What AI Made Cheap, and What It Made Priceless

Two years ago, I read Arthur Brooks's book “From Strength to Strength.” A friend had gifted it to me while I was considering launching a new online learning business. The last one I launched at 38, so doing it again at 56 was scary to say the least. The typical questions: What if I blow through a large chunk of my savings with no real recourse at my age? What if my cancer returns and I have to re-gear my whole life back into therapy, and worse yet, I leave my wife encumbered with unwinding a business she knows nothing about? And at 56, shouldn't I be traveling and seeing the world anyway? My friend felt a #1 NY Times bestseller might give me some clarity and direction.

Well, I never got any.

Arthur Brooks asserts that from our late 30's on into our 50's, our brain power shifts from fluid intelligence, our analytical horsepower and ability to innovate, to crystallized intelligence, our ability to synthesize what we've built over years of experience. The advice that follows from that [and I’m paraphrasing here]: accept the decline early, and move from the notion of “success” to that of “significance”, investing more in love and deep connection than accomplishment, and treating the back half of life as a different and arguably richer form of contribution rather than a decline.

While the book had plenty of clinical research to back its claims, the message never really sat well with me. All it did was create more questions and exacerbate my career confusion. Ultimately, I took the plunge and launched a year after reading that book. Absolutely no regrets, but admittedly, many of Brooks's assertions on processing capacity and mental speed have turned out to be true. I'll get into the trials and tribulations of my first year as a 50-something entrepreneur in an upcoming newsletter.

Then last month I set off to explore an idea that had been nagging at me for a while. Does AI level the playing field between the 20-somethings acing software at breakneck speed and us slow-but-wise old sloths? So I ran an experiment. Could I truly write a program using Claude Code and hit my goals in a few afternoon hours, as the hype claims? Because if indeed I could, then there is something to be said for the playing field getting quickly leveled. Further, the odds of building a successful business will in fact have turned in our favor, as the mechanics of mental exertion get trivialized, leaving experience and skill as the only differentiator.

So for the experiment, I had a few rules: 1) it had to have a persistence layer, a real database that enabled me to retrieve data; 2) it had to solve a problem outside my realm of knowledge and expertise; 3) it had to have a user layer that met the highest usability standards; and 4) Claude had to guide me through every step of the process from scoping to coding, to testing, to deployment, and ultimately, a live site.

The problem I was solving was simple. Collect every non-stop round-trip airline fare from New York to London over the past 90 days using available free data sources, and collate the information to tell me the lowest, highest, and average fares over that stretch. Then estimate the range of fares I can expect over the next 7 days with a recommendation on the day, time, and specific site I should go to to buy. Put it all on a graph so I can see the trend. Save it all so I can retrieve it anytime.

So I set off to run the experiment, and voila!! 3 hours later, I actually had a live public website (crude but live and error-free nonetheless) fully deployed on GitHub. I was absolutely blown away by what I was able to accomplish on that Friday afternoon. I started at 1:30pm and had it live by 4:30pm.

What does that all mean?

Here's the part that stings a little. Brooks was right about me. I am slower than I was at 35, and I feel it. But here's what he missed. That Friday afternoon, I wasn't the one supplying the speed. Claude was. All that fluid-intelligence horsepower, the fast building, the code I couldn't write to save my life, the analytical grind that would've taken me weeks and a 24-year-old a day, the machine did all of it. So yes, my fluid intelligence is in decline. The trouble for Brooks is that everybody's just got a whole lot cheaper, and when the thing that's declining is suddenly cheap and everywhere, the decline stops mattering much.

What Claude couldn't do was decide why a New York-to-London fare tracker was worth building in the first place. It couldn't tell me why I needed the “highest usability” in the way the data was being displayed. It couldn’t have scoped the problem the way I needed it. That's crystallized intelligence. That's the exact stuff Brooks says we spend 30 years piling up.

So here's where I've landed. His research can be dead-on and his conclusion still be wrong, at least for now. His argument assumes your fluid intelligence stays valuable while your supply of it drops, so every year you're worth a little less. But AI just knocked its value toward zero for everybody, young and old. What's left as the one real differentiator? The one thing that only goes up with age.

Which, if you follow it out, flips Brooks on his head. The back half of life isn't some graceful retreat from building into “significance.” It might be the first time in a career that all that hard-won experience compounds instead of quietly depreciating, because the execution that used to bottleneck you is now handled for you.

So if you're sitting on 25 or 30 years of judgment, quietly assuming you missed your window to build something of your own, I'd think hard about that. You're not too late. I know I’m not. Odds are you're showing up at exactly the right time, holding the one thing that just got scarce.