How to Eat Glass: Google’s Big AI Launch of Meta is to Make Sense of Facebook, Instagram, and the Future of AI Search

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. You run into a problem when you get to eat your cheese, because the cheese falls off right away. You turn to Google for a solution that’s frustrating.

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.

Since it was launched in May of this year as the search Generative Experience, and since it has been tested for a year, there is an odd situation here.

Take Google I/O’s big launch of this feature, for instance. The demo was highly controlled, and yet, it delivered a questionable answer about how to fix a jammed film camera. It said that they should open the back door and gently remove the film, but that they should not do that if they want to ruin your photos.

Look, it’s a tough spot for Google to be in. Bing went big on AI before Google did with Satya Nadella’s famous “we made them dance” quote, OpenAI is reportedly working on its own search engine, a fresh AI search startup is already worth $1 billion, and a younger generation of users who just want the best experience are switching to TikTok. The company is clearly feeling the pressure to compete, and pressure is what makes for messy AI releases. One of the things Meta had to do when it released its artificial intelligence system was to convince people to eat glass. Sounds familiar.

Companies developing artificial intelligence are often quick to avoid taking accountability for their systems with an approach much like a parent with an unruly child — boys will be boys! The companies claim that it is not their responsibility to predict what this machine will do.

That is a problem for users. The future of search, according to the company, was artificial intelligence. What’s the point, though, if the search seems dumber than before?

AI optimists argue that we should embrace the hype because of the rapid progress made so far, trusting that it will continue to improve. I really do believe that this technology will continue to get better, but focusing on an idealized future where these technologies are flawless ignores the significant issues they currently face — and allows companies to continue delivering subpar products.

Why AI is so cool: Why AI should be dumb? Why AI shouldn’t be dumb compared to AI, or how Google has gotten better at AI

“[These models] are constitutionally incapable of doing sanity checking on their own work, and that’s what’s come to bite this industry in the behind,” Marcus said.

The feature currently available is a small part of what the company announced last week. There is a lot of ambition here, for example, multistep reasoning and the ability to create anAI-organized results page. But right now, the company’s reputation hinges on just getting the basics right, and it’s not looking great.

“You actually need to do some reasoning to decide: is this thing plausible? Is this source legitimate? Marcus said that artificial general intelligence may be required to do things like a human fact checker. Marcus and Yann LeCun agree that large language models that power current artificial intelligence systems will not be what creates AGI.

“A company once known for being at the cutting edge and shipping high-quality stuff is now known for low-quality output that’s getting meme’d,” one AI founder, who wished to remain anonymous, told The Verge.

But Pichai has also said that Google’s brought the cost of delivering AI answers down by 80 percent over that same time, “driven by hardware, engineering and technical breakthroughs.” It appears that the technology wasn’t ready for that kind of work.

Many users on social media are talking about how weird the product is, from telling them to glue on their pizza and to eating rocks. 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.