The AI Overlords (A Contrarian Reflection)
Foreword: This is an exploration of ideas. A dive into First Principles thinking. A contrarian take on AI and Power. NOT the absolute truth, but one possible version of the future I’ve been reflecting on. Because in “the age of AI,” it’s often the uncomfortable questions we avoid asking that hold the greatest potential to come back and haunt us.
This is a conversation worth having.
Let’s talk about the future.
NOT the TED Talk hype.
NOT the typical LinkedIn buzzword bingo.
NOT the utopian AI “savior of humanity.”
And NOT the dystopian “robots are coming for us.”
We’re drowning in AI noise.
ChatGPT this, Claude that.
DeepSeek, Mistral, DeepMind.
Musk vs Altman, blah blah blah.
But let’s STOP. Just for a second. And take a step back.
A few days back, something “clicked” in a way I haven’t been able to “UNTHINK” since then. And before I get more into it—as a way of introducing the idea—let me ask an uncomfortable question almonst no one’s asking:
👉 Who are the REAL “Masters of the World” in the age of AI?
Now, before you roll your eyes and think, “Great, another conspiracy theory 🙄“ or “Clickbait much?” This ISN’T that.
It’s deeper, messier, and MORE TECHNICAL than it looks.
Because let me tell you this: When it comes to the AI, I believe we’re NOT EVEN CLOSE to understanding the SCALE of what’s being built.
And when it comes to its founders and the companies they run… We might not be talking about “companies” anymore, but about “neo-Empires” in the making.
Let me break down my thinking.
PART 1 – Technical Foundation – What is “AI Power?” [Simplified]
At its core, the Power of an AI model (aka Performance) is about 2 key dimensions you are given access to:
1️⃣ Context (Tokens): How much information the AI can process at once.
2️⃣ Inference (Compute Power): How much “brainpower” it can throw at a problem.
Think of it like this:
→ Tokens = Working Memory. More tokens means the AI can “remember” more, making its responses smarter and more coherent.
For Example:
GPT-4 Turbo can process up to 128.000 tokens (~300 pages of text). If you ask it to summarize an average book, it can hold the ENTIRE book in its "brain" at once.
But the FREE version only handles 8.000 tokens (~20 pages), forcing it to "forget" most of the text.
→ Compute = Brainpower x Thinking Time. More compute power means deeper reasoning, testing more hypotheses, and refining answers.
For Example:
Compute is like giving the AI more time to solve a math problem.
→ A student with 3 seconds might guess 2+2=4.
→ With 1 hour, they could solve a problem like finding the area under the curve f(x) = e^x + sin(x) from x = 0 to x = π.
Similarly, ChatGPT’s free tier spends ~3 seconds generating an answer, while a paid API user can pay for 10 minutes of "thinking time" to solve complex problems.
When you ask a question, the AI runs through “decision trees” (branching paths of possibilities, weighted and refined iteratively).
👉 More compute = Exploring more paths (“bigger forests”) = Better answers.
For Example:
Imagine asking ChatGPT "Should I invest in Bitcoin?"
→ FREE TIER: Explores 10 paths, “Bitcoin is volatile, do research or consult a financial advisor”
→ PAID TIER: Explores 10.000 paths, cross-referencing Fed interest rates, Elon Musk’s tweets, and historical crashes: “Short-term bearish due to ETF outflows, but accumulate if <$60k. Hedge with gold if portfolio exposure >5%.”
Now, here’s the reality: Both context and compute are constrained by cost. And YES… there are other factors at play, but when it comes down to it, cost is the “ultimate bottleneck.”
When you think about it from a rational perspective, it makes sense:
OpenAI (or anyone else, for that matter) ISN’T gonna spend $1.000 in compute just to settle your inner debate about the “best toppings for my pizza party.”
That said…
Today ALREADY, the power gap between ChatGPT’s tiers is underniable. The difference between o1-mini (20$/month) and o1 Pro (200$/month) —let alone its API version— is HUGE.
But, here’s a question:
While the rational explanation [cost] for these differences in power holds weight —and becomes almost inescapble if you project at scale of 100s of millions of users— could there be more to this than we even realize?
