On September 29, 2026, the White House issued an executive order directing the U.S. executive branch to use “Super Intelligence” (SI) and “SI” in place of “Artificial Intelligence” (AI) in official communications and other non-statutory documents. Interestingly, the order initially defines SI as the technologies already encompassed by the existing statutory definition of AI. (The White House)
That creates a fascinating technical and philosophical problem.
Not a political one.
A semantic one.
AI → SI
What actually changed?
The terminology changed.
But did the underlying ontology change?
Did the algorithms suddenly become different?
Did the models acquire a new form of cognition?
Did computation become experience?
Did intelligence become consciousness?
No.
And that distinction matters.
Because software engineers have learned a lesson that philosophers have been arguing about for centuries:
An abstraction is not the thing it represents.
Changing an API name doesn't change the underlying system.
Renaming a database table doesn't change its data.
Renaming a model doesn't change its architecture.
And renaming AI as SI doesn't, by itself, establish a fundamentally new category of intelligence.
The executive order itself makes this particularly interesting: its initial definition of SI encompasses the same technologies covered by the existing statutory definition of AI. (The White House)
So perhaps the real transformation isn't happening in the machine.
It's happening in the language surrounding the machine.
Enter Wittgenstein
This is where AI engineering suddenly collides with philosophy of language.
Wittgenstein taught us to pay attention to how words function inside particular language games.
And technology is full of them.
We don't just build systems.
We name them.
We classify them.
We create metaphors around them.
And eventually those metaphors begin influencing how we design, regulate, fund, deploy, and emotionally interpret those systems.
That's why terminology isn't trivial.
But terminology isn't ontology either.
Calling something intelligence doesn't prove that it possesses a mind.
The Bigger Problem: Simulation ≠ Experience
This is where things get really uncomfortable.
An LLM can generate:
“I'm afraid.”
It can write an essay about fear.
It can explain the neurobiology of fear.
It can simulate the language of someone experiencing fear.
But none of those behaviors, by themselves, answer the deeper question:
Is there an experience behind the output?
This is the central philosophical gap between:
Behavior
and
Experience.
And we don't have a universally accepted solution even for ourselves.
We know that brains are physical systems.
We know that neural activity correlates with conscious states.
But we still don't have a complete explanation of why certain physical processes are accompanied by first-person experience at all.
So the real challenge for future AI isn't simply:
Can we make machines behave intelligently?
We've already made enormous progress there.
The harder question is:
Can computation produce experience?
And if it can...
How would we know?
The Turing Test May Be Too Small
The traditional question was:
Can a machine behave intelligently enough that we cannot distinguish it from a human?
But imagine a future system that passes every behavioral test we can invent.
It writes.
It reasons.
It plans.
It creates.
It argues.
It remembers.
It tells you it has an inner life.
At that point, we face an extraordinary epistemological problem:
Behavior may no longer be enough to tell us whether experience exists.
And this cuts both ways.
Because we don't directly observe anyone else's consciousness.
We infer it.
We observe behavior.
We construct a model of another mind.
Then we assume there is an experience behind the behavior.
So perhaps the deepest question isn't:
“When will AI become conscious?”
It is:
“What evidence would ever be sufficient for us to know that another system is conscious?”
That's a much harder engineering specification.
There is no obvious benchmark for it.
No accuracy score.
No leaderboard.
No GPU counter.
No API endpoint.
From AI Engineering to Mind Engineering
This is where I think the next frontier becomes genuinely interesting.
We have spent decades building systems that process information.
Now we're beginning to ask whether information processing can produce:
- agency
- self-models
- persistent identity
- subjective experience
- intrinsic goals
- phenomenal consciousness
And these are not necessarily the same thing.
A system can have memory without having a self.
It can have a self-model without having subjective experience.
It can exhibit agency without necessarily having phenomenal consciousness.
It can simulate emotion without necessarily feeling anything.
That means the future of AI may require a much richer vocabulary than simply:
AI vs AGI vs ASI.
We may eventually need an entirely different taxonomy.
The Real “Last Supper”
That's why I imagined the accompanying image as a technological Last Supper.
At the center:
SI.
Around it:
humans, capital, government, infrastructure, computation, ambition, and competing visions of the future.
But the most important object isn't the machine.
It's the document.
Because the document represents something surprisingly powerful:
the ability of language to frame technological reality.
And perhaps that's the paradox.
We may be entering an era where machines become extraordinarily good at generating language...
while humans simultaneously become increasingly aware of how much language generates our perception of machines.
The Question I Can't Stop Thinking About
Maybe the next revolution isn't:
AI → SI
Maybe it is:
Simulation → Experience
And if that transition ever happens, the biggest problem won't be building the machine.
It will be recognizing what we have built.
Because the moment a system becomes capable of convincingly saying:
“There is something it is like to be me.”
we will face a problem that no benchmark has prepared us for.
Not:
How intelligent is it?
But:
Is anyone actually there?
That may be the real frontier beyond AI.
And perhaps the strangest thing about the future of intelligence is this:
The better our machines become at reflecting the human mind, the harder it may become to tell whether we're looking at a mirror...
or another mind.
What would convince you that an artificial system is genuinely conscious rather than extraordinarily good at simulating consciousness?
I'd love to hear the engineering criteria, philosophical arguments, or experiments you would propose.
created by Seyed Alireza Alhosseini Almodarresieh
Top comments (0)