People have been giving instructions to computers for a long time now. 

I’m doing it right now. Every time I make my finger press down on one of these keys, it’s like a little instruction: “Computer, please insert the letter ‘r’ into the place where the cursor is.” 

When I take a picture with my smartphone, I’m giving the computer an instruction: “Computer, use your on-board camera to take a photo right now, using the same instructions (zoom, flash, etc.) that I’ve already given to the Camera app.” 

“Zoom in a little bit.” 

“Submit this form.”

“Scroll down.” 

“Send message.” 

We don’t really think about it when we feed these instructions to the computer. We just do it.

People don’t want a drill. They want a quarter-inch hole. 

The product design community loves this quote, but in reality it’s not quite right. We’re a little too preoccupied with the drilling process. People don’t particularly care about making quarter-inch holes. In fact, they’d skip it if they could. What they really want is the happiness they get when they look at a sentimental picture hanging on their wall.

It’s very easy for us to lose sight of why we’re using computers. It’s very easy for us to lose sight of the computer’s place in the world.

We don’t like to think about the computer’s place in the world because it is a boring one. Are you zoomed in 52.1% or 53.4%? Who cares? But that’s an actual job the computer is doing. You just don’t think about it. And it’s a good thing that you don’t have to think about it. That tiny job is one of the many thousands of little jobs the computer is (and has been) doing for you whenever you take a photo! That is computing at its best - saving or modifying millions of zeros and ones that you don’t have to think about in service to your meaningful, human end goal.

You should never have to think about lighting up pixel #13,423. You should have think about if you want to take a picture of your dog, and why or why not. 

People don’t want a quarter-inch hole. They want the happiness they get from a picture on their wall. 

We people feed instructions to computers because of things we have going on in our lives, things we are trying to accomplish, and we are using the computer to accomplish those things. 

“Computer people” often make the mistake of thinking that the computer is an end in itself, because to a computer person, it is! Computer people enjoy computers because they enjoy computers. It’s their hobby. It’s the same thing as how golf people enjoy golf because they like playing golf.

The difference between computers and golf, however, is that you’re not forced to play golf in order to function in society. Non-computer people are forced to use computers to function in daily life.

Computer people are prone to enjoying the process of drilling a quarter-inch hole. Computer people like a good challenge. Writing software code by hand is an example of an instruction that is difficult to give to the computer. For computer people, the difficulty gives them the challenge they’re seeking. It lets them sink more effort into their hobby. 

On the other hand, non-computer people wish they could avoid the process of drilling altogether. They want it to be as easy as possible to give the computer instructions. The instructions are the means to the end. This is the majority of people. For them, it’s a problem if it’s difficult to feed the computer instructions. That wastes their time and energy at the expense of their goal. 

I suspect that LLMs will go down in history primarily as an innovation in the easing of feeding instructions to computers. It’s in the name: “language models”. Language is the medium of communication. When a person is giving instructions to a computer, they’re communicating. “Computer, can you help me find this news article?”

What I see missing from the LLM companies’ thinking is the human end goals. “You should learn AI so that you can use AI more! Why? To use AI!” The Silicon Valley companies are making the same mistake computer people tend to make: they assume that using the computer is the end goal. It isn’t. Not for most people, at least.

Someday, there might be a business leader who dares to ask, “What if we centered our focus on what AI helps people accomplish in their daily lives?” Questions like this have been asked before. But lately, we’ve forgotten about them.