AI News: Perplexity Personal Computer, Nemotron Super 3 + Healer & Hunter Alpha!

Okay, so let me just jump straight in because today something happened that I've been watching for a long time and the AI company just announced that it will run a computer for you 24 hours a day, seven days a week whilst you sleep, whilst you eat, whilst you literally do anything else and it never stops, it never takes a break, it never asks for a raise. This is not science fiction, that happened today on the 12th of March 2026 and it is one of five massive things that dropped in the last 24 hours that I need to walk you through right now. So, my name is Julian Goldie, if you're new here, welcome, I cover AI news every single day and I try to explain all of it in plain English because this stuff matters for everyone, not just tech people, not just coders, everyone and if you run a business, this matters. If you have a job, this matters and if you've ever worried that AI is going to change the way you work, you're right, it is and what happened today is one of the clearest signs yet.

So, let me start with the biggest story. Perplexe just announced something called personal computer. Now, I know that sounds like a normal thing, everybody has a personal computer but this is different, this is not a piece of hardware, this is not a laptop or a desktop, this is a piece of software that you install on a small computer, specifically a Mac mini and then that computer runs an AI for you all day and night. Think about what that means.

It gives an AI constant access to your files, your apps, your active sessions, running 24-7 for a small desktop computer that sits on your desk and never switches off. So, instead of you opening your laptop and typing something into an AI chatbot and waiting for an answer, you just have an AI that's already working, already thinking, already doing tasks on your behalf before you even wake up. Perplexe CEO Aravind Srinivas describes it this way, it never sleeps. That is the whole point.

Now, here's the part that's wild to me. This AI has 20 different models inside it, 20 separate brains all working together. It's connected to many different things and it's designed to be the most powerful orchestration machine ever built for a regular person or a small business. Let me explain what orchestration means because I know that sounds technical.

You know how a movie director does not do everything themselves. They have a camera, a lighting person, a sound person, a makeup artist. Well, the director just says what they want and all those different people go and do their jobs. This is exactly what this AI does.

So, for example, if you say I want to build a financial dashboard that tracks my portfolio and emails me every single week and the AI breaks that into 20 steps, sends different steps to different AI models, waits for them to finish, puts it all together and delivers the final result to you. You do nothing, it does everything. And Perplexe says this setup is actually more private than tools like OpenClaw because every single action the AI takes has to be confirmed by you. And there's a full audit trail of everything that it did.

So you can see what happened. You can check the work. You can turn it off at any time. Now, this is still early.

You have to join a wait list to get access and it's only available on Mac right now. It costs $200 a month for the full version. That $200 price tag tells you something important. This is not a toy.

The real unlock here is business shaped and teams drowning in repetitive work. Founders who cannot yet afford to hire. Operations running on spreadsheets and duct tape. If you're a founder, a freelancer, a small business owner, you want to pay close attention to this because right now you might be doing 10 hours of admin work a week, scheduling emails, doing research, building reports, summarizing documents, managing follow ups and a system like this could do all of that for you whilst you sleep.

And we're not there yet for everyone. You know, the $200 price tag is steep. The wait list is real, but the direction is obvious. In 12 months, this is going to be available for everyone, probably for like $50 or $20 or maybe even free at some level.

And when it is, the people already understand how to use this are going to have an enormous advantage over the people who are still figuring out what it is. That's not prediction. That's a pattern. Every single AI tool that's launched in the last two years has followed that exact curve.

So story number two, NVIDIA just released something called Nemetron 3 Super. I want to be very clear about why this matters, because on the surface, it just sounds like another AI model, another brain, another thing people can plug into their apps. But this one is different. NVIDIA released Nemetron 3 Super, a 120 billion parameter reasoning model built specifically for complex multi-agent applications.

Let me put that in plain English. A parameter is like a brain cell. The more you have, the smarter the model. 120 billion is a massive number, but here's the thing.

Only 12 billion of those 120 billion parameters actually running at any given moment. So think of it like a huge library. You don't read every book at once. You go to the shelf you need, grab the right book and use it.

