Innovating with AI - Prompt engineering doesn't matter

Hey Innovator,

The common wisdom among AI enthusiasts is that “prompt engineering” is the mission-critical skill of the future.

After all, if AI is taking over everything, the people who know how to talk to AI are going to see huge career growth. Just ask The Washington Post, which says:
​
“Tech’s hottest new job: AI whisperer. No coding required.”

Or Forbes, which adds:

“The Hot New Job That Pays Six Figures: AI Prompt Engineering.”

There’s just one little problem: the days of prompt engineering are already over.

That’s not because AI isn’t a big deal – as we’ve all noticed, the use of language models, image models and (most recently) video models is spreading like wildfire into every nook and cranny of our economy and society.

Instead, it’s because the skill of writing AI prompts, which seemed like science-fiction just 18 months ago, has already become a commodity that everyone will (soon) know how to do.

Let’s contrast this with the hot tech skill of the last generation: coding. Until very recently, simply “learning to code” was widely viewed as a ticket to career success, to the point that even baby toys were marketed as ways to teach your kid to code. (My favorite is the code caterpillar.)

Of course, AI has shaken up software development, with tools like GitHub Copilot making senior coders much more efficient and obviating the need for a lot of junior-coder roles. But even so, coding had a really solid, decades-long run, during which it was an in-demand skill that was legitimately difficult to master – which meant that good coders could fetch high prices for their time and attention.

But prompt engineering isn’t like coding.

Instead, the best comparison is typing.

Flash back a few decades, and “typist” was a very important, in-demand skill. In fact, it was very common to hire people whose entire job was typing things. The Federal Reserve even tracks “word processors and typists” as a distinct job category – here’s the graph of typist employment over the last 23 years:

In 2000, there were 282,000 Americans employed as typists. Today, 89% of those jobs no longer exist.

What happened?

There certainly isn’t less typing happening today than there was 24 years ago – in fact, there’s clearly been an explosion of computer use (and thus a huge increase in the amount of typing being performed).

Instead, typing has just become a skill that is baked in by default to all other jobs.

That means far fewer discrete typists.

And as AI permeates everything, prompt engineering will become a basic, universal skill as well.

We’ll teach kids how to write prompts in third grade, just like we do for typing today. (My 10-year-old son is already getting good at image-generation prompts!)

And everyone in the current workforce will be required to learn, just as they learned to type a few decades ago.

Even more crucially, prompt engineering is a skill that builds naturally on existing language and communication skills, so the learning curve is significantly shorter than it is for something like typing, which requires memorization of an arbitrary set of motor skills. Prompt engineering simply isn’t much different from talking or writing, which makes it relatively easy to learn (and, as a result, makes it relatively low-value in the long-term).

Yes, two years ago, the idea of writing a prompt was foreign to nearly all of us. But today, OpenAI has distilled it down to a few pages of documentation – and reading that is really all you need.

So, when readers ask me why I rarely explicitly “teach” prompt engineering, this is why.

It’s not really the skill of the future – instead, it’s quickly becoming a table stakes resumé requirement, like being able to use e-mail or Microsoft Word. The skills that IWAI students learn are the bigger and more valuable ones – like rapidly generating and prototyping new ideas, and building AI products or processes in ways that result in large-scale businesses without the risk of getting crushed by Big Tech.

That’s what separates today’s AI pioneers from the rest of the pack.

See you next time,

– Rob Howard
Founder of Innovating with AI

​

PS. I'm putting the finishing touches on my next AI course. It's called "AI Idea Pipeline".
​
I started putting it together when I realized that the first step to launching an AI idea in 30 days was [insert drumroll 🥁 here] to make sure you had a really valuable idea to work on.
​
If you want to be the first to hear about it, click here to join the waitlist for AI Idea Pipeline.