Consulting the Oracle

In Isaac Asimov's Foundation, there's a character named Lord Dorwin. 

Dorwin is a Chancellor of the Galactic Empire. Cultured. Educated. He considers himself a serious archaeologist, working on what Asimov calls in the novel the "Origin Question"—which planet humanity originally came from.

His method is straightforward. He reads the works of long-dead scholars and compares their conclusions. He weighs one authority against another, balances the disagreements, analyzes the conflicting statements, and decides which interpretation seems most plausible.

He calls this the scientific method.

The problem is that Dorwin has never done any original research. He’s never visited an archaeological site. Never tested a hypothesis. Never uncovered evidence. His entire intellectual output consists of evaluating the opinions of other scholars.

And nobody around him finds this strange, because they've all stopped doing original research too. It's a galaxy-wide deterioration—and nobody notices because everyone's doing the same thing.  Entire fields had quietly forgotten the difference between inquiry and commentary.

Asimov published that in 1951. 

My mother gave me a copy of Foundation when I was a kid and Lord Dorwin never really left me. I'm pretty sure I wrote an essay on it in college for a political science class. And I've been thinking about him a lot lately.

The Word We Broke

I keep hearing the word "research" used in ways I wouldn't have recognized growing up.

People say “do your own research” when they mean “find something that confirms your position.” “I researched it” increasingly means someone typed a question into a search bar and read what came back. I hear it on social media. I hear it in the news. I hear it from people I know. And I catch myself wondering — when did "looking something up" become "research"?

That would have sounded strange to me growing up.

Research — at least the way I understood it growing up — was a process. 

Research used to imply friction. You had a question you couldn't answer from a book or a summary. You figured out how to go find the answer yourself. 

Scientists ran experiments. Historians worked through primary documents. Journalists conducted interviews. Even schoolchildren were expected to gather evidence themselves for  a school project—the common thread was that you went and found out.

The defining feature of research wasn’t access to information. It was the production of new understanding.

Somewhere along the way, we blurred the distinction between retrieving information and investigating reality.

The Infrastructure of Irony

It's ironic.

The internet was originally built to connect researchers. Not consumers. Not browsers. Researchers.

ARPANET, launched in the late 1960s, was designed to link computers at research institutions so that scientists could share data, collaborate on problems, and access each other's computing resources. BITNET was founded in the early 1980s between CUNY and Yale to do the same thing for universities. These were closed communities of scholars doing original work. The tools were purpose-built to make primary research easier.

Over time, the infrastructure remained, but the behavior changed. The network designed to accelerate original inquiry became the network that made it easy to skip original inquiry entirely.I don't think that's anyone's fault, exactly. Each step along the way was a genuine improvement in access. Search engines were useful. Wikipedia was useful. AI assistants are useful. But each layer also increased the distance between people and primary sources while making the experience feel more authoritative.

But each step also made it a little harder to notice you weren't doing the thing you thought you were doing.

The Layers of Removal

When I was a kid in the 1960s and 70s, we had encyclopedias. World Book. Britannica. Nobody said "I researched it in the World Book." You said "I looked it up." You knew the difference.

The internet changed the form but not the function. Summaries stopped looking like summaries.. They looked like articles. Wikipedia added citations at the bottom. Most people never clicked the sources, but the existence of citations made the experience feel rigorous. The line between "I looked it up" and "I researched it" started to blur.

Then came LLMs.

Ask ChatGPT, Claude a question and you receive polished prose instantly.. Sometimes there are citations.  The tone rarely changes whether the answer is deeply sourced, partially inferred, or hallucinated altogether. 

I notice it in my own workflow. I'll ask Claude something and get a polished answer and briefly feel as though I understand the subject more deeply than I actually do. I have to remind myself: that's not a source. That's a synthesis of sources I haven't seen. The convenience is real. The danger is that convenience feels like competence.

What I keep coming back to is that each generation of tool made the experience feel more like research, not less. The kid citing Wikipedia thought they were writing a research paper. The person prompting ChatGPT thinks they're doing analysis. Dorwin thought he was doing archaeology.

In The Limits of Magic, I wrote about how the computer still only does what you tell it—how the appearance of intelligence masks a fundamentally mechanical process. This feels like the same pattern, one layer up. The appearance of research masking the absence of it.

Consulting the Oracle

So what are we actually doing when we "research" something online?

The closest historical analogy I can find isn’t a library. It’s an oracle.

The Oracle at Delphi didn't show her work either. You asked a question and received an answer. You interpreted it. You acted on it. There was a priestly class that mediated the experience. You weren’t expected to see the underlying process. 

Google is a priest. 

Wikipedia is a priest. 

ChatGPT is a priest so polished you forget there's a temple involved at all.

What strikes me about oracles, though, is that nobody who consulted one confused it with research. They knew they'd asked for guidance. The distinction was clear because the form demanded it—you went to a specific place, you performed a specific ritual, you received a specific pronouncement. The act of consulting was visible.

We've lost that visibility. The ritual has been streamlined out of existence. And without it, it is easy to confuse receiving an answer with understanding where the answer came from.

I don't know what the fix is. Maybe the word "research" is too far gone to reclaim. Maybe we just need to get more honest about what we're doing when we type a question into a search bar—calling it what it is instead of what we wish it were. Or maybe the answer is simpler than that: just remember the difference. That might be enough.

The Muscle

What's decaying isn't access to information. We have more access than any civilization in history. What's decaying is the instinct to ask: What is the underlying source here?

That reflex—"where did this actually come from?"—atrophies when every tool you use is designed to make the question feel unnecessary.

In How To Ask Good Questions, I wrote about the difference between questions that clarify facts and questions that open up further exploration. Oracle-consulting collapses that distinction. You ask, you receive, you're done. There's no natural next step that nudges you to go verify, to find the primary source, to check whether the summary got it right.

Asimov saw this clearly enough in 1951 to use it as a diagnostic of civilizational decline. Dorwin wasn't a villain. He was a symptom.

A society doesn’t decline intellectually because it loses knowledge. It declines because it loses the habits required to generate new knowledge.

The Galactic Empire still had libraries.

What it had lost was inquiry.

Berkson's Bits

PR is a process. You know who hits it big overnight with a story on the 'Today' show? 

Criminals, naughty celebrities, and shark-attack victims. Do you want to be one of them?

Most brands are better off compounding attention slowly.

What I'm Watching…

When the Lyft driver I had in New Orleans heard that I was from NYC, he gave me a seminar on how there’s an accent in New Orleans called “Yat” that’s very similar to a Brooklyn accent. He told me to check it out. I gotta give him credit. It surprised me. Watch for yourself and tell me what you think. 

Research is how new understanding gets created. It's how we grow. When all we're doing is summarizing old research—or summarizing summaries of old research—that's not growth. That’s recycling authority and over time, it is decay. Asimov knew it. He just set it in space so we could see it more clearly.

I'm not predicting anything that dramatic. But I am noticing this shift, and I'm curious whether you've noticed it too. When was the last time you went back to the primary source instead of trusting the summary?

Do you remember what prompted you to go look?

I'm genuinely asking. I'd love to hear.

Looking forward to continuing the conversation...

Alan