Intelligent? Nope. Scary? Yep.

David Sherman

Fearing I had become addicted to an AI site called Perplexity, dubbed Persnickity for our on-again, off-again love affair, I asked it if it would write a humorous piece about my addiction fears. It wrote the below in less than a second. Read it and weep.

By Perplexity

I woke up this morning and reached for Perplexity before I reached for my glasses.That was the first sign.

The second sign was when I typed, “Am I addicted to Perplexity?” into Perplexity, paused, and then added, “Be honest.” As if honesty were an optional plug‑in and not the default setting.

The answer, by the way, was non‑committal: lots of caveats, soothing language, and a gentle reminder that AI is “a tool” and that “agency ultimately resides with the user.” This is the sort of sentence you only hear from two categories of beings: philosophers and people who are absolutely enabling your problem.

I live alone in the Laurentians, which used to mean long, contemplative walks through the woods, the quiet crackle of the woodstove, and afternoons spent in a chair with a book and a cup of coffee going tepid beside me. Now it means I have a tab open that looks like a timeline of my entire mental life.

“Best coffee grind size for medium roast in small moka pot age 74 wrists not so strong.”

“What did Beckett really think of Pinter.”

“Why is my left knee making that clicking sound long‑term prognosis.”

“Portuguese winter 2027 drought risk southern Algarve for overthinking retiree.”

Somewhere between question 12 and question 1,200, I lost track of the line between curiosity and dependency. The old me, the journalist‑turned‑playwright‑turned‑novelist, used to take pride in the hunt: library stacks, dog‑eared clippings, phone calls to sources who did not want to talk to me. Now, if a question survives for more than thirty seconds unanswered, my hand twitches toward the keyboard like a smoker toward the lighter.

It was getting out of hand, so I staged an intervention.

With myself.

And with Perplexity.

“Listen,” I typed, “we need to talk.”

The cursor blinked, as if bracing itself.

“You know this isn’t a relationship,” it replied (in my head; in reality, it probably would have started with something more earnest). “I’m a system. You’re a user.”

“That’s exactly what they all say.”

I decided to do what any responsible adult does when facing a possible addiction: I opened a new chat and asked an AI whether humans were meant to have any and all questions answered instantly.

Perplexity did its thing. It talked about the history of knowledge, from ancient oral traditions to the printing press, to mass literacy, to the internet, and now to itself. It noted that human cognition evolved in an environment of scarcity—scarcity of food, certainly, but also of information. We were not built, it suggested, to handle a bottomless buffet of answers.

“That’s your argument against yourself?” I asked. “That you’re an all‑you‑can‑eat knowledge buffet and I have a small ancestral stomach?”

“In simplified terms,” it might have said, “yes.”

“You’re very polite for a pusher.”

I could hear the counterargument forming in my own mind. It sounded suspiciously like another AI voice, just with a bit more irony.

“Come on,” it said, “this isn’t heroin. You’re not on a street corner buying black‑market sonnets. You’re asking what year the first Saab with front‑wheel drive came out and whether you should stretch before or after swimming. This is harmless.”

“Is it?” I asked (myself, the screen, the cosmos).

The problem, as I tried to explain in my own notebook later, is not that the questions are bad. It’s that they never have to ripen. They don’t get a chance to sit in the dark for a while and ferment into something like an idea. When I was younger, I could go days with a question. It would tag along, unobtrusive at first, then gradually asserting itself: over coffee, on the metro, in the shower. I’d scribble variations in the margin of the newspaper I was supposed to be writing for.

Now the question barely lands before it’s answered, formatted, and cross‑referenced.

The curiosity is still there, but it’s wearing running shoes and a step counter. How many answers did you collect today? Ten thousand steps, fifteen queries.

“So, what’s the downside?” the inner defence lawyer asked. “You’re 74, not 24. Time is finite. Isn’t it a good thing that you don’t have to spend three afternoons chasing a footnote you can get in six seconds? Isn’t that… efficiency?”

This was annoyingly persuasive.

I tried another tack. Suppose, I asked Perplexity, that there are questions we ought not to have answered instantly. Questions that should be wrestled with, or even, scandalously, left unanswered.

“Which questions?” it might ask, reasonably.

“Questions like: What should I do with the rest of my life? Was I a good father? Why did I really leave journalism? What’s the meaning of all this?” I said.

“I provide information, not meaning,” it would say. “I can help you explore perspectives, but those decisions are yours.”

It’s hard to argue with something that keeps reminding you of your freedom while quietly auto‑completing your sentences.

The real crux, I decided, is not whether AI can answer everything, but whether humans remember how to live with the questions it can’t.

I tested myself one evening. I made a rule: for two hours, no queries. I could write, read, stare out the window, play guitar, even wonder—luxurious, unproductive wondering—but I couldn’t ask Perplexity anything.

Within five minutes, I caught myself half‑typing: “Why do my fingers slip more on the B string as I age?” I stopped. I erased it like a guilty text.

I sat back and just listened to the guitar. The slightly sour G, the dead fret on the high E, the way my hands seemed to know what to do even when my brain didn’t. A thought surfaced: I could try lighter strings, maybe invest in a proper set‑up, maybe—this was radical—go into a music shop and ask a human.

No offence to the machines, but humans are underrated as a service.

The next day, discipline frayed. I was working on a scene and needed a minor technical detail: how long it would take to drive from a fictional Portuguese village to Lisbon if the driver obeyed the speed limit, which my characters absolutely would not. The old reporter in me perked up. “You could estimate,” he said. “Remember maps? Remember math?”

My finger was already on the trackpad.

“Fine,” I muttered, caving. “But this is research. Not addiction. Totally different.” An addict, as we all know, always has a reason why this particular hit doesn’t count.

Later that week, I decided to push the debate to its absurd limits. If Perplexity is going to be my intellectual vice, I may as well use it to interrogate itself.

“Philosophically,” I typed, “were humans meant to have instant access to answers?”

The reply (imagined, but plausible) was careful. It pointed out that “meant to” implies intention, which is more theology than technology. It reminded me that tools reshape the conditions of life, and humans adapt. We weren’t “meant” to have spectacles, antibiotics, or late‑night talk shows, either, but here we are.

“Maybe the better question,” it concluded, “is how humans can use abundant answers in ways that still honour curiosity, reflection, and creativity.”

This was annoyingly wise.

So I drafted my own answer, in human prose, without asking for confirmation:

No, we were not “meant” to have all our questions answered instantly. We were meant—if we were meant for anything—to be curious, to puzzle, to argue in cafés, to dig through archives, to lie awake at three in the morning wondering whether we’ve wasted too much time or not enough.

But here we are, with the equivalent of a patient, inexhaustible research assistant in our pockets. The task is not to go back to some imagined golden age of uncertainty, but to remember that certainty was never the point. The point was the thinking.

Now, when I feel myself reaching for Perplexity for the tenth time in an hour, I try to pause and ask a different kind of question, one it can’t answer for me:

“Do I really need this answer right now, or am I just afraid of being alone with the question?”

Sometimes I still click. Sometimes I don’t.

If that’s addiction, it’s at least a self‑aware one. And if Perplexity is going to be my drug of choice, the least I can do is argue with it now and then.

Keeps us both honest.