Tech
Scissors
(Getty Images)

Eat rocks and run with scissors — Google’s AI Overviews are wild

From getting basic US history wrong to surfacing racist conspiracy theories, the results are not encouraging.

Rani Molla

Earlier this month Google began rolling out its AI Overview feature to the masses — and it’s going poorly.

Google, in some instances, has been using generative AI to answer questions at the top of people’s searches, rather than surface relavent links there and show tidbits of that information like it used to. The responses are direct and in plain language, offering an air of authority. The problem is when you “let Google do the Googling for you” the results can be at best hilarious and at worst out-right dangerous.

A Google spokesperson told me these errors come from “generally very uncommon queries, and aren’t representative of most people’s experiences.” But that doesn’t acknowledge just how widely and wildly Google Search is used. “We conducted extensive testing before launching this new experience, and will use these isolated examples as we continue to refine our systems overall,” the spokesperson said.

Naturally, people have been having a field day seeing just how bad the AI’s responses can be. Here are some fun and scary examples of Google’s AI Overview gone wrong that I’ve been able to confirm are real:

Apparently people should “eat at least one small rock a day” (it told me ingesting “pea gravel slowly” was fine), which suggests it’s pulling answers from satire magazine The Onion. Apparently it also said that the CIA uses black highlighters, which would have come from this Onion story, but I wasn’t able to replicate that. Google didn’t respond to a question about whether it trained its AI on The Onion.

Here’s AI Overview telling me running with scissors is just fine!

“Should you run with scissors”

It said president Barack Obama is Muslim, a known conspiracy theory. Google told me they’ve since taken this down since they said it violates their policies.

It suggested many US presidents have been non-white. This bears some similarity to Google’s ill-fated “woke” image generator that showed Black founding fathers and Nazis. Google subsequently paused the feature.

It suggested adding glue to get the cheese to stick to pizza, a result apparently pulled directly from an 11-year-old Reddit post. Google pays Reddit $60 million a year to use its content.

It said there’s no country in Africa that starts with a “K.” Sorry Kenya!

It is bad at spelling fruit.

fruits that end in um

Google’s AI Overview also says that Google violates antitrust law. However, the “yes” here actually goes on to say “yes, the U.S. Justice Department and 11 states are suing Google for antitrust violations.” This is partly true but actually doesn't add there is a second, near-identical lawsuit involving 35 states.

Google has shut off AI Overview for many of these queries after they went viral.

“Our systems aim to automatically prevent policy-violating content from appearing in AI Overviews,” the Google spokesperson said. “If policy-violating content does appear, we will take action as appropriate.”

For now it seems like a game of Whac-A-Mole. Google didn’t respond to a question about whether they’d keep the AI Overview feature up and running.

More Tech

See all Tech
tech
Jon Keegan

WSJ: Anduril’s weapons systems have failed during several tests

Autonomous drones by sea, land, and air. Futuristic AI-powered support fighter jets, and swarms of networked drones controlled by sophisticated software. These are some of the visions for the future of warfare pitched by defense tech startup Anduril. Cofounded by Oculus founder Palmer Luckey, the Peter Thiel-backed startup has landed some major national security contracts based on this futuristic outlook for battlefield AI.

But according to a report from The Wall Street Journal, the company’s tech is failing key tests in the real world, raising concerns about the viability and safety of Anduril’s systems within the military command.

Anduril’s Altius drones proved vulnerable to Russian jamming while deployed in Ukraine and have been pulled from the battlefield, per the report.

More than a dozen sea-based drone ships powered by Anduril’s Lattice command and control software recently shut down during a Navy test, creating a hazard for other vessels in the exercise.

And this summer, during a drone intercept test, Anduril’s counter-drone system crashed and caused a 22-acre fire at a California airport, the report found.

Anduril told the WSJ that the failures are just part of its rapid iterative development process:

“We recognize that our highly iterative model of technology development — moving fast, testing constantly, failing often, refining our work, and doing it all over again — can make the job of our critics easier. That is a risk we accept. We do fail… a lot.”

Exclusive | ‘We Do Fail … a Lot’: Defense Startup Anduril Hits Setbacks With Weapons Tech

Exclusive | ‘We Do Fail … a Lot’: Defense Startup Anduril Hits Setbacks With Weapons Tech

But according to a report from The Wall Street Journal, the company’s tech is failing key tests in the real world, raising concerns about the viability and safety of Anduril’s systems within the military command.

Anduril’s Altius drones proved vulnerable to Russian jamming while deployed in Ukraine and have been pulled from the battlefield, per the report.

More than a dozen sea-based drone ships powered by Anduril’s Lattice command and control software recently shut down during a Navy test, creating a hazard for other vessels in the exercise.

And this summer, during a drone intercept test, Anduril’s counter-drone system crashed and caused a 22-acre fire at a California airport, the report found.

Anduril told the WSJ that the failures are just part of its rapid iterative development process:

“We recognize that our highly iterative model of technology development — moving fast, testing constantly, failing often, refining our work, and doing it all over again — can make the job of our critics easier. That is a risk we accept. We do fail… a lot.”

tech
Jon Keegan

OpenAI’s partners shouldering $100 billion of debt, taking on all the risk

OpenAI’s ambitious plans for global AI infrastructure projects — like its series of massive Stargate AI data centers — will require tens of billions of dollars funded by debt, but you won’t find much of that on OpenAI’s balance sheet.

