Rebel Technologist

Most people are using the most powerful cognitive tool
in human history to fix their email subject lines.

That's a tragedy.

Because what AI is actually doing (quietly, underneath
all the productivity hype) is something that hasn't been
possible at scale since the 15th century.

It's turning determined, curious, adaptable people into
genuine polymaths.

Renaissance Men and Women who can build, create, and
develop real skills across disciplines they never thought
were available to them.

Da Vinci painted the Mona Lisa, designed flying machines,
mapped the human body, and wrote treatises on hydraulics.

We call him a genius.

But what he actually had was ACCESS.

Access to knowledge across disciplines. Time to pursue
multiple paths. The intellectual permission to be more
than one thing.

The industrial era took all of that away.

Pick a lane. Specialise. Repeat until retirement.

AI just tore up that contract.

The barriers to entry across almost every discipline have
collapsed. Simultaneously. In three years.

Coding. Music. Writing. Design. Finance. Architecture.
Law. Biology. Whatever lights you up.

Every single one now has an infinitely patient expert
who will meet you where you are and walk with you as
far as you want to go.

This isn't AI doing things for you.

This is AI giving you back the parts of yourself you
shelved. And the tools to explore the parts you've
never been yet.

The only question is whether you have the nerve to stop
optimising your existing self and start building a new one.

Da Vinci didn't wait for permission.

Neither should you.

1 week ago (edited) | [YT] | 1

Rebel Technologist

Most companies will tell you they're doing AI.

Ask them where. They'll point to a chatbot that summarises emails and a Copilot licence nobody trained anyone to use.

That's not agentic AI. That's AI as a productivity supplement. Like a vitamin you take but don't absorb.

A December 2025 HBR survey of 325 enterprise leaders found that only 12% of organisations have AI embedded inside the actual flow of work. The majority (39%) are still running standalone tools alongside their processes rather than inside them.

This is the 12% problem.

And it's not a technology problem. It's a design problem. A governance problem. An organisational courage problem.

The companies clearing 12% aren't just using better models. They're doing something different. They've rebuilt workflows around the assumption that AI participates, not assists. They've defined what the agent can decide, what it must escalate, and who is accountable when it gets it wrong.

That accountability piece is the tell. Most organisations skip it because it's uncomfortable. Defining accountability means admitting that agents can fail. And that someone will be responsible when they do.

MIT's Thomas Malone frames it plainly: if an agent takes an action, someone in the organisation should be accountable for that action. He compares it to owning a dangerous animal. You're responsible for what it does.

Most enterprises are releasing the animal without reading that part of the contract.

The 12% have read it. They're building accordingly.

The rest are still updating their LinkedIn bios to say they're AI-first.

3 weeks ago | [YT] | 3

Rebel Technologist

Here's the uncomfortable truth about your new AI agent: It's going to break by Tuesday.



Building a skill is easy. Building one that survives contact with reality is the real challenge. While you're still patching fragile scripts every time an agent hallucinates, the smart money is building resilient systems.



Perplexity just dropped their internal methodology for agentic workflows. The core rebel insight: Stop starting with the prompt. Start with the tests.



𝗧𝗵𝗲 𝗣𝗿𝗲-𝗕𝘂𝗶𝗹𝗱 𝗕𝗮𝗰𝗸𝗹𝗼𝗴: Don't write instructions until you have the test cases. Pull 10 to 20 real queries from your production logs to create a hero query set. Find five the agent should handle flawlessly, and five neighbor queries it must refuse. The gap between those sets is your backlog.



𝗜𝗻𝘁𝗲𝗻𝘁-𝗕𝗮𝘀𝗲𝗱 𝗧𝗿𝗶𝗴𝗴𝗲𝗿𝘀: Phrase your routing descriptions the way your team actually talks. Instead of robotic commands like "monitor pull requests," use phrases like "babysit a PR" or "make sure this lands." Keep these descriptions under 50 words so the system adapts to human users. If it requires a specific command, it is already broken.



𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲-𝗗𝗿𝗶𝘃𝗲𝗻 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻: The model already knows how to execute basic commands. It needs direction on your specific intent. Tell it to "preserve intent while resolving conflicts" rather than listing 15 literal steps to click through a repository. You are guiding a reasoning engine, not writing a Python script.



