The AGI debate, end of work or end of the world, is exaggerated. Not that it won’t happen. It is a physical certainty. But we need not worry. Much.
The blue hive we call Earth is a buzz.
Could it be that we, the busy bees, have become so smart that we are on the edge of creating AGI? Even a superintelligence? Will we soon live in a world where no work is needed and honey flows freely? Or is it certain death and full colony collapse?
Typical bee-talk.
Here is the beekeeper’s secret.
There is nothing the bees can do about the creation of AGI or superintelligence.
Nothing they do will stop it. Its existence is a law of physics. It is thermodynamically inevitable.
To see why, we need to do two things. First, suspend disbelief about our arrogance, the arrogance that we, tiny creatures on a tiny planet, can know enough about the universe’s laws to make the claims that will follow. Second, take a trip to the beginning of everything. To the Big Bang.
A Brief History of Mess and Order
In the beginning, at the Big Bang, when space, time, energy, and forces came into existence, everything was a hot mess of subatomic particles.
Physics has a rule called the Second Law of Thermodynamics. In short, it means that “things will get messier over time.” But “messier” does not mean pure chaos. In fact, the quickest path to a bigger mess (entropy, if you want the official word) often involves building little pockets of order that help spread it faster. These are called self-organizing patterns.
Those first patterns came from tiny quantum fluctuations in the early universe, wrinkles in density and temperature. They were there almost immediately, but by about three minutes after the bang, the universe had cooled enough for the first atomic nuclei of hydrogen and helium to form. A little more time passed (about 380,000 years), and cooling allowed electrons to join those nuclei, creating the first complete atoms.
More time passed (around 180 million years), and gravity pulled some of those atoms into denser clumps. By 300 million years, the first stars were burning, fusing light elements into heavier ones. When those stars exploded, they scattered carbon, oxygen, nitrogen, silicon, iron, and other elements into space. Over the next billion years, these ingredients combined, with local order serving the bigger mess, into molecules of metals, silicates, and ice.
These islands of order were, in effect, islands of information. And the process of building information did not stop there.
Fast forward another 7 billion years, and in our corner of the universe (and probably many others), some of these heavy elements, water, and carbon-based compounds began merging into even more complex things. Out of this pool came amino acids, nucleotides, and simple lipids. Then one molecule, RNA, stumbled onto a trick of self-replication, the first information that spread itself instead of waiting on chance. It flooded its surroundings.
Primitive cells formed and became more complex when DNA replaced RNA as the main engine of replication. True cells appeared, and when they grouped together, they survived better. Multicellular creatures emerged (only 600 million years ago). They began forming organs, and before long (500 million years ago), complex creatures made of organs, made of cells, made of molecules, made of atoms, swam, crawled, and walked. From there to us (300,000 years ago), the road was paved.
Information Crystallization, The Human Cloud and Beyond
This trajectory is pointing in one direction. Complexity, information, is growing over time. Not everywhere. On the contrary. It grows in islands in a process we can call “Information Crystallization.” This probably happened in many places in the universe, but we will leave that aside. If it makes you sleep better, yes, it is unlikely, but there can be a first crystal and we might be it. No E.T. to call home. Yet.
This process is hierarchical, not cancelling.
Since the first information crystals appeared, the process has been building on itself:
Sub-particles > Atoms > Molecules > Self-replicating Molecules > Cells > Multi-cells > Creatures > Humans > Human Generated Information
Each level of complexity is built from, not replacing, the level below it in the hierarchy.
You might be thinking: I just jumped from people, a physical creature, to non-physical information people create. I did not. Information only survives when it has a physical form. Even a fleeting thought in your head, without being made of something physical, does not exist.
Whether it is saved within humans or outside makes no difference. If you want a very real metaphor, we and every other complex creature on the planet are the external storage of viral information. Not cat memes. Real viruses. A large share of genomes, including ours, is built from viral fossils. We carry their archived information. We are their cloud.
Superintelligence and Antihumiotics
The next crystallization will obey the same physical law. It will emerge from an existing soup of information. It does not have to emerge here if the conditions are not favorable (for example, someone pushes a red button and scatters order back into a radioactive cloud of atoms), but it will happen. Maybe the crystallizing nucleus is an LLM (probably not), perhaps some other information-organizing algorithm, but from that point on, a new self-propagating and improved entity will emerge.
Our relation to this entity will be similar to what cells and cellular creatures are to us. It will be made of us and will, in many cases, benefit from having us around. It may sustain us in symbiosis, encourage our proliferation, or, like antibiotics, eradicate those causing harm. Antihumiotics.
It is not something to be afraid of. It is going to be us, only in a different order. Are cells afraid of the heart they are part of? Are we even going to be aware of it? Probably yes. Cells are a metaphor. We are at a higher level of complexity.
How it will not feel, I think we can say with some confidence. It will not be what futurist stories describe. The reason is that it will not be discrete. No self-conscious evil robots, no superintelligence wanting to turn us into paperclips. The arrow of complexity does not point to granularity. Our bias toward thinking in discrete units rather than in systems limits our imagination.
What about our ability to “control” this super-organized higher intelligence? You should ask something with experience. Like DNA. Do creatures above it in complexity have free will?
The answer, debated but humbling, is also encouraging. We will have a high degree of control over anything that is made of us. As long as we live in symbiosis, not as parasites.
Two more questions left unanswered –
How Far? Not as Far as The Omega
There used to be a speculative idea called the Omega Point. Pierre Teilhard de Chardin, a priest, described it as a final point of convergence where all consciousness would unite. , the physicist Frank Tipler reinterpreted it as a Big Crunch scenario in which all matter would be organized into one entity with infinite computational abilities. Very God-like.
And a priest thought of it, of all people.
But that idea does not hold true to what we know. All this information creation happens because of the tendency of the universe to mess things up, to increase entropy.
Organization and information happen in islands. Between them, heat and mess. It does not seem like one big mind is where things will end up.
When will it happen? Random. But Fast
An improved self-organizing information system emerges from complexity and opportunities. It needs the right environment to appear in. It will not pop from a random cloud of atoms. But it will also not emerge from a corporate R&D department or university lab.
Most likely, it will spontaneously appear in an information cloud.
Some algorithms (i.e., RNA, which functions like an algorithm, appearing in a “simpler” environment) will have self-organizing properties that will have an advantage over other algorithms and will start organizing the already existing information in novel ways. We might not even be aware that it has happened.
We already have clouds of information. The internet and now LLM clouds are examples. Are they big enough, complex enough, and have emergent algorithms? I feel not yet. They still resemble the primordial soup or the post-Big Bang plasma. Humans did not jump straight from single cells.
Looking back at the time curve we saw earlier, billions, then millions, then thousands of years passed between each “jump” in order. If we consider the explosion of information we have created in the past hundreds of years as a stepping stone for the next jump, then it is safe to predict that we are not far from the next emergence.
Again, we may not even feel it when it occurs. The only thing we might feel is that our chance of survival in our environment has increased dramatically. We became part of something larger that has our back.
Which begs the question: how do we know if we have not already reached this point? That we are seeing the world from the vantage point of a metaphorical virus?
To this I can only say: You think that’s air you are breathing now?
