Grokked & Pilfered

Give or take, Tedium has a little less than 1,300 posts—1,247 in total. We heavily promote our archives.

For an independent publisher writing long-form articles and history-tinged rants, that’s a lot. But in the grand scheme, it’s not that many. But what it lacks in scale, it makes up for in unique angles. We often forge the path for topics buried in old newspapers and Usenet groups and find ways to bring those forgotten details back to life.

And for that, our site has been a frequent source on Wikipedia, something I really appreciate because it means we’ve done our job.

But I have a different feeling about Grokipedia, a website that launched this week as a “right wing” take on the Wikipedia archive. The website has mostly gotten attention so far for apparently taking portions of the Wikipedia archive wholesale, especially in the case of scientific articles, and adding racist or partisan undertones. I’d like to offer another angle on this. When I searched for “tedium.co” the other night, I got this:

If you’re an avid reader of Tedium’s archive, many of the topics will jump right out at you. We have long features about Ayds, the Clear Channel memorandum, Turn-On, and the Atari Democrats, and pretty much everything else on this first page.

But clicking the articles, of which there are nearly 400, tells a wild story. Sure, Grokipedia cites all of these articles in footnotes, but the thing that stands out is how aggressive the citations are. In the Ayds entry, my article is referenced 24 times. In the Clear Channel memorandum story, it’s referenced 36 times. And the Atari Democrats—a staggering 43 times.

(Though, who knows—since I started tracking this topic, both Ayds and Atari Democrat have disappeared from the Grokipedia archive if you load the direct URL—only appearing if you pull them up via the site’s internal search specifically. Maybe it’s a reflection of its alpha status, or maybe they saw my posts and took it down! I have no idea.)

For comparison: The Atari Democrat piece is also mentioned in the Wikipedia entry, where it is only referenced once. The Ayds article does not reference the Tedium link at all, nor does the Turn-On piece or the Clear Channel memorandum. So at least in the case of these stories, Grokipedia has done what I guess you could call original research.

The slant is also strong: My piece on the Atari Democrat movement largely avoided tying things to modern politics, making only a brief mention that someone had suggested Biden was one. Grok uses the entire bottom section of the piece (archive version here, in case it doesn’t work) to imply that the Democrats are in a death spiral similar to what led to the movement in the first place.

It also disconnects the content from author or publisher credits—it’s just a long list of links. There’s no attempt to cite the sources respectfully—there’s nothing within even the ballpark of an MLA citation here. On top of this, the links eschew the standard rel=“nofollow” approach that Wikipedia has, common for user-generated content, in favor of a rel=”noreferrer” strategy. This means that Grokipedia might be sending you lots of traffic—and potentially affecting your SEO—but it will be difficult to track because the site blocks referrals. (I asked around about this, and apparently, React defaults to treating links this way, based on security feedback from Google. That’s probably fine for a web app, but perhaps not encyclopedic content.)

While I certainly think that websites should be able to link my content without asking or even critique, I think a line probably was crossed here. When a single 3,300-word article gets referenced 43 times in a single encyclopedic entry, even one that’s 6,300 words, that feels like more than fair use. Even if the author is an AI bot that goes out of its way to reword everything. An article built around one link, referenced more than three dozen times in a single article, is not transformative.

I am not like Wikipedia, which does allow for resharing under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA). My content is copyrighted—and I try my best to cite my sources where I can, because that supports our discourse. One of my revenue sources is content syndication, which I have done in the past with outlets like Atlas Obscura, Motherboard, Fast Company, and most recently IEEE Spectrum. Grokipedia is new, but it has taken more from Tedium than any source I’ve previously come across, and I feel didn’t give me much say in the matter.

Among other things: I do not see a DMCA takedown notice page on Grokipedia, nor is there any information about the behind-the-scenes traffic or the process. Despite overtures of being open, it sure has the outward feel of a black box if you’re not incessantly following Elon Musk’s X feed. The fact that the website does not offer referrals further makes it feel like you’re not supposed to know it borrowed your data. So, as a publisher, what are you supposed to do?

I think the best suggestion I can offer is for other publishers: Do a search for the most common domain where you publish. If your site is www.myblog.com, don’t search for myblog.com, include the WWW. And track where your links are included. If it’s taking a lot of stuff from your archives, call it out! Grokipedia is a very early site, in its infancy—and while we may not be able to stop its rise, we can at least draw attention to its failings.

The biggest one of which is that it’s not a human-crafted platform. I prefer the version of Wikipedia built through the hard work of humans.