My Life With AI—Part VII: The Joy Of Discovering The Hidden Value Of GenAI

Have you ever hired an intern? Besides having a good heart, you probably also hoped to get more work done. An intern can help you do it.

Hold this thought. I’ll come back to it in a moment.

Despite piquing my curiosity, I quickly abandoned the first generation of GenAI platforms. They were too inconsistent. That frustration led me to build my own local AI environment, which taught me Python but also taught me how demanding AI infrastructure really is. Ironically, while I struggled to improve my own system, the commercial GenAI platforms were improving far faster than I expected.

So why did I return to GenAI?

At first, it was because of speed and resource consumption. The advantages here should be obvious. GenAI had evolved to the point where it had significantly reduced (but not wholly eliminated) the problems that initially caused me to move away from it. It could now accomplish what I was trying to do locally at the prompting of a plain English command (called, appropriately enough, a “prompt”). That was the “speed” thing.

The “resource” thing simply refers to the fact that the computing power (namely storage space) on my laptop was limited. It wasn’t just the AI system I was building; it was the infrastructure needed to build that system. I was already pushing my disk capacity to its limits with my video production.

I couldn’t conveniently expand the disk space on my laptop and I didn’t want to switch to the desktop. GenAI platforms, which exist in the horrid cloud that I totally detest, offered welcome respite.

With that, I decided to once again venture into the undiscovered country of GenAI. Sure, it might not pass the AI test, but I quickly discovered a hidden value all the PR folks seemed to have left off their talking points list.

It’s not just speed. It is the nature of the speed.

Speed, at its rawest and most blunt, is brute force. In computing, brute force simply means throwing enormous amounts of processing power at a problem. Imagine a firefighter trying to break through a locked door to rescue a baby trapped inside. Now imagine ten firefighters attacking that same door at once. Which door gets opened first?

Certainly the one with ten firefighters. That’s brute force. That’s the kind of speed GenAI offers.

But here’s the kicker. That brute force is useless if the output offers no value. If the firemen only end up bringing out a copy of yesterday’s newspaper and leave the crying baby in the crib, does it matter how fast they break through that door?

But suppose those firefighters carry the baby safely out instead. Suddenly the speed matters enormously. Saving the right thing quickly is where brute force becomes valuable.

That has been the dilemma of GenAI. The industry promised a feast, but too often delivered little nourishment. It was all sizzle and no steak. Or at least not the kind of steak you wanted to eat.

What happens, though, when you finally learn how to order exactly the steak you want? Suddenly, getting it fast offers tremendous value.

That’s the hidden value of GenAI. If you know how to get what you want, GenAI can get it to you fast. Brute force has never been valuable by itself. It’s only valuable when guided by human judgment.

For example, I have GenAI act as a copy editor for my Commentary columns. At first, it was quite frustrating. It kept wanting me to “correct” my voice. It was like a relentless high school English teacher who insists you write “properly.”

I’d hate to see how GenAI would edit the likes of William Faulkner, Charles Dickens, or Ernest Hemingway. Each of these unique voices would be polished into the same safe, formulaic prose that so much modern best-selling writing already resembles.

OK, so maybe I need to decide if I want a distinctive voice or to sell a lot of books. Interesting. Do I want to collect old newspapers or save babies?

Eventually I realized the problem wasn’t GenAI. It was my prompt. I wasn’t asking it to preserve my voice. Once I explained that preserving my voice was part of the assignment, the results improved dramatically.

People make the mistake that GenAI represents an authority, some never-ending font of perfect knowledge. Folks shouldn’t think that way. For one, that only makes them more susceptible to fear tactics based on popular science fiction stories. More importantly, the idea that you can use GenAI as an encyclopedia of information is wrong. Don’t believe me? Then give it the Hamburger Test (see “My Life With AI – Part V: Why GenAI (And All Search Engines) Fail,” Mendon-Honeoye Falls-Lima Sentinel, October 23, 2026).

The real value of GenAI is productivity. The caveat is that the task must be one you’re already qualified to evaluate. If your work entails researching patents, GenAI can sift through the vast digital archive far quicker than you could. And, because you’ve done this yourself, if GenAI spits out garbage, you’ll notice immediately.

If someone without experience did the same thing, they might not notice the garbage. That’s what I mean when I say it only increases productivity for tasks you know.

Think of GenAI as the brightest intern you’ve ever hired. The intern works at incredible speed, never gets tired, and can digest mountains of information in minutes. Unfortunately, that same intern occasionally invents facts with complete confidence. You’d never let such an intern publish anything without your review. But you also wouldn’t waste time doing yourself what that intern can accomplish in a fraction of the time.

You gain productivity by only having to look at the finished product and avoiding the steps leading up to that finished product.

Tools amplify competence; they don’t create it.

The hidden value of GenAI isn’t that it knows more than you do.

It’s that, when you’re already knowledgeable, it lets you reach useful results dramatically faster.

And maybe that means you’ll have time to save a few more crying babies.