Sometimes to save the internet, you must eat glue

The glue pizza search result was traced back to a comment from a Redditor who went by “fucksmith” making an obvious joke on the subreddit r/Pizza.

Because I’m both brave and a genius, naturally, I had to try to make the glue pizza myself.

Now, Liz Reid, the head of search at Google, wrote in Thursday’s blog post that Google will limit the use of satirical content in its AI-generated search results (no more answers from The Onion, for instance). It will also limit its use of user-generated content for AI-generated advice (being more careful with answers from Reddit, whose content Google is paying $60 million a year for, according to reports).

There was a whole lot of attention paid in the past week about how bad many of these AI-generated search results were — particularly because they were wacky and funny. Was Google’s response to tamp down its big AI search ambitions just because a few jokesters on X made silly queries? Maybe.

Could it be because Google took the feedback seriously and realized that there were use cases they hadn’t expected and they needed to retool based on this new information? Maybe.

Could it be a combination of those two things — that a small minority of trolls abusing the system for laughs revealed some serious flaws and dangers of putting AI in search results? That it wasn’t just a PR disaster for a week but made Google seriously rethink the safety of the AI Overviews product and what it would actually be used for? Most likely.

But let’s not overlook one crucial factor here: ME! I actually ate the glue pizza. (It did not taste good, and please do not do this at home).

In my dreams, I like to believe that Google CEO Sundar Pichai saw a picture of my gaping maw ingesting polyvinyl acetate and cheese, fell to his knees, and cried out, “What have I done?!?” Sacrificing my palate and the equilibrium of microplastics in my bloodstream was not in vain or merely for clicks; it slowed the steamroller of AI that is destroying all that we loved about the old internet.

The ridiculousness of those AI answers does call into question the entire concept of using AI for Google search results.

What is the point of it — these AI-generated results, the idea that you should “let Google do the Googling for you” — instead of using your own judgment to pick the link that looks like it has the best answer? As Max Read writes, “It is possible I am in a minority here, but speaking for myself I want to see a selection of different possible results and use the brain my ancestors spent hundreds of millions of years evolving to determine the context, tone, and intent.”

Similarly, for New York Magazine, John Herrman points out that humans have become pretty good at parsing Google results on their own: “Understanding that you’ll encounter some nonsense, scams, jokes, and ads on the way to finding what you’re looking for, or realizing that you won’t, is part of the job of using Google. By attempting to automate this job, Google has revealed — and maybe discovered — just how hard it is and how alien its understanding of its own users has become.”

I don’t necessarily think that AI search results are a terrible idea and will never be good. I believe Google when it says that most people, most of the time, have enjoyed AI search results. That’s mostly been my experience, too.

The fact that Google rolled this out with such easily exploitable flaws? That was bad. But fixing it? That’s good. And I like to convince myself that my eating glue pizza was part of the noise that prompted Google to act.

Please, please … I don’t need your thanks. I’m just doing my job! As they say, not all heroes wear capes. Some just eat glue.