That's how this model works. It pulls the right knowledge for the task, uses it and puts it back. That makes it incredibly fast. It delivers up to seven times higher throughput than previous models and doubles the accuracy of the last generation.

But the thing that really matters here is not the speed. It's the openness. NVIDIA is not just releasing the model itself. They are releasing everything, the weights, the training data, the research reports, the entire process.

Anyone can look at this model, look inside of it and build on top of it. That's a massive shift. And Nemetron 3 Super can run on a single GPU, one graphics card, not a data center. Think about what that means.

A developer of a decent gaming PC can now run a model that outperforms some of the best closed models in the world. From the bedroom for zero dollars per query. That's not a small thing. That's a massive shift in who can build with AI and what they can build.

And there's a third layer to the NVIDIA story I need to mention before I move on. The same day that Nemetron 3 Super dropped, NVIDIA announced a significant strategic investment in Mira Morati's AI startup Thinking Machines Lab. Who is Mira Morati? She used to be the chief technology officer at OpenAI.

She helped build ChatGPT. She helped build GPT-4. She left OpenAI in September 2024 and then started her own company. As part of the deal, Thinking Machines Lab committed to deploying at least one gigawatt of NVIDIA's next generation Vera Rubin compute systems to train its models, starting in 2027.

NVIDIA also made a significant financial investment in the company, though the exact amount was not disclosed. One gigawatt of compute. That is roughly the amount of energy needed to power 750,000 homes. That's not a small bot.

That is not a small bet. That is NVIDIA saying, we believe Thinking Machines is going to be one of the most important AI labs in the world. Thinking Machines has grown from about 30 employees a year ago to around 120 today, with more hires coming up from top AI labs than leaving for rivals. The company's stated mission is to build AI that's more understandable, more customizable, and more collaborative.

Something that individuals and businesses can actually shape to their own needs. That positioning is interesting. OpenAI, Ananthropic, and Google build powerful fixed products. You take what they give you.

Thinking Machines is betting that future wants something you can mold yourself. We'll see. But this deal tells you NVIDIA is betting the same way. Story number three, Google launched Gemini Embedding.

And I want to make sure you understand what that actually is, because the name makes it sound boring. And it's not. Gemini Embedding 2 is Google's first fully multimodal embedding model. It maps text, images, video, audio, and PDF documents into a single undefined space.

And I know what you're thinking. What is an embedding model and why should I care? Let me explain it like this as an example. So Gemini, and with this embedding model, you want to imagine that you have a library.

Not a normal library. But this library has books, photos, music, recordings, videos, and voice memos. Thousands of each. And you want to find everything in that library that's about docs.

In the old system, you would need a different search engine for each type of content. One for books, one for photos, one for videos, four separate searches, four separate tools, a lot of time. Gemini Embedding 2 puts everything into one shared space. You make one search and it finds the relevant books, photos, videos, and audio all at once.

You no longer need separate systems for different media types. And that sounds like a convenience. It's actually much more than that. Think about a company with thousands of hours of recorded customer service calls.

Right now, to know what customers are unhappy about, someone has to listen to those calls or transcribe them into text first. That costs time and money. With Gemini Embedding 2, you could just type a question. The AI would listen to all the recordings directly without transcription, understanding what's being said and give you the answer.

Early access partners are reporting latency reductions of up to 70%, meaning answers come much faster and significantly reduce the cost. The model supports over 100 languages and is available right now in public preview through Google's API and Vertex AI. Google is also rolling out a completely separate set of AI updates to Google Workspace this week. Gemini is now deeply integrated into Docs, Sheets, Slides, and Drive.

You can describe what you want, like draft a newsletter using the meeting minutes from my January HOA meeting and get a full personalized first draft in seconds. These updates are available now in beta for Google AI Ultra and Pro subscribers in English worldwide. The point I want to make here is bigger than any individual feature. So a year ago, Google was behind in the AI race.

Everyone said so. Google was slow. Google was cautious. Google was losing to OpenAI.

Now look at March 2026. Gemini Embedding 2, Gemini Workspace, new features in Docs, Sheets, Slides, Drive, a pixel drop with AI features. You can book your rides and order your groceries, new image generation tools. Google moved.

