model people want. That is a model that feels like having a real assistant instead of just a fancy autocomplete tool. Claude Code Channels is Anthropic saying, we can do that too, except we're not going to make you learn a whole new framework to do it.

We're going to bring it to the tools you already use every single day. Telegram, Discord, your phone. And this is exactly why Anthropic built channels, because the model wars and the agent wars are the same war now. If your model cannot run as an AI agent, it doesn't matter how smart it is.

If your agent cannot stay connected to when you walk away, it's not actually useful yet, right? And Claude Code closes that gap. It's in research preview right now, which means it's early, which means the people who learn this stuff today are going to be miles ahead of everyone else who waits, right? I keep saying this, the gap between people who understand these tools and people who don't, it's not staying the same size.

It's getting bigger every single week. And this is one of these weeks. Okay. Story number two, Google completely rebuilt its AI coding platform.

Most people have no idea this happened. Now, tools like Lovable, Bolt, Repl.it, they all do a form of vibe coding, right? Where you can just describe what you want and then you can build what you want. And they've been growing insanely fast, right?

And now Google watched all of this happen. And on March 19th, 2026, yesterday, Google completely overhauled its AI studio platform. Google was launching a completely upgraded vibe coding experience in AI studio designed to turn prompts into production ready platforms from multiplayer experiences and installing external libraries to ways to save progress and log in securely. So you can now truly build functional full stack apps, right?

That's the official Google announcement. Production ready. Not just a demo, not just a prototype, but a real app with a real backend. And here's what's actually new.

The core addition is Google Antigravity. So this is a coding agent that handles real-time multiplayer functionality, database provisioning, and third-party library installation, right? It's a coding agent that just works. Now, previously vibe coding tools you could build with Antigravity, but it didn't have that backend functionality.

Now, when you're building an app in AI studio, the agent will proactively detect when you need data storage and user identity based on your queues and your prompt, and it uses something called Firestore and Firebase authentication with your approval, right? So you just build the app, you click enable Firebase, the agent sets up your Firebase project, provisions Firestore, enables authentication and connects your app to a backend with one click, right? And the full app is ready to go. So people can log in, they can sort of message each other, they can check out each other's data, they can see their data and log in and come back tomorrow and it's still there, right?

Now, Google showcased several example apps built entirely through prompts. So there was a multiplayer laser tag game with real-time leaderboards. There was a collaborative 3D particle visualization using 3DS and a recipe organizer with Gemini powered generation. So this new experience in Google AI studio has already been used internally to build hundreds of thousands of apps over the last few months by Google itself.

And here's the part that I think matters for you mostly, right? Google AI studio is a browser-based AI development platform that allows users to build full-stack applications using plain language prompts. Browser-based, right? It's free to start, no downloads, no setup, you just go to AI studio and you start describing what you want, right?

This is for non-developers who just want to build whatever they want. And a year ago, if you wanted to build a real working app with a database and user logins, you needed a developer, you needed a lot of money, you needed tweaks, right? Today you need a browser and a quick tech description. That is a category shift and Google just made it free.

Quick break. If you're watching this and thinking I need to understand exactly how to use these tools, not just watch someone talk about them, that is exactly what the AI Profitable Boarding is for. It's my private community where we go deep on the practical stuff, how to use these tools in your actual business, how to build things, how to make money, real tactics, not just news coverage. You can check it out, link in the comment description or just go to the AI Profitable Boarding.

Let's get back to the news. Story number three. This one is personal, like actual personal. Perplexe launched Perplexe Health and this is the one that made me stop and just sit with it for a minute, right?

Here's the problem Perplexe is trying to solve. Your health data is everywhere and nowhere at the same time. Your fitness stuff is an app or health or Fitbit. Your lab results are in some hospital portal you log into twice a year and can never remember the password for.

Your prescriptions are somewhere else. Your doctor's notes are probably somewhere else, right? And your blood work from three months ago is in a PDF you downloaded and probably lost as well, right? So all the information exists but none of it talks to each other.

And when you try to ask a question about your own health you are either googling generic information that has nothing to do with the actual situation or you're waiting three weeks to see a doctor for a 15 minute appointment. Today Perplexe is launching Perplexe Health, a suite of connectors to personal health data that allows for personal and powerful personal uses of Perplexe computer, right? So it gives you accurate health information, all of it in one place. Now this connects to Apple Health, electronic health records from over 1.7 million care providers, platforms including Fitbit, Ultra Human, Withings, Clue and many more.

Integrations with Aura Function are also expected soon. And so Perplexe Health can actually track your metrics and trends over time across biomarkers and activity data for a personalized dashboard. You can ask it a health question and the answer draws from your medical records and wearable data at once. So it could be like a question about resting heart rate for example, which can factor in your recent activity, your cardiac history and your latest blood work.

Your resting heart rate question then gets answered using your recent workouts, your cardiac history going back years and your blood work math from last month, all at the same time, all personalized to you. It's not a Google search, it's not a generic WebMD article written for everyone and therefore it's actually useful because it's your data, right. And so here's the context that makes this even more important, right. Microsoft launched Copilot Health on March 12th, joining OpenAI's ChatGPT Health and you've got Amazon Anthropic who are also doing this sort of stuff.

So this is a race and the finish line is being the AI that knows the most about you personally, so it can be the most useful to you personally, right. And so what Perplexe is about is that they can do better than the search engines that have been optimizing for ad clicks for health care related data for 20 years, right. And answers in Perplexe Health as well, they also draw from premium medical literature including clinical guidelines and peer-reviewed journals, right. So each response includes citations linked directly to the source material so you can read what the answer is based on.

