Google’s AI Search and “Web” View

Ernie Smith (Hacker News):

Simply put, Google has started adding “AI overviews” to many of its search results, which essentially throw pre-processed answers that often do not match the original intent of the search. If you’re using Google to actually find websites rather than get answers, it $!@(&!@ sucks. Admittedly though, it’s not the first time Google has adulterated its results like a food manufacturer in the 19th century—knowledge panels have been around for years.

But in the midst of all this, Google quietly added something else to its results—a “Web” filter that presents what Google used to look like a decade ago, no extra junk. While Google made its AI-focused changes known on its biggest stage—during its Google I/O event—the Web filter was curiously announced on Twitter by Search Liaison Danny Sullivan.

[…]

Google does not make it easy, because its URLs seem extra-loaded with cruft these days, but by adding a URL parameter to your search—in this case, “udm=14”—you can get directly to the Web results in a search.

John Gruber:

Safari, uniquely amongst popular web browsers, doesn’t allow users to configure custom search engines. There are ways to get custom search engines in Safari using extensions — Kagi, my default search engine of choice since late 2022, does just this — but it’s kludgy. Why doesn’t Safari support adding custom search engines like every other browser does?

On the Mac, I initiate most web searches from LaunchBar, not Safari’s location field, and LaunchBar makes it trivial to add a custom search using this &udm=14 URL trick. Similar utilities like Alfred and Raycast do too. The downside compared to LaunchBar’s built-in Google search action (and Safari’s location field) is that a simple custom query URL doesn’t provide as-you-type suggested results.

Jeff Johnson:

Can you perform the trick with StopTheMadness Pro? Yes! Use the redirects feature.

John Gruber:

Expert users won’t need this site, but typical users might love it as their home page.

Kylie Robison:

Imagine this: you’ve carved out an evening to unwind and decide to make a homemade pizza. You assemble your pie, throw it in the oven, and are excited to start eating. But once you get ready to take a bite of your oily creation, you run into a problem — the cheese falls right off. Frustrated, you turn to Google for a solution.

“Add some glue,” Google answers. “Mix about 1/8 cup of Elmer’s glue in with the sauce. Non-toxic glue will work.”

So, yeah, don’t do that. As of writing this, though, that’s what Google’s new AI Overviews feature will tell you to do. The feature, while not triggered for every query, scans the web and drums up an AI-generated response. The answer received for the pizza glue query appears to be based on a comment from a user named “fucksmith” in a more than decade-old Reddit thread, and they’re clearly joking.

John Gruber:

We’re all rightly dunking on the Elmer’s Glue suggestion, but it’s just as wrong to suggest mixing cheese into the sauce. No one does that.

crumbler:

I thought AI Overviews would be disastrous but I never imagined they would be this funny

This answer apparently came from The Onion.

Matt Birchler:

What the AI responses have done for me is add more bullshit above the actual search results I want. Now I’m scrolling past the AI vomit at the top of the page, then past the ads, and then to the links that get me what I want. They’re pushing the valuable content lower and lower down the page, which is driving me nuts.

To their credit, sometimes the AI answers are useful, and they do a decent job of linking to the source that gave them the info that appeared in the AI vomit, but the hit rate is too low in my experience. Not to mention those answers take a few seconds to load, so I’m often scrolling down to the web results since they’re available instantly and I don’t have to wait to maybe get the right answer.

Previously:

• Reddit AI Training Data and IPO
• Unlimited Kagi Searches
• Brave Search Public Beta

Update (2024-05-28): Kylie Robison (Hacker News):

The messy rollout means Google is racing to manually disable AI Overviews for specific searches as various memes get posted, which is why users are seeing so many of them disappear shortly after being posted to social networks.

It’s an odd situation, since Google has been testing AI Overviews for a year now — the feature launched in beta in May 2023 as the Search Generative Experience — and CEO Sundar Pichai has said the company served over a billion queries in that time.

Maxwell Zeff (Hacker News):

In my experience, AI overviews are more often right than wrong. However, every wrong answer I get makes me question my entire experience on Google Search even more – I have to asses each answer carefully. Google notes that AI is “experimental” but they’ve opted everyone into this experiment by default.

[…]

What is clear is that Google felt pressured to put its money where its mouth is, and that means putting AI into Search. People are increasingly choosing ChatGPT, Perplexity, or other AI offerings as their main way to find information on the internet. Google views this race existentially, but it may have just jeopardized the Search experience by trying to catch up.

Update (2024-05-29): John Gruber:

LLM-powered search results are a bauble. The trust Google has built with users over the last 25 years is the most valuable asset the company owns. Google most certainly does have a choice, and they’ve chosen to erode that trust just so they can avoid accusations that they’re “behind”.

Alfred Artificial Intelligence Google Gemini/Bard Google Search LaunchBar Raycast Safari Search StopTheMadness Web

5 Comments RSS · Twitter · Mastodon

Leave a Comment

Name

E-mail (will not be published)

Web site

Δ