And that matters not just because Google is a company competing with OpenAI, but because Google Workspace is the tool that hundreds of millions of people use at work every day. Every AI feature they add to Docs, to Sheets, to Gmail, that reaches everyone, not just tech people, not just developers. Your neighbor who works in HR, your school secretary, your kid's school secretary, your parents running a small catering business. This is how AI goes from something tech people talk about to something that changes how everyone works.

Now let's talk about story number four. This one is fascinating and weird, and I think you're going to enjoy it. Two mystery AI models just showed up on OpenRooter this week. OpenRooter is basically a market where AI companies can put their bundles out for developers to test.

And they have a program called Stealth Models, where they test new free models on the market right now. There's no name for these models. There's no branding. There's zero explanations where it came from.

The community just gets to use it and try to figure out how to build with it. This week, two dropped at once. The first one is called Healer Alpha. It's described as a frontier omni-modal model, meaning it can see, hear, reason, and take action.

So it can natively perceive visual and audio inputs and process multiple types of data at one time. The second one is Hunter Alpha. It's described as a one trillion parameter model with a one million token context window built specifically for agentic use. And it's completely free to use right now, too.

One trillion parameters. GPT-5, one of the most powerful AI models in the world, is estimated at around 700 billion parameters. Hunter Alpha claims one trillion, and nobody knows who built it. That's the whole point.

The identity is hidden. The community tests it, runs benchmarks on it, and over the following weeks, the mystery gets solved. Previous stealth models on OpenRooter turned out to be early versions of GPT-5, new models from OpenAI's experimental labs, early checkpoints from other major companies. So when you see a one trillion parameter model drop for free with no explanation, the AI community pays attention.

The way I think about this is simple. Companies are releasing frontier-level AI models for free, anonymously, just to see how people use them. A year ago, that sentence would have sounded insane. Today, it's just a normal Thursday.

Story number five. Olama announced this week that you can now run prompts on a schedule inside Claude Code. So some of you might not be developers, and I want to make sure this lands for everyone because it matters beyond the technical world. Claude Code is Anthropic's coding tool.

It lets AI write code, fix bugs, and manage software projects. Olama is a tool that lets you run AI models locally on your own computer. What they just announced is this. The forward slash loop command runs a prompt on a recurring schedule inside Claude Code.

So you can tell it to check something, run a task, or give you a reminder, and it repeats automatically on whatever schedule you set. So you could say, for example, give me the latest AI news every morning, and it would do just that every morning without asking you ever again. Or you could tell it to monitor your Slack messages and give you a summary at 7am. Or watch your GitHub repository and automatically address feedback from code reviewers.

Or, for example, run a weekly report on your deployed apps. Set it once, it runs in the background, whether you have the app open or not. And that last part is important. These tasks are durable.

They survive restarts. You set it and genuinely forget it. Why does this matter beyond developers? Well, this is the direction that everything's heading.

The idea that you set up a task once, and an AI handles it forever, on a schedule, without reminders, without follow-ups, is going to apply to email management, customer service, social media, finance, research, and everything else. What Olama and Claude Code just shipped is an early version of that future. But it's a working version. You can use it today.

Story 6. There's a new open source tool called Paperclip that's been climbing fast on GitHub this week. And I think it deserves its own moment in this episode. Paperclip is an open source orchestration tool for running an entire company made of AI agents.

It gives those agents an org chart, a budget, goals and governance, the whole corporate structure, minus the humans. Let me explain what that means. Right now, when people use AI agents, for example, tools like OpenClaude or Claude Code that can work autonomously, they usually run one agent at a time. You tell it what to do, it does it, you check the result, you give it the next task.

And that's like hiring one employee and managing themselves. But Paperclip is asking a different question, which is, what if you had a CEO agent, a CTO agent, a marketing agent, a developer agent, a designer agent, all working together, coordinated through an org chart, with defined responsibilities and a shared company mission? You set the mission. So for example, build the number one AI note-taking app to $1 million in monthly revenue.

