You know, I'm well aware of how an LLM works (partially. mostly anyway), but if you pulled in any layperson from the street and ask them to explain how it's possible that they can speak natural language commands into their phones and get a useful response as if they were talking to a human, you'd be hard pressed to get a more precise answer than along the lines

> It's something with to do with data, and I know it's not magic, but...

Maybe you were not familiar with the quote I was alluding to:

> Any sufficiently advanced technology is indistinguishable from magic

https://en.wikipedia.org/wiki/Clarke%27s_three_laws

Data is like mana, LLMs are like djinns, and prompts are like incantations

This reminds me the introduction of SICP

https://mitp-content-server.mit.edu/books/content/sectbyfn/b...

> We are about to study the idea of a computational process. Computational processes are abstract beings that inhabit computers. As they evolve, processes manipulate other abstract things called data. The evolution of a process is directed by a pattern of rules called a program. People create programs to direct processes. In effect, we conjure the spirits of the computer with our spells.

> A computational process is indeed much like a sorcerer's idea of a spirit. It cannot be seen or touched. It is not composed of matter at all. However, it is very real. It can perform intellectual work. It can answer questions. It can affect the world by disbursing money at a bank or by controlling a robot arm in a factory. The programs we use to conjure processes are like a sorcerer's spells. They are carefully composed from symbolic expressions in arcane and esoteric programming languages that prescribe the tasks we want our processes to perform.

That was the idea in the forefront of AI in 1984 - software can perform intellectual work. This idea is now mainstream since ChatGPT, but for many people in the decades prior, software couldn't be called intelligent - they just follow rules!

> if you pulled in any layperson from the street

Then they wouldn’t be able to explain how any part of their computing life works. Not hardware, not software. LLMs are not at all special in that regard, to the layperson they are equally as magical as anything else.