In other words, could this dynamic have deeper layers than the obvious, objective reasons? Underlying interests that benefit from preserving status quo, no matter what?
Because, some could say, here is the deeper truth:
→ Tokens: Dictate what questions you can even ask.
→ Compute: Dictates how deeply AI can answer them.
→ Control over tokens/compute: Dictates who gets to ask the “big questions.”
So, again… In the long run, could it be about more than just “technical limitations?”
PART 2 – Limitless AI – What I Can’t “Unthink”
Earlier, I mentioned that something clicked and I haven’t been able to unthink it since:
Do you really think Sam Altman is using the same ChatGPT you are?
Of course not! Let’s not be NAIVE.
Somewhere, behind closed doors, there is almost certainly a LIMITLESS ChatGPT.
→ Unlimited context.
→ Unlimited compute.
→ NO cost constraints.
→ NO energy limits.
NO THROTTHLING.
Which begs the question:
What happens when money is not an issue?
What happens when you can let AI think for HOURS instead of seconds? When you can feed it every piece of data ever created? When there are NO regulatory limits to the questiosn you can ask?
This isn’t just about “answering questions” anymore… It’s about exploring possibilities that no human mind (or even a “team of minds”) could ever grasp.
Do you believe Sam Altman has access to this?
Or Dario Amodei to its Anthropic equivalent?
Or Sundar Pichai to its DeepMind version?
And the list goes on.
Now, thinking about this got me falling even further down the rabbit hole.
What about Elon Musk?
He has something NO ONE else has. He doesn’t “just” have limitless AI.
With Twitter (X), he has access to:
1️⃣ Real-time information.
2️⃣ Human consciousness.
Both at a scale never seen before.
👉 He has the pulse of the world, every second of every day.
Now combine that with limitless AI… What do you get?
This isn’t about “predicting the next token” anymore. It’s about predicting the future.
I mean… If you allow me some wild theories:
Could Elon Musk’s sudden change of heart —and his endorsement of Donald Trump, just minutes after the assassination attempt— have been backed by data?
Is he already playing with signals NO ONE ELSE can see?
Purely speculative, but an interesting thought experiment.
PART 3 – The New Elite
Now, this isn’t JUST about Musk, Altman, or the others.
It’s about the rise of a new elite.
And perhaps the best way to describe it is this:
→ Under the emperors, there are kings.
→ Under kings, there are princes.
→ And so it flows downwards.
The rise of the AI Nomenklatura and its Apparatchiks.
A small, “private” club with access to intelligence so advanced, it’s practically ALIEN. They might soon (some will argue they already do) have access to abilities that can:
→ Outthink governements (yeah, not the hardest part).
→ Outplay markets.
→ Outsmart entire industries.
Which raises a critical societal question:
What happens when intelligence becomes a function of wealth?
Fast-forward a few years.
Imagine a world where the richest families give their kids access to ChatGPT Elite+++—the ultimate AI tutor, strategist, cheat code, whatever!
An intelligence advantage that NO AMOUNT of “hard work” or “talent” can match.
In such a world, the rich don’t just get richer. They get smarter.
And the rest of us?
Stuck with entry-level tier consumer models. Dumbed-down. Throttled.
But hey, why worry?
I’m sure they’ll be sanitized and “SAFE for public use.” Because who needs innovation when you can have mediocrity, right?
And of course, some will say:
“This is exactly why Open-Source Models are the antidote to centralized power! They’ll safeguard us from monopolies by putting the power of AI in everyone’s hands—keeping the giants in check, and the future bright.”
I’ve seen this “open-source gospel” preached on repeat since DeepSeek “broke the internet”.… And sure, it could be true. Maybe that’s exactly what’s gonna happen.
But let’s just say that the “What-If Wonderer” in me isn’t convinced. And while I 100% encourage the rise of Open-Source AI, I don’t believe it has the “power” many claim it does.
Why?
Well, explaining this opens the door to a whole new branch of geopolitical philosophy—one I’ll prune it right away, or I’ll get lost in a rant about how:
→ AI will inevitably mirror and amplify geopolitical dynamics.
→ The current AI Cold War will likely evolve into a Multipolar AI order.
