Chances are when I say alchemy, you begin thinking of old people hunched over in a distillery, attempting to create The Philosopher Stone. A stone which could be used as a transmutation agent, or could be mixed into an elixir to create an elixir of immortality. What you may not know, is that for many:
Alchemy [was] the understanding of the relationship between consciousness and matter1
Alchemy was a lot more than a simple hunt for The Philosopher Stone. Many alchemists were searching to build an understanding of the way matter functions.
In fact, alchemy originally began as a practice to make certain metals appear like another. In other words, alchemy began as mostly con artists hoping to trick people. It originally began as the study of imitation, not the creation.2 For example, instead of turning lead to gold, they would tint silver to make it appear golden.
A very common test for the consciousness of AGI is known as The Turing Test. It boils down to creating a machine / artificial intelligence which is capable of convincing a human that the machine itself is a conscious human. We are following the same paths as the alchemists, we are not yet attempting to create, we are trying to imitate.
Now the Turing Test has received much criticism - often opting for a more robust version of the test, yet it still boils down to imitation. We are still in this phase, whether we have completed the Turing Test or not is still not clear. We yet to create an AI which can convince any human it is real, but we have made some which may be capable of convincing inexperienced humans that it is in fact also a human.
Eventually, some time during the 3rd century BCE, someone (we’re not quite sure who) got the brilliant idea that perhaps if we are capable of tinting metals (and imitating other metals), then surely, we must be capable of changing more of their properties, perhaps even replicating them! 3
We are following the same steps in the AGI space, perhaps instead of simply imitating human intelligence, we could copy it? This has lead to approaches such as Whole Brain Emulation, which is the process of scanning, reconstructing, and simulating an entire human brain.4 Perhaps from the physical matter of the brain, if we are to replicate it, consciousness may emerge!
There are many other approaches, but I am not very well versed in their understanding, so I will hold off a little before explaining them (perhaps in future letters)
In the Alchemical world, once this idea of replicating metals came about, they started theorizing as to how it could be done. Hence, The Philosopher Stone. A stone which could be used to transmute lead into gold, or any metal into another. Alchemists would inevitably claim to have created the stone, whether they were lying or if they believed it themselves we are not sure. But the governers of the time leaned more towards optimistic skepticism.
Having the capacity to turn less precious metals like lead into gold could make kingdoms prosperous! And so many Alchemists were put to the test. Kings wanted to hire alchemists and practically turn them into contractors in order to create gold and get paid. But they were also skeptics (and for good reason). In order to get hired as a contractor, Alchemists needed to display true transmutation in front of a public audience.5
This proved to be incredibly difficult, as without nuclear reactions, it is impossible to reach transmutation.
Still, Alchemists stepped up, attempted transmutation, and a large portion of them were caught as frauds (and hung…) But, a few slipped through, again, whether they had planted gold, or if they accidentally brought gold embedded lead is still up for debate.
I feel as though a similar thing is beginning to happen with AI systems. Those who claim to have reached AGI capabilities are put to the test via public exhibition (publication). Which is a great thing, I am very glad that this is how we test things. Imitation of intelligence no longer passes in these stages, as now we have tests we are capable of running in order to determine whether a program has pseudo-intelligence, or real intelligence. A program might appear intellectual on the surface, but after more digging, it is clear it is not. The same was the case for the Alchemists who tinted silver into gold. Perhaps it passed originally, but as they developed tests for the purity of metals, then they could know when they were being lied to.
Alchemy began with practically nothing concrete, beginning with the 4 elements outlined by Aristotle:
Air
Water
Fire
Earth
But even with such an abstract understanding of what was happening, alchemists still discovered many elements which are now on our periodic table.
Alchemists never reached their magnum opus, The Philosopher Stone was never created. It was never a real science, but it was the building blocks to modern science and medicine.
I hope AGI Engineers accomplish their magnum opus, but if we don’t, I’m sure it will lead to revolutionary advances in future fields.
This is a fascinating historic insight. There is a parallel in the quest. The difference, however, lies in the purpose of the quest. The alchemical quest, though at times, was couched in the noble goal of the key to life was really about gold, or personal enrichment, the AI philosopher's stone is a more noble thing, it a quest for the quests sake. As such the delusion of the Turing test is not deliberate fakery by charletons but well meaning delusion about what is really being saught.
Is it intelligence you want? Or is it sentience? Intelligence you have, by many verifiable definitions. So perhaps what you seek is sentient intelligence, self aware, self reflective.
And in this understanding, perhaps, lies the true path to the philosophers stone. Intelligence, what creates it and how it functions we seem to be getting, but sentience is the illusion. And that may lie in organics of the sences and the interpretation thereof.
But this undoubtedly has already been thought of.
Interesting article! I wouldn't have thought to relate the developmental process of AI to that of alchemy ... that's cool!