Any sufficiently advanced technology is indistinguishable from magic.
Sir Arthur C. Clarke.
When I was about 6 or 7 years old, I wrote my first computer program — a simple fortune-telling game. The program requested inputs such as name and the current day of the week, then printed text based on those inputs, like “Hi Darek, nice to meet you! Tomorrow will be a very sunny day, so get ready for a hot weekend.” It was a series of if-then statements written in BASIC and executed on my Atari 800XL. Though I understood how the program worked, I thought real programs couldn’t be that simple, as the games I played seemed more complicated. Now, with a computer science degree under my belt, I realize there’s no magic dust that distinguishes true games from that small project. To a certain degree, it’s if-then statements all the way down, complemented by data structures, but there’s nothing magical about them either.
Two decades later, in my early twenties, the Internet was gaining popularity (pre-Google), and it felt magical. Curious about how webpages were created, I bought a book and taught myself HTML and JavaScript. As I learned, the magic dissipated. I understood how browsers worked, what they did to display content, and how to control their actions. Once again, it seemed like simple steps, applied diligently, produced predictable results.
Computers are like large, fast calculators — they don’t do anything particularly creative or crazy. They’re consistent and obedient, with every “if” accompanied by a “then.” There’s no magic, just the execution of instructions.
Modern Large Language Models certainly feel magical. These models, built on vast amounts of data, are capable of generating human-like text, answering questions, and even creating art. They seem to possess a level of creativity and complexity that goes beyond the if-then statements and data structures we’re familiar with. They feel incomprehensible and almost miraculous.
But will this feeling of wonder dissipate over time? Are we simply too early in the game to appreciate that the same principles underlie these new technologies, even if their results are more surprising and confusing?
Looking at LLMs, I can’t help but feel that we’re witnessing the birth of an amazing and magical technology. Yet, I often wonder if this is similar to my childhood experience with “real games” or the Internet before I learned HTML. Perhaps it’s just a matter of getting accustomed to it. How will my children relate to this tech in 5 or 10 years? Will they share this feeling of wonder and amazement? Or maybe they will just be so familiar with it that the magic just won’t be there. After all, we don’t think radio is magical. Or electricity. Or fire. And yet, all of them at some point in history must have felt like something wonderful and superhuman. Ultimately, we are all humans on the path of exploration and growth. Embracing the unknown is just a part of our journey. We should be used to it by now. Or should we?