The Computer as a Communication Device
I’m reading the 1968 paper “The Computer as a Communication Device” by J.C.R. Licklider and Robert W. Taylor and I can tell it is going to change my life, because it is going to change how I see the future of tech, which is my future.
Yesterday I wrote a blog post about an organization’s cultural conditions that, in my estimation, are required to move towards a problem’s true formulation.
The quote that served as the article’s epitaph was from “The Evolution of Physics” by Einstein and Leopold Infeld:
“The formulation of a problem is often more essential than its solution, which may be merely a matter of mathematical or experimental skill. To raise new questions, new possibilities, to regard old problems from a new angle requires creative imagination and marks a real advance in science.”
“The Computer as a Communication Device” is a problem formulation paper. You can sense in their words and phrases that they are envisioning a future they can’t fully implement (and don’t need to) but have certainly formulated. As a kid who grew up in the analog-digital boundary it’s nothing short of a spiritual experience to read this paper for the first time as an adult.
They start by describing what it means to communicate:
“But to communicate is more than to send and to receive. Do two tape recorders communicate when they play to each other and record from each other? Not really-not in our sense. We believe that communicators have to do something nontrivial with the information they send and receive.”
How can you not think of Human-LLM interaction after reading this? By their definition, “communicators have to do something nontrivial with the information they send and receive.” We can argue to death about whether LLMs are intelligent, or conscious, or can/cannot reason. But I don’t think we can argue that LLMs (though they are tools in my opinion) are not just tape recorders— they do something nontrivial with the information we send to them. LLMs are communicators, by that definition. The question “what do I want to build with LLMs” or “how do I want to use with LLMs” becomes: what do I want to communicate to LLMs? What do I want them to communicate back? and what do I want to do with that information (trigger a function call, execute a script, render the image, etc.)?
I’ll continue to post more thoughts from this paper as I come across them.