josiah jennings

I’ve always approached the subject of artificial intelligence with the skepticism of the “Chinese Room Argument” (https://plato.stanford.edu/entries/chinese-room/). How could we ever hope to replicate the complexity of the human brain such that this replication exhibits consciousness? Defining “intelligence” notwithstanding, artificial intelligence must not only appear intelligent but actually be intelligent. Of course, we must sufficiently define intelligence before we can test for it. Moreover, we must sufficiently understand consciousness, as it seems to be inextricably tied to intelligence. For this, the Turing Test alone is insufficient. The most advanced large language models are virtually indistinguishable from having a conversation with the average human, yet few if any with an understanding of computing or neurology would define these models as “intelligent” or “conscious.” Our answers to these questions are often predicated on a set of narrowly defined parameters. That said, advancements in modern computing and AI discourse show exciting promise.

I say all of this only to say these days my skepticism of replicating consciousness falls as much in favor of the possibility as it does against it. One of my biggest mental hurdles to accepting artificial intelligence as indeed possible has always until recently been the sheer scale from which consciousness emerges as a complex system. How in the world do we possibly engineer something like that, something nearly 4-billion years in the making that operates at a near infinitesimal scale? Lately, I’ve been studying a lot about the intricacies of CPU and GPU architecture, given my pursuits in cybersecurity and the rapid ascent of NVIDIA, and my skepticism, at least in this regard, has vanished. I admit I’m late to the party, but it was not until lately when my professional and educational endeavors aligned in such a way that I’ve been able to fully appreciate what’s happening right now. Previous skepticism aside, I’ve always held the opinion that artificial intelligence should at least be theoretically possible. After all, our own consciousness emerged from the same universe as the one in which our technology is engineered. Shouldn’t it then be possible to reverse engineer such a phenomenon?

Transistors are the building blocks of the microprocessor, and transistors within modern microprocessors are astonishingly small. State-of-the-art transistors are typically 5 to 7 nanometers in size. 1 nanometer is 1-billionth of a meter. Comparatively, neurons, the fundamental building blocks of the brain and nervous system, range anywhere from 4 to 100 micrometers in diameter. 1 micrometer is 1-millionth of a meter. This means that even the smallest neurons are thousands of times larger than the smallest transistors, implying that if we scaled a transistor up to the size of a small coin (around 1 centimeter in diameter), a neuron would accordingly expand to the size of at least a large beach ball, ranging anywhere from 40 centimeters all the way up to 10 meters in diameter. Brains remain incomprehensibly complex to us, but if scale is any indication, modern microprocessors have the potential to be so much more complex, especially considering quantum computing. I realize there are serious problems with this argument when taken in and of itself, but this is where computer engineering and architecture come into play.

Essentially, computing and consciousness reduce to the manipulation of information. Despite their size differences, transistors and neurons serve similar functions in their respective domains. Transistors control the flow of current in electronic circuits, forming the backbone of digital logic and computing. Similarly, neurons transmit information through electrical and chemical signals, enabling brain function and nervous system operations. This stark size contrast between modern transistors and neurons underscores the incredible miniaturization achieved in electronic components compared to the larger, intricate structures of biological cells, and thus, also, the possibilities. As we continue making advancements in information technology, the boundary between engineered intelligence and natural consciousness blurs. The very universe that birthed our own consciousness might indeed provide us the tools to recreate it.