When I was a kid, my father purchased a thirteen-year-old Austin Healey, though by modern standards of automotive wear and tear, it looked like it might have been 130 years old. I recently found the bill of sale in a box of his things, neatly typewritten and noting a purchase price of $500.
Very slowly, between 1976 and 2014, he and I (mostly he) tinkered with the car and eventually got it back up and running. Rusted body panels were repaired with new steel, the wiring harness was replaced, and the fuel pump, carburetors, and brake lines were all rebuilt or replaced. We tackled much of the cosmetic work as well: British racing green paint, a new mohair top, sandblasted wire wheels, and re-chromed trim. It was never a particularly efficient project. There were long stretches when very little happened, punctuated by weekends spent trying to solve problems that probably should have taken a few hours.
By the time he passed away, the car was running smoothly and looked terrific. And it was incredibly fun to drive.
For those unfamiliar with old English roadsters, the experience of operating one is unlike anything on the road today. The engine temperature is perpetually problematic, and the heat coming through the firewall melts your feet even with the heater turned off. The fuel pump ticks loudly. You can’t shift into first gear unless you’re at a dead stop. There are no safety features to speak of, no air conditioning, no power steering, and certainly nothing resembling a touchscreen. The windows are operated by hand, the top requires a fair amount of wrestling, and you feel every bump in the road. Ground clearance is maybe three inches, and you’re never entirely confident that you’ll make it home without something breaking.
Driving an Austin Healey is one of the most frustrating and genuinely enjoyable experiences you can have on the road.
Nowadays, I drive a 2024 Tesla Model Y, which is about as far as one can get from a 1963 Austin Healey Mark II. It’s safe, fast, comfortable, quiet, and remarkably reliable. It requires almost nothing of me beyond pointing it in the right direction. There’s no clutch to operate, no engine temperature to worry about, and no particular reason to understand how anything under the hood works. It just works.
And while it’s fast and fun in a tech-nerd sort of way, it’s a rather soulless experience compared with bombing down a country road in autumn, top down, exhaust rumbling, wind buffeting, and occasionally wondering whether that new noise is something to worry about.
By almost every measurable standard, the Tesla is the better car. And yet, given the choice on a crisp October afternoon, I’d take the Healey every time.
I’ve been thinking about that a lot lately, particularly as I find myself using artificial intelligence more and more.
I have mixed feelings about AI. I’m excited by what it makes possible, but also apprehensive about some of its less desirable consequences, particularly for the professions and skills it may eventually displace. And yet, whatever reservations I have, I’ve become an enthusiastic user of the technology.
At Parlon, the software company I’m helping build, I’ve used AI to do things I wouldn’t have dreamed of attempting just a few years ago. Our website, for example, was built almost entirely using Claude. I’m not a web developer, and while I’ve always been reasonably comfortable around technology, the notion that I could build and maintain a sophisticated website largely on my own would once have been laughable.
Now I describe what I want, argue with the machine a little, make adjustments, and watch it materialize. Things that once would have required a developer, a designer, and several rounds of back-and-forth can often be accomplished in an afternoon. It’s extraordinary, and I would be foolish to pretend otherwise.
The same applies to writing, research, analysis, and any number of other things that occupy my working day. I can do more, faster, and with considerably less effort than I could before. In a small company where there are never enough hours or hands to do everything that needs doing, that’s an enormous advantage.
Still, I find myself wondering what we’re giving up in exchange for all this newfound efficiency.
I started my career at a time when knowing how to do something mattered at least as much as knowing what needed to be done. If you wanted to build a website, you learned enough HTML to get yourself into trouble. If you wanted to make a presentation look good, you fiddled with it until it did. If you wanted to write something worthwhile, you stared at a blank page, wrote something terrible, deleted half of it, and tried again. There were frustrations, false starts, and plenty of wasted hours, but there was also the satisfaction that came from figuring something out.
Today, increasingly, we can skip much of that process and go straight to the result. And for a great many tasks, that’s wonderful. I don’t miss spending hours troubleshooting a broken bit of code any more than I miss worrying about whether an old British fuel pump is going to leave me stranded beside the road.
Still, there’s a difference between the satisfaction of creating something and the satisfaction of having created it. I’m beginning to wonder whether we’re confusing the two.
Consider writing. I enjoy it, not because I’m especially good at it, but because it forces me to organize my thoughts. I’ll often begin with a vague idea, only to discover halfway through that I don’t entirely agree with myself. A sentence leads somewhere unexpected, an argument falls apart under scrutiny, or a memory comes back that changes the direction of the piece. The finished article is nice to have, but much of the value comes from the process of writing it.
AI can make that process enormously easier. It can improve the structure, identify weaknesses, suggest better phrasing, and spare me the embarrassment of publishing something riddled with grammatical errors. But if I hand over the entire exercise, I may end up with a perfectly respectable article without having done much of the thinking that made writing it worthwhile in the first place.
It’s tempting to describe this as a loss of craftsmanship, though I’m not sure that’s quite right. My father wasn’t restoring the Healey because he rejected modern automotive engineering. He used the best tools he could afford, bought replacement parts when it made sense, and would undoubtedly have welcomed anything that made the more miserable jobs easier. If someone had invented a machine capable of removing forty years of rust in an afternoon, I’m fairly certain he would have been first in line to buy one.
But I suspect he would have been less interested in a machine that restored the entire car while he went off and did something else.
Of course, there’s another side to this. The tools that have made building a website almost effortless have also allowed me to spend more time thinking about the business, talking with customers, and working on problems that I find considerably more interesting than debugging HTML. AI is moving the boundary between the work we have to do and the work we want to do. That’s a good thing, and in many ways it’s what technological progress has always been about.
The challenge, I suppose, is recognizing the difference between work that’s merely an obstacle and work that’s rewarding in its own right. There’s no virtue in making something unnecessarily difficult, but there isn’t much satisfaction in outsourcing every interesting problem, either.
My father and I spent nearly forty years working on a car that any modern family sedan could outperform without breaking a sweat. From a purely practical standpoint, the entire project was absurd. We could have bought something newer, faster, safer, and more reliable for a fraction of what we eventually spent bringing that old Healey back to life.
But then we would have missed forty years of working on it.
I have no intention of abandoning AI, just as I have no intention of trading in my Tesla for something that might require a tow truck every third Sunday. But perhaps we ought to be a little more thoughtful about which frustrations we eliminate and which ones we choose to keep.
Efficiency is a wonderful thing when the objective is simply to arrive somewhere. Sometimes, though, the experience of getting there is the point.
Which is why I drive the Tesla most days.
And why, on a crisp October afternoon, I’d rather have the keys to the Healey.