Big Tech has paraded AI models for years, like they were the second coming of the internet. Neural networks, deep learning, transformers, chain-of-thought, and more recently, jargon like chain of draft, model distillation, AI-first operating systems and much much more.
But here’s the dirty little secret.
The model was never the product.
It was the bait.
It was that shiny thing dangled in front of investors. The mirage promising an all-knowing digital oracle.
But now, the conversation has shifted.
Because the real power is the system.
The architecture.
The agents, apps, and entire ecosystems that think, and do.
They act.
They execute.
They replace.
The AI model is just the brainstem. The real beast is what happens around it. And it’s about to consume everything.
If you like my rants and want to support me:
- Comment, or share the article; that will really help spread the word 🙌
- Connect with me on Linkedin 🙏
- Subscribe to TechTonic Shifts to get your daily dose of tech 📰
- TechTonic Shifts also has a blog, full of rubbish you will like !
AI models were useless until they grew arms and legs
Remember when ChatGPT and its cousins first hit the scene? They could generate paragraphs of text, answer your questions, even pretend to be Shakespeare. But ask them for real-time stock prices, flight schedules, or anything requiring a fresh scrap of reality, and they would fumble.
The problem was that they were mere parrots.
They were repeating only what they had been trained on.
They were locked in time like an encyclopedia printed last year.
Then something changed.
LLMs got tools.
They learned to check facts as they worked.
Well, kinda.
They reached out to the web, pulled in real-time data, even cited sources like smug little know-it-alls. OpenAI, Anthropic, Google, and the rest gave them arms, gave them legs, let them roam beyond the model. And suddenly, they weren’t text generators anymore. They became systems.
Operators.
Orchestrators of information.
This isn’t just about answering questions. These models are beginning to control the digital world itself. OpenAI’s Pro version with Operator and Claude’s Computer Use, can interact with your PC, search files, summarize documents, execute commands.
You give them access to your calendar, your emails, your task manager, and suddenly, they’re not assistants.
They are managers.
Meet Manus, your new AI buddy (but, like, in the cloud)
And more recently we have Manus, and that’s a system built by a small Chinese company called Butterfly Effect. It’s not another AI chatbot like DeepSeek was.
No, this thing is different.
It’s not one intelligence.
It’s a swarm.
Manus is a hive of AI agents, and all are working under a single “executor” AI.
Think of it like CEO with its nose all white, managing an army of unpaid workers. It takes a task, breaks it down, assigns subagents, and lets them loose on whatever your problem is.
Some focus on research.
Others analyze data.
Some write reports, others handle automation.
A few even generate code and deploy it. It’s like a full-blown company running itself, except no one is asking for a raise.
And the nice thing is that it runs completely autonomously in the cloud.
No human oversight.
No need for you to babysit it.
You give it a task, then walk away while it works. If that doesn’t send a shiver down your spine, congratulations, you’ve already made peace with your robotic overlords.
And yes, you could’ve seen this coming if you read this:
Anticipating AI’s next move • article ② • | LinkedIn
Anticipating AI’s next move • article ③ • | LinkedIn
Rise of the machine customers | LinkedIn
Especially pay attention to the AI-native companies that are on the rise. They are quietly starting up, but the real AI revolution will be in this space.
Do you want to know how this will pan out? Watch this video:
The system Is the power. Not the model
Here’s the thing you need to know. Manus isn’t a secret, ultra-advanced AI model. It’s not even using its own. It’s just calling up Claude 3.5 Sonnet from Anthropic and Qwen from Alibaba. Those are both models that anyone with an API key can access.
So their magic isn’t in the model.
The magic is in the system.
The network.
The orchestration.
The way that these agents coordinate, fetch data, fact-check, process workflows, and execute commands in a way that no single AI model could ever do alone.
That’s the real future.
Not better models.
But better systems.
SaaS is dying, and vertical AI is eating its corpse.
For years, businesses have been addicted to SaaS – Software as a Service. You needed a CRM, you bought Salesforce. You wanted some Email marketing, you got Mailchimp. Accounting, QuickBooks.