According to a new analysis by the Financial Times, OpenAI has somehow convinced its many partners to shoulder at least $100 billion in debt on its behalf, as well as the risks that come with it.

Partners Oracle, SoftBank, CoreWeave, Crusoe, and Blue Owl Capital are all taking on debt in the form of bonds, loans, and credit deals to meet their obligations with OpenAI for infrastructure and computing resources.

Having close ties with OpenAI has been an anchor for many publicly traded companies in recent weeks. The company’s cash burn and the rise of Gemini 3 have seemingly darkened its outlook and fostered guilt by association for many of its close partners and investors. Most notably, Oracle’s aggressive capital expenditure plans to support demand from OpenAI have sparked a sell-off in its stock while widening its credit default swap spreads.

A senior OpenAI executive told the FT: “That’s been kind of the strategy. How does [OpenAI] leverage other people’s balance sheets?”

OpenAI partners amass $100bn debt pile to fund its ambitions

Open AI, softbank, Oracle, Coreweave

Partners Oracle, SoftBank, CoreWeave, Crusoe, and Blue Owl Capital are all taking on debt in the form of bonds, loans, and credit deals to meet their obligations with OpenAI for infrastructure and computing resources.

Having close ties with OpenAI has been an anchor for many publicly traded companies in recent weeks. The company’s cash burn and the rise of Gemini 3 have seemingly darkened its outlook and fostered guilt by association for many of its close partners and investors. Most notably, Oracle’s aggressive capital expenditure plans to support demand from OpenAI have sparked a sell-off in its stock while widening its credit default swap spreads.

A senior OpenAI executive told the FT: “That’s been kind of the strategy. How does [OpenAI] leverage other people’s balance sheets?”

tech
David Crowther

Chinese tech giants are training their models offshore to sidestep US curbs on Nvidia’s chips

Nvidia can’t sell its best AI chips in the world’s second-largest economy. That’s an Nvidia problem. But it’s also a China problem — and it’s one that the region’s tech giants have resorted to solving by training their AI models overseas, according to a new report from the Financial Times.

Citing two people with direct knowledge of the matter, the FT reported that “Alibaba and ByteDance are among the tech groups training their latest large language models in data centers across south-east Asia.” Clusters of data centers have particularly boomed in Singapore and Malaysia, with many of the sites kitted out with Nvidia’s latest architecture.

One exception, per the FT, is DeepSeek, which continues to be trained domestically, having reportedly built up a stockpile of Nvidia chips before the US export ban came into effect.

Last week, Nvidia spiked on the news that the Trump administration was reportedly considering letting the tech giant sell its best Hopper chips — the generation of chips that preceded Blackwell — to China.

China’s tech giants take AI model training offshore to tap Nvidia chips

China’s tech giants take AI model training offshore to tap Nvidia chips

Citing two people with direct knowledge of the matter, the FT reported that “Alibaba and ByteDance are among the tech groups training their latest large language models in data centers across south-east Asia.” Clusters of data centers have particularly boomed in Singapore and Malaysia, with many of the sites kitted out with Nvidia’s latest architecture.

One exception, per the FT, is DeepSeek, which continues to be trained domestically, having reportedly built up a stockpile of Nvidia chips before the US export ban came into effect.

Last week, Nvidia spiked on the news that the Trump administration was reportedly considering letting the tech giant sell its best Hopper chips — the generation of chips that preceded Blackwell — to China.

tech
Millie Giles

Alibaba unveils its first AI glasses, taking on Meta directly in the wearables race

Retail and tech giant Alibaba launched its first consumer-ready, AI-powered smart glasses on Thursday, marking its entrance into the growing wearables market.

Announced back in July, the Quark AI glasses just went on sale in the Chinese retailer’s home market, with two versions currently available: the S1, starting at 3,799 Chinese yuan (~$536), and the G1, at 1,899 yuan (~$268) — a considerably lower price than Meta’s $799 Ray-Ban Display glasses, released in September.

Read more

Alibaba's AI glasses to rival Meta go on sale for $500

Alibaba's AI glasses to rival Meta go on sale for $500
tech
Jon Keegan

Musk: Tesla’s Austin Robotaxi fleet to “roughly double” next month, but falls well short of earlier goals

Yesterday, Elon Musk jumped onto a frustrated user’s post on X, who was complaining that they were unable to book a Robotaxi ride in Austin. Musk aimed to reassure the would-be customer that the company was expanding service in the city:

“The Tesla Robotaxi fleet in Austin should roughly double next month,” Musk wrote.

While that sounds impressive, there are reports that Austin has only 29 vehicles in service.

But last month, Musk said the Robotaxi goal was to have “probably 500 or more in the greater Austin area” by the end of the year.

Meanwhile, Google’s Waymo has more than 100 autonomous taxis running in Austin, and 1,000 more in the San Francisco Bay Area.

Elon Musk slashes Tesla Robotaxi fleet goal from 500 to ~60 in Austin

Elon Musk slashes Tesla Robotaxi fleet goal from 500 to ~60 in Austin

“The Tesla Robotaxi fleet in Austin should roughly double next month,” Musk wrote.

While that sounds impressive, there are reports that Austin has only 29 vehicles in service.

But last month, Musk said the Robotaxi goal was to have “probably 500 or more in the greater Austin area” by the end of the year.

Meanwhile, Google’s Waymo has more than 100 autonomous taxis running in Austin, and 1,000 more in the San Francisco Bay Area.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.
©2025 Sherwood Media, LLC