The "Museum" at the Best Western

A Facebook page called "History Haven, " — 165,000 followers, listed on the platform as a "Museum," with a publicly posted address at 2111 N Lacrosse St in Rapid City, South Dakota — published a post on May 4 about three of the most famous Westerners who ever lived. Within a few days it had picked up over 2,000 likes, 124 comments, and 249 shares.

Two things you should know up front.

The first is that 2111 N Lacrosse St in Rapid City is the Best Western Ramkota Hotel. Unless the museum is operating out of a conference room next to the indoor waterpark, History Haven is not a museum. It is a Facebook page.

The second is that the post itself — the image at the top, and most of the words underneath it — was generated by AI from material I wrote and posted on this blog years ago.

This is the actual photograph History Haven was supposedly describing. It was almost certainly taken in early September of 1873, around the time the three men first performed together on stage as The Scouts of the Plains. Wild Bill Hickok is seated on the left. Texas Jack Omohundro is standing in the center, his right hand resting familiarly on Hickok's shoulder. Buffalo Bill Cody is seated on the right. It is the only known image of all three together.

Here is the image History Haven posted instead.

Three different men, all standing, in clothes that almost-but-not-quite resemble the originals, with faces that almost-but-not-quite resemble the real Hickok, Omohundro, and Cody. It is a sepia-toned hallucination of a photograph that already exists.

I want to walk through how I know the post is AI, because the giveaways say something about how these pages work.

The caption is describing a different photograph than the one they posted

History Haven's text identifies the three men by position: "On the left is Wild Bill Hickok... At the center stands Texas Jack Omohundro... On the right stands Buffalo Bill Cody." Those identifications are correct — for the real 1873 photograph, the one I posted on this blog years ago, the one whose composition I described in my own writeup. The AI that wrote History Haven's caption was working from that writeup. So the caption gets the positions right.

It just gets them right about a photograph History Haven didn't post.

In the actual 1873 photograph, the man on the left (Hickok) is seated and bareheaded; the man in the center (Texas Jack) is standing and wearing a wide-brimmed hat. In the AI fabrication at the top of the History Haven post, the man on the left is the one wearing the hat. The caption tells you that man is Wild Bill Hickok. In the picture above the caption, he isn't.

You only fail to notice that mismatch if you don't actually know what these three men looked like. And you don't know what they looked like if you've never looked at the photograph you're claiming to describe. The page owner ran my old writeup through a chatbot for a polished caption, generated a generic black and white image to put above it, and posted the result. The history isn't the product. The engagement is. The history is just whatever borrowed material happened to do well on someone else's page first.

The factual claims that aren't from my post are wrong

History Haven's text calls Cody "the youngest of the trio." He wasn't. Cody was born on February 26, 1846. Texas Jack was born on July 27, 1846. I don't expect History Haven to know this, but historically speaking, in the year 1846 July came after February, which made Cody five months and a day older than his friend Omohundro. Hickok, born in 1837, was the eldest by nearly a decade.

This is exactly the kind of confidently-asserted error you get when a language model is told to write atmospherically about something it doesn't actually know. The same paragraph that promotes Cody to "the youngest" also says he "carried an energy that hinted at what lay ahead" — and that's the real tell. Real history rarely needs that much soft-focus filler, because real history has actual specifics. It has facts and knowledge and provenance. AI history has energy and posture and gaze.

The structure mirrors mine, the specifics are filed off

I'm not going to go line by line, but the History Haven post hits the same beats as the writeup I published years ago: the three men's distinct contributions, the scout / lawman / cowboy framing, the through-line about how the West would be remembered. The phrasings are reworded, the citations are gone, the Earl of Dunraven quote I used to anchor the cowboy point has been sanded off, and a few new wrong things have been added in the gaps. That is what you get when you feed someone else's research to a chatbot and ask it to rewrite this for Facebook.

Why does any of this matter?

It matters because History Haven's version got 2,000 likes, 124 comments, and 249 shares. Posts I make these days — written by a human who has spent years on these men, who has written a book about Texas Jack, who can tell you who the other men are in the only other extant image of these three men — does not get those numbers. Almost nothing written by actual human historians does anymore.

The pages that win in the algorithm are the ones that can produce a polished-looking picture and a frictionless paragraph in under two minutes, then post a half-dozen of them a day, then monetize the audience on the back end. They aren't interested in being right. They're interested in being upstream of an ad. The history is just bait.

I've watched this accelerate over the last couple of years and I'm not going to pretend it isn't discouraging. There are days I wonder what the point is of spending a week on a piece that a bot will paraphrase in twenty seconds for a page with eighty times my reach. I have, honestly, mostly stopped trying to compete on that front.

But the thing I keep coming back to is this: if you're reading these words, you are part of whatever answer there is. Be skeptical of any "history" page that posts more than once a day. Reverse-image-search the photographs. When the prose drifts through atmosphere instead of telling you something specific, it usually is drifting because there is nothing specific underneath. And when you do find someone doing the work — a real historian, a museum that exists at the address it lists, a writer who can respond when you ask a question or tell you why a particular detail caught their eye — share their post instead of the slop. That's how the algorithm starts to learn there's an audience for the real thing.

History Haven is not a museum. It is a content farm pretending to be a museum listing an address in a hotel room. And Wild Bill Hickok, Texas Jack Omohundro, and Buffalo Bill Cody? The three men in that 1873 photograph? They deserve better than what was posted under their names last Monday morning, and so do the people who showed up to read about them.