Right now most health answers online are grounded in like SEO optimized content to rank. SEO optimized content to rank is just like, you know, garbage written by people who want clicks, not accuracy. Whereas Perplexe is thinking, you know, people want something more personalized, people want something ready to go. So right now Perplexe Health is ready for pro and max subscribers in the US.

It lets you connect data from medical records, wearables, wellness apps, and Apple Health so it can provide you context-aware answers to your health questions. Story number four, OpenAI just bought a company called Astral and if you're a developer you might be tempted to skip past this one or if you're not a developer, sorry, but do not because here's a quick context. OpenAI has a product called Codex. It is their AI coding tool.

Think of it as like their version of core code and it's been growing really fast. It's already seen three times user growth and five times usage increase since the start of the year. It has over 2 million weekly active users, right. And OpenAI wants to make sure that momentum doesn't slow down because the coding tool war is the most important war in AI right now, right.

Developers are the people who build everything else. So they're the ones who integrate AI into products. They're the ones who create the tools. Everyone eventually uses it and if you win over developers, you win the future.

So by acquiring Astral, they bring open source developer tools into the Codex ecosystem, right. And they're not just buying a company, they're buying a presence, they're buying trust. It's buying a direct line into the daily workflow. Millions of developers already use these tools every single day.

And this is not just about building better AI models anymore, it's about ownership of the developer environment itself. It's like this, right. Imagine your word processor also own the spell checker, the grammar tool, the formatting library, and file saving system. At some point you stop thinking of those as separate things, right.

They are just your word processor. They're just how writing works. This is what OpenAI and Anthropic are both trying to do with that coding. They're trying to become so embedded in how developers work that switching away feels like learning to write with your other hand.

And the race is real, it's accelerating. The competition between Anthropic's code and OpenAI's Codex is fierce. Those two $200 per month subscriptions add up to billions of dollars per year in revenue for companies that very much need that money. So this is not a features competition anymore, it's a land grab and the Astral acquisition is OpenAI planting a flak.

Now if you're a non-developer, here's what this means. Every tool you use that has AI in it, that AI was probably built by a developer using one of these coding tools. The faster and better those coding tools get, the faster and better everything else you use gets, right. So the war between core code and Codex is not a developer story, it's everyone's story because developers build the world that you live in and the AI helps them build what's shaping that story, okay.

The final one is that there was a new model called Hunter Alpha, right, and it was it was completely stealth. Nobody knew what it was or who it was, but it was a powerful model that everyone was loving it. People were thinking it's DeepSeq and it actually surpassed 1 trillion tokens in total usage and climbed to the top of open routers leaderboards, right. Now people were thinking, ah maybe it's DeepSeq, maybe it's for example new OpenAI models, but actually the reveal came and Xiaomi's AR model team, MIMO, actually run by former DeepSeq researcher Luo Fuli, announced that Hunter Alpha was actually an early internal test build of MIMO V2 Pro from Xiaomi.

Not DeepSeq, Xiaomi, the company you might know for making affordable smartphones and electric vehicles, right. And so Xiaomi built a trillion parameter AI model specifically to power agent frameworks like for example OpenCore and they are giving developers free access to it for a week, right. You can actually use it inside Kilo Code and OpenCode this week for free. And think about who is building frontier AI models right now.

OpenAI, Anthropic, Google, Meta, those are the names everyone knows, those are names everyone expects to be at the frontier. Xiaomi is not on that list. Xiaomi is a phone company, it's a car company, except now they have a model that tops the open router usage charts, processed over a trillion tokens in one week and it's directly competing in the agent framework with the best models in the world. A phone company.

That's a story that should make everyone sit up straight. Okay, so let me pull together all these stories because I think they're telling one, right. AI is moving from being a thing you go to, to a thing that comes to you. The old model was you open a computer, you go to an AI tool, you type a question, you get an answer, you close it, done.

New model is your AI is running in the background whilst you live your life. It texts you when it needs you, it knows your health data, it builds apps for you, it works inside your coding tools for you if you ever open them. And it's powered by models from companies you would never have predicted were in the AI race at all. Now the last thing I want to say is there's a version of watching today's news where you just feel overwhelmed, right.

Five things happen, they're all big, they all matter, and there are five more tomorrow and five more the day after that. And I get that feeling, I genuinely do. Most of the people I talked to in the airport for warning felt exactly that way when they started learning this stuff systematically. The overwhelm is not because there's too much happening, the overwhelm is because you don't have a framework yet for processing what matters and what does not.

Once you have a framework, today's news doesn't feel like chaos, it feels like evidence. Evidence of direction, evidence of where things are going, evidence that helps you make better decisions about your time, your money, your career, and your business. And this is what I'm trying to give you inside each of these episodes. Not just a framework, sorry, not just the news, but the framework, right.

The world is changing faster than most people realize. The only thing I know for sure about times like this, having watched a lot of them in a row, is that the people who stay informed, stay curious, and keep learning are the ones who come out the other side in a better decision than when they started. That's the whole point of this channel, that's the whole point of the AR Profit Warning, and that's why I'll be back tomorrow with whatever comes next, because something always comes next. See you then, thanks for watching, and if you haven't already, check out the AR Profit Warning, link in the comment description, or just go to thearprofitwarning.com.

Cheers.