And the agents figure out the tasks, delegate them to each other, track progress, stay within budget and report back. The project already has over 14,000 GitHub stars and is trending across multiple AI communities in just its first week. And there is already a real world example of this kind of thing working. Nat Eliason built an AI agent called Felix using OpenClaude.

Felix operates autonomously, earning over $100,000 in revenue with a target of $1 million, handling content creation, research and business development without daily human oversight. Eliason talked about this publicly in March 2026. One human, one AI agent, $100,000 in revenue. I want you to sit with that for a second.

Not because I think everyone should fire their team and replace them with AI. That's not the point. The point is this. The bar for what one person can build and operate with AI assistance is rising dramatically.

Month by month, week by week. Now, just to be clear here, those numbers could be exaggerated. They're not fully verified. But I want you to understand the trend, which is that these AI agent companies are growing massively.

Even on TrustMR, for example, there's over $388,000 per month being generated just through OpenClaude wrappers, which are non-human AI agent teams, right? They're companies where they just resell OpenClaude access to people who are not technical enough to set it up. So think about this. Two years ago, a one-person business was capped at what one person could physically do.

Now, one person with the right AI tools can run something that used to require a team of five or ten. That's the economic story underneath all of this. 100% of enterprises surveyed by crew AI say they plan to expand their use of AI agents this year. 65% already using AI agents in production.

Organizations report they have automated an average of 31% on their workflows. 31%. Almost a third of the work that used to require humans is already being handled by AI in these companies. And they are just getting started.

Now, let me step back for a moment and talk about something that I think ties this all together. Because when you look at everything that just dropped in the last 24 hours, perplexity always on computer, right? The personal computer. NVIDIA's free open source model that beats the best closed models.

Google's embedding all media types into one AI system. Mystery, trillion parameter models, automated AI schedulers, and a tool to run entire companies with AI agents. It's tempting to think this is just a lot of separate news stories. It's not separate.

It's one story. And the story is this. AI is moving from a tool you use occasionally to a system that runs in the background of your life and your business all the time. A year ago, you opened a chatbot when you had a question.

Now, AI can run a computer for you whilst you sleep. Now, AI can watch your code and fix problems before you wake up. Now, AI can run an entire company of agents working towards a shared goal. The gap between people who understand this and people who do not is widening every single week.

And I'm not trying to scare you here. Genuinely, I'm not. I talk to a lot of people who get anxious when they hear this kind of stuff and I understand why. The pace is dizzying and it feels hard to keep up.

But here's the thing I keep coming back to. The people in the AI Profit Boarding, the community I run for exactly this purpose, they're not anxious about this. They're excited because they have context. They understand what's happening, how it works, and how to use it to build a better business or better career.

Their knowledge is not that complicated. It does not require an engineering degree. It doesn't require 10 hours a day. It requires the right community, the right practical training, and the willingness to stay engaged.

That's what the AI Profit Boarding is. It's a place where real people, creators, founders, marketers, business owners, learn how to actually use this stuff. Real workflows, real results, and real money saved and made. So if you want in, the link is inside the comments inscription, or you can go to the AIProfitBoarding.com and come and join us.

Now, let me go a level deeper on something I think people are missing in all of this news. There's a pattern across every single story today. Perplexy's personal computer is not just about doing tasks faster. It's about the idea that AI is becoming the operating system.

Not just a feature inside your computer, but a thing that runs the computer itself. NVIDIA's Nemetron release is not just about open source. It's about flipping the control of who owns the most powerful AI tools. Right now, the biggest companies own the most powerful models.

NVIDIA just put something that competes with them into the hands of anyone with a computer. And Google's Gemini embedding update is not just a developer feature. It's about AI becoming able to understand the real world. Not just text, but everything humans create.

Video, audio, images, documents, all at once. And Paperclip is not just a project management tool for AI agents. It's a blueprint for a completely new kind of business. One where human owners set the mission and AI employees now do the work.

Now, I want to be fair about what's hard here. So these tools are still early. Perplexe's personal computer is on a waitlist and costs $200. Paperclip is an open source project that requires a technical setup.

