JamesKonderla.net

JamesKonderla.net

Original Works, Blog and Random Thoughts

The Recipe is at the Bottom

The Irony of Modern Search

Scrolling for the Chicken

We have built the most powerful information-retrieval tools in human history. I can ask a machine to summarize a 300-page report, translate it into Portuguese, and explain the statistics in the appendix. Yet when I want to know how long to roast a chicken thigh, I still scroll past a childhood memory, a trip to a farmers market, six photos of the same plate, and three autoplay videos before I find the number.

That is the cosmic joke at the center of this post. The truth, verified and edited, increasingly costs $20 a month. The free version is a labyrinth whose walls are made of 1,000-word SEO-optimized recipe pages that don’t tell you how to cook the chicken until the final paragraph.

The recipe page is funny, but it is also a small, familiar model of a much bigger problem I wrote about in The Toll for Truth: the open web is shaped by what pays, and what pays is rarely what the reader came for.

Why the Recipe Is at the Bottom

The wall of text is not an accident. It is a business model. Food blogger Elizabeth Lindemann explained it candidly in 2020: ads are her largest source of income, and “the longer the content is, the more advertisements can be inserted into the post” (Lindemann, 2020). Most of her traffic comes from Google, and placing the recipe at the bottom “helps users stay on the page for more time, which can help with SEO.”

To be fair to bloggers, she also notes that the body text often answers real questions: substitutions, troubleshooting, what went wrong the first time. Some of it is genuine craft. But the structure is set by two outside forces: an advertising market that pays by the scroll, and a search engine that rewards time on page.

There is also a popular theory that the stories exist for copyright. It is half right. The U.S. Copyright Office states that “a mere listing of ingredients or contents, or a simple set of directions, is uncopyrightable,” while “the written description or explanation of a process” may be protected (U.S. Copyright Office, 2021). The part of the page the reader actually needs is the part the law treats as free to copy. Keep that in mind for the next section.

In other words, the reader is not the customer. The reader is the inventory. The page is built to hold attention long enough to sell it, and the answer is the bait at the end.

Then the Machines Learned to Cook

The SEO recipe page was annoying, but at least a person usually stood behind it. Someone tested the recipe, took the photos, and answered the comments. That is changing.

In August 2026, Pew Research Center found significant signs of AI authorship in 10% of a random sample of 10,000 web pages, and in more than a third of pages published since ChatGPT’s release (Bestvater et al., 2026). On .com pages, the share rose from about 1.1% in January 2021 to 9.35% in January 2026. Pew is careful to note that detectors are imperfect, and I argued in Poisoning the Well that we should treat these numbers as estimates rather than facts. The direction, though, is hard to miss.

Recipes are an easy target. Remember the copyright detail: the ingredient list and simple steps are free to copy. When Bloomberg reporters spoke with 22 independent food creators in late 2025, at least four had found AI-generated copies of their recipes on other domains. Pinch of Yum’s Bjork Ostrom found a German-language AI mirror of his entire site, with altered photos of his own family (Alba & Arroyo, 2025).

The copies are not always edible. Inspired Taste’s Adam Gallagher called the AI results “Frankenstein AI recipes,” stitching his ingredients to other blogs’ instructions. Easy Peasy Foodie’s Eb Gargano saw an AI summary of her Christmas cake recommend baking a 6-inch cake for 3 to 4 hours (Alba & Arroyo, 2025).

Meanwhile, the people who wrote the real recipes are losing the traffic that paid for them. Gargano’s turkey recipe traffic was down 40% year over year. Carrie Forrest of Clean Eating Kitchen lost 80% of her traffic and revenue in two years, went from about ten staff to zero, and is closing her business. Google tested AI recipe summaries in search, called Recipe Quick View, and ended the experiment in July 2025 after food writers warned it would cut their traffic and ad revenue (Tameez, 2025). AI Overviews, however, are still answering the question before anyone clicks.

So the joke gets a second punchline. We spent a decade complaining that the recipe was buried at the bottom of the page. Now the page might not have been written by a cook, and the cook might not be in business next year.

Truth Costs $20 a Month

Here is the irony, stated plainly. The verified version of almost anything, whether tested recipes, edited news, or peer-reviewed research, increasingly sits behind a subscription. The free version is a maze built for advertisers, and it is now being refilled by machines.

In The Toll for Truth Part 6, I described this as three tiers. The first is the verified record, available to people who pay or have an institutional login. The second is derivative access: headlines, snippets, screenshots, and free AI summaries that point toward the record without delivering it. The third is the subsidized and synthetic layer, produced cheaply and paid for by ads (Konderla, 2026). The recipe page lives in all three at once. The tested recipe is in the cookbook or the paid app, the AI Overview is the snippet, and the content-farm copy is the third tier.

Who pays for the first tier is not random. Pew found that only 17% of Americans paid for news in the previous year: 30% of adults in the highest income tier, compared with 8% in the lowest (Pew Research Center, 2025, as cited in Konderla, 2026).

Everyone else pays in a different currency: time. Scrolling past the ads, checking whether the cake really bakes for four hours, comparing three versions of the same recipe to see which one a human wrote. Call it a time tax. For a recipe, it costs a ruined dinner. For a medical question, a court date, or a news story before an election, the stakes are higher, and the filtering work falls hardest on the people with the least time and money to spare.

Finding the Chicken

I don’t think the answer is to blame food bloggers. Most of them are doing exactly what the ad market and the search engine pay them to do, and many are now losing even that. The problem is a system in which the free path is shaped for advertisers and the verified path is shaped for subscribers.

A few things help in the meantime:

  • Support the cooks you trust directly. Buy the cookbook, join the newsletter, or use the “Jump to Recipe” button so their ads still load.
  • Check for a human. Real recipe pages usually have a named author, original photos, and comments from people who made the dish.
  • Use your library. Many public libraries offer free access to cookbooks, magazines, and news databases that would otherwise cost a subscription.

The bigger question is one I want to measure rather than just joke about. How much time does the free path cost compared with the paid one, and who ends up paying it? That is the Time Tax, and it may be the next thing I research.

Sources

In

, ,
, , , , ,