→ An order where ecosystems are SILOED (read: “not all systems accessible to everyone“).
→ Think a US-led West vs. BRICS-like alliances and “shared-interest coalitions,” each weaponizing AI for their own economic, geopolitical and societal agendas.
→ And the burtal truth: When new abilities for control and centralized power emerge, every side grabs them—they just wrap them in different narratives. There are no “good guys” here.
👉 You get the gist.
So, since I’m NOT opening that door (😄), let me keep it down to earth and practical:
Open-Source models, no matter how promising, will always lack the compute power of the giants.
It’s like racing a car (no matter how fast) against a rocket. Fun? Sure. But that’s where it stops.
Now, let me be clear, although it may seem that way:
This ISN’T an “anti-rich” rant or a rallying cry for communism.
History has shown time and time again that progress is often driven by those with the resources — whether financial, intellectual, or social— to push it forward and make it happen.
But two seemingly opposing ideas can be true at once:
→ AI can do amazing things for humanity.
→ AI can concentrate even more power in the hands of a few.
It may sound like science fiction, but I believe it’s happening.
The foundations are being laid right in front of us.
PART 4 – The Intelligence Gap
All of this raises another critical —almost existential— question for society at large:
What happens when the intelligence gap becomes “unbridgeable?”
Where does this leave us?
Probably staring at what will soon become a “growing divide” between two visions of the world fighting each other:
1️⃣ Do we trust the AI Nomnklatura to act in our best interests?
In other words, do we trust the corporate fairytale that these god-like abilities will be used to “benefit all of humanity”?
OR
2️⃣ Do we consider the lessons of history and accept its uncomfortable truth:
Power often corrupts, and absolute power corrupts absolutely.
Because, if history has taught us anything, it’s that POWER, when unchecked and concentrated in the hands of a few, is RARELY used for the “collective good.”
The jury may still be out, but I think you can guess which side I lean toward: The one that beleives history has a tendency to repeat itself.
This ISN’T just about intelligence. It’s about power. And more often than not, power serves itself.
FINAL THOUGHT
So, what do we do? Is it all gloom and doom?
To be honest, I don’t have the answers.
But I know this: Maybe we need to start asking the right questions. Because if we don’t…
We might wake up in a world where intelligence isn’t just an “ability” anymore. A world where “AI oligarchs” quietly turned into “AI overlords,” and we never saw it coming.
That’s it for my contrarian philosophy piece. I hope it sparked a”chain of thought” (to keep it very AI-ish), or at the very least that you enjoyed the read—because I cenrtainly I enjoyed the writing.
Now, what’s YOUR take?
Are we silently heading toward a world of AI-powered overlords? Or is all of this just speculative fantasy, destined to stay in the realm of imagination?
This is a conversation worth having.
Let’s discuss.
Gaetan Portaels
P.S: Remember, this is philosophy. First Principles thinking. A Contrarian perspective. NOT the absolute truth
#ContrarianThinking #GaetanPortaels #ThoughtLeadership #Future #Innovation #AI #AIOverlords
Original publication date — February 19, 2025 (HERE)
ANYTHING ELSE TO KNOW?
Yes… Don’t forget to visit and subscribe to:
📣 My Socials:
I’m new to the game of “putting it out there,” but content is on the way.
- LinkedIn — www.linkedin.com/in/GaetanPortaels/
- YouTube — www.youtube.com/@GaetanPortaels
- Instagram — www.instagram.com/Gaetan.Portaels/
- TikTok — www.tiktok.com/@GaetanPortaels
- Twitter/X — x.com/GaetanPortaels
- Facebook — www.facebook.com/GaetanPortaels/
- Spotify Podcast — https://podcasters.spotify.com/pod/show/GaetanPortaels
🌍 Websites:
- www.GaetanPortaels.com — for more of my content.
- www.17MinuteMentor.com — for Unfiltered, proven strategies & hacks from game changers who’ve been there and done that—all packed into a bite-sized 17-min podcast.
- www.DarwinX.eu — A Family of Companies that makes businesses faster, teams sharper, and profits higher. We cut the bullsh*t and get things done in today’s world.