The stealth models may or may not be what they claim. So not everything is going to work perfectly. There will be mistakes. There will be hallucinations.

There'll be moments where the AI does something dumb and you have to fix it, right? Anyone who tells you AI is already flawless is selling something. But here's what I keep coming back to. The direction is undeniable.

ChatGPT now has 900 million weekly active users. That's over 10% of the global population using it every single week. A year ago, it was 400 million. That is 500 million new users in 12 months.

And that's not a niche technology anymore. This is the internet in 1997. This is the smartphone in 2009. The people who figured out early what the internet was good for built incredible businesses.

The people who figured out early what smartphones were good for built incredible apps and careers. And we're in that moment right now with AI and the window is still open. Now, let me talk about something that I know a lot of you are thinking about, but maybe haven't said out loud. What does this mean for my job?

And I get this question all the time. I want to give you a real answer, not like a corporate speak answer. Here's the truth. Some jobs are going to change dramatically.

Jobs that are mostly repetitive. Jobs that are mostly information processing. Jobs that are mostly coordination and scheduling. Those are going to look very different in three years.

And that's uncomfortable to say, it's true. At the same time, and this is the part I really want you to hear, the people who are learning AI, who are getting comfortable with these AI tools, who are figuring out how to combine their human judgment and creativity with AI execution power. Those people are not threatened. They're supercharged.

Think about what that means to be a marketer who understands how to use AI, to do the work of a five person team. Or a lawyer who uses AI to do research in minutes that used to take days. Or a small business owner who uses Flexi's AI computer to handle their admin whilst they focus on clients. Those people are not worried about AI.

They're ahead of AI, right? And the gap between the people who are ahead and the people who are not is the fastest winding gap I've ever seen in any industry. I'm not trying to scare anyone. I'm just trying to wake people up, right?

Because right now in March 2026, there's still time to get ahead of this. There's still a learning curve that is accessible to normal people. And the tools are still early enough that being an early adopter gives you a real advantage. But in 12 months, 18 months, that window is going to be much smaller.

Now, let me wrap this up by talking about what you can actually do right now. If you're a complete beginner and this all feels overwhelming, start simple. Pick one AI tool. Could be ChatGPT, Claude, or Gemini, doesn't really matter which one.

And use it for one hour daily to do something in your job or business that you currently do manually. Could be writing emails, could be summarizing documents, could be planning a meeting. I would just start there. You'll be surprised by how fast it changes your thinking about what's possible.

If you're already using AI regularly and you want to go deeper, look at the stories from today. Perplexes, personal computer, paperclip, alarm and scheduled tasks. These are tools for people who are ready to move beyond basic chatting and into actual automation. If you're a business owner, the question you need to ask this week is not should we use AI?

The question is already answered, right? The question is, where is the 30% of our workflow that AI should be handling right now and is not? Because if you do not answer that question, your competitor will. And if you're a developer, Nemetron 3 Super just landed, it's free, open source, 1 million token context window, runs on a single GPU.

If you haven't looked at it yet, you should today. The thing I keep coming back to after covering AI news every single day, after watching every model launch, and every product announcement, and every funding round is that the people who are going to do the best through this usually aren't the smartest. They're not the most technical. They're not the people with the most resources.

They're just the people who stay curious, who keep showing up, who are willing to learn something new every week, even when it's uncomfortable. Who understand that in a world changing this fast, the biggest risk is standing still. The AI industry is moving at a speed no one fully predicted over 18 months ago. The tools dropping today, an always on AI computer, a trillion parameter mystery model, a tool to run a zero human company.

These were science fiction conversations just two years ago. Now they're headlines. Now they're the products you can download. Now they're the companies being built by real people making real money.

That's where we are. And the people who recognize that and take action on it, not next year, not when it gets easier, right now. Those are the people who are going to look back on 2026 as a year, everything changed for them. I hope you're one of them.

Thanks for being there. That's your AI news for today. If you haven't already, check out the AI Profit Boardroom, link in the comments description, or go to the AIprofitboardroom.com. And this is my AI automation community that teach you how to learn, scale, and grow with AI automation.