Every problem had its SaaS tool, and each costing a neat little subscription fee.
But now?
That is looking kinda ancient.
Because vertical AI is here.
And it’s not another SaaS tool.
It’s a full replacement for entire teams of people.
Just think of an AI agent that automates a task, AND it owns a function.
It has a real job.
A legal agent that drafts contracts, negotiates deals, and files paperwork. A financial AI that manages budgets, forecasts profits, and executes trades. A marketing AI that analyzes customer data, generates content, and optimizes ad spending.
All on its own.
This is bigger than SaaS.
SaaS gave you tools. Vertical AI does the work itself. Instead of juggling 50 SaaS subscriptions, you get one AI system that just handles everything.
Business is changing. The workforce won’t survive it.
Mistral CEO Arthur Mensch sees the writing on the wall. He’s been talking to enterprise clients, and guess what he is saying.
He doesn’t care about “cool AI features” anymore.
Companies don’t want flashy pilots.
They want ROI.
Companies are done playing around with AI as a gimmick. They want systems that actually make them more money. That means AI is not an “assistant” anymore.
It is now running operations.
Mensch says the transformation is obvious, the biggest pain in AI isn’t the model. It’s everything around it. The governance. The security. The way data flows between departments. The businesses that get this, the ones that build complete AI-powered systems, are the ones that will win.
The ones that don’t, well, they’ll be the next Blockbuster.
AI won’t answer questions. It’ll ask you the questions.
In a few years, once all the problems have been ironed out, and the frameworks are enterprise ready, every department will have its own AI agents. They will handle research, write reports, build marketing campaigns. But the gnarly thing is that humans won’t be asking AI for information anymore. AI will be asking humans.
Think about that.
Think hard, and long.
You won’t be logging into a system to pull sales data.
The AI will ask you for strategic input.
If the company needs a final sign-off, the AI will ping you.
When it needs feedback on a proposal, the AI will send it to the right person.
At some point, you’re not managing AI anymore.
AI is managing you.
That, my intellectual friend, is the SaaS of the near future.
The AI revolution is here. You’re just in the way.
So here we both are.
AI isn’t a tool anymore.
It has become a system.
A force.
To be specific, a silent, tireless workforce that doesn’t sleep and doesn’t get paid. It only sucks energy and water. But it never ask for vacation time.
Tech leaders don’t dream of a world where AI makes our lives easier. They dream of a world where humans are optional.
This isn’t sci-fi. This isn’t the future. It’s happening right now.
The only question is, when the AI starts running the show, will you even notice? Or will it already be too late?
Signing off from the inevitable.
Marco.
Oh? Did I ask you to take the survey?
I build AI by day and warn about it by night. I call it job security. Stick around if you like what I write. If not, don’t worry, the AI already already knows you were here.
To keep you doomscrolling 👇
- The AI kill switch. A PR stunt or a real solution? | LinkedIn
- ‘Doomsday clock’: it is 89 seconds to midnight | LinkedIn
- AIs dirty little secret. The human cost of ‘automated’ systems | LinkedIn
- Open-Source AI. How ‘open’ became a four-letter word | LinkedIn
- One project Stargate please. That’ll be $500 Billion, sir. Would you like a bag with that? | LinkedIn
- The Paris AI Action summit. 500 billion just for “ethical AI” | LinkedIn
- People are building Tarpits to trap and trick AI scrapers | LinkedIn
- The first written warning about AI doom dates back to 1863 | LinkedIn
- How I quit chasing every AI trend (and finally got my sh** together) | LinkedIn
- The dark visitors lurking in your digital shadows | LinkedIn
- Understanding AI hallucinations | LinkedIn
- Sam’s glow-in-the-dark ambition | LinkedIn
- The $95 million apology for Siri’s secret recordings | LinkedIn
- Prediction: OpenAI will go public, and here comes the greedy shitshow | LinkedIn
- Devin the first “AI software engineer” is useless. | LinkedIn
- Self-replicating AI signals a dangerous new era | LinkedIn
- Bill says: only three jobs will survive | LinkedIn
- The AI forged in darkness | LinkedIn
Leave a Reply