𝗧𝗵𝗲 𝗙𝗮𝗶𝗹𝘂𝗿𝗲 𝗙𝗹𝘆𝘄𝗵𝗲𝗲𝗹: The highest-value asset in your system is the failure log, or what Perplexity calls the gotchas. When the agent screws up in production, that mistake becomes the new guardrail. Turn every off-target load or hallucination into a standing instruction that inoculates the system against repeating the error.



𝗧𝗵𝗲 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗧𝗮𝘅: Ask yourself if the agent would fail without a specific line of instruction. If the answer is no, cut it immediately. Every extra word is a context tax that slows the system down and confuses the model. If a task is easy to explain, the model already knows it.



Stop treating agents like interns you have to micromanage. Start architecting them like systems that learn.



Your agents will figure out the work. Your architecture decides if they survive the real world.



(Full playbook here: research.perplexity.ai/articles/designing-refining… )

1 month ago | [YT] | 0

Rebel Technologist

There is no 'wait and see' strategy in 2025. There's only 'move now' or 'cut later

The market isn't waiting for certainty. It's a relentless force, and in 2025, it presents a stark, binary choice for every business leader.

Here's the flip:

- Side A: Invest in Agentic AI now. Integrate with your legacy systems. Accelerate your teams. Capture the revenue growth that's available in your market.

- Side B: Wait for more certainty. Watch competitors execute. Watch them capture your margin. Then cut costs desperately just to survive.


There is no third option.



The notion of a "wait and see" approach? That's just Side B in slow motion. The market's velocity is too great. Examples of explosive growth from early automation are multiplying. Every quarter of hesitation means margin transferred directly to someone faster, hungrier.

The choice isn't whether to automate; it's whether you do it from a position of strength or desperation.

Here's what you need to realise:

1. Proactive Investment: Agentic AI isn't a future concept. It's here, ready to integrate with your existing infrastructure, not replace it.

2. Team Acceleration: Empower your people, don't replace them. Agentic AI frees them from the mundane, allowing them to focus on high-value, creative work.

3. Revenue Capture: Early movers are already seeing significant gains. This isn't speculation; it's happening now.

Remember this:

✅ "Hesitation isn't strategy. It's just slow-motion defeat."

✅ "The margin isn't waiting for you to feel ready."

This is the binary choice of 2025. The coin won't spin forever. It will land. Call it.

P.S. That spinning coin? It's not a gamble if you understand the odds.


#AgenticAI #BusinessStrategy #Innovation #FutureOfWork #Leadership

7 months ago | [YT] | 0

Rebel Technologist

I call bullshit on your AI-driven layoffs.

Every earnings call sounds the same lately. "We're strategically reducing headcount as AI automation enables greater efficiency."

Translation: "We found a trendy excuse to do what we wanted anyway."

Let's talk infrastructure reality.

Most companies can't even automate their expense reporting, yet they expect us to believe they've automated entire departments?

Their digital transformation roadmaps measure timelines in years, not months.

Their API strategies are PowerPoints, not platforms.

Their legacy systems require teams of consultants just to keep the lights on.

💡 You can't automate what you haven't digitised.

💡 You can't digitise what you haven't understood.

The companies genuinely deploying AI at scale—Google, Meta, Microsoft—they're not hiding behind vague announcements.

✅ They're demonstrating specific use cases.

✅ They're showing measurable productivity gains.

✅ They're genuinely replacing certain functions with automation.

The infrastructure is real.

The rest? Corporate theatre using AI as a convenient scapegoat for restructuring they should have done years ago.


P.S. The real AI revolution is built on code, not corporate spin. Don't fall for the theatre.

8 months ago | [YT] | 1

Rebel Technologist

AI in 2024: Beyond the Hype!
Everyone's talking about AI, but what does it really mean for your business in 2024? Find out in my recent article, "AI's Real Deal: 24 Predictions for 2024 to Cut Through the Hype!" From game-changing predictions to practical advice, I've got you covered. Join the conversation and be prepared for the future. #AIBusinessTrends #TechInsights #FutureReady

www.linkedin.com/pulse/ais-real-deal-24-prediction…

2 years ago | [YT] | 0

Rebel Technologist

Where do you spend most of your time?

3 years ago | [YT] | 1

Rebel Technologist

Wake up call !

1. The least time you get is with your family, so spend time with them.
2. I worry about this graph as there is no time spent with Mates, make time for friends too, not just on Facebook.
3. Marry your best mate, you gonna spend your old age with them, have fun doing so, don't be a burden on each other and look forward to it
4. Finally, get comfortable with yourself; you are awesome and enjoy the time with you !

3 years ago | [YT] | 1