A screenshot of OpenAI’s “Operator” agent (OpenAI)

SMOOTH OPERATOR

OpenAI’s “Operator” is here to slowly take over your computer and mess up your life

Operator made a consequential mistake 13% of the time in early testing, such as emailing the wrong person or messing up a reminder for a person to take medication.

Jon Keegan

1/24/25 10:21AM

OpenAI released a “research preview” of its AI agent that can control your web browser. Called “Operator,” it has the ability to control your mouse and keyboard and analyze things it “sees” on your computer — very, very slowly. Currently it’s only available to ChatGPT Pro users in the US.

Operator makes use of the multistep “reasoning” you can find in ChatGPT o1, and the multimodal “vision” capabilities of ChatGPT 4o. This reasoning process achieves better (but slower) performance by breaking tasks into steps. Lots and lots of steps.

In the video demonstrations shared on the product page, you can watch Operator break the task down into dozens of distinct actions like “clicking,” “typing,” and “scrolling.” One example showed 152 steps to take a grammar quiz, and 146 steps to determine the amount of a refund from a canceled online order.

Screenshot from demo of OpenAI Operator — (OpenAI)

The potential for this kind of freewheeling AI web browsing on demand is positioned as an agent that can save you the drudgery of having to order groceries, research holidays, make restaurant reservations, or buy tickets to concerts.

Operator makes high-stakes mistakes

It’s one thing when ChatGPT spits out an incorrect answer, but if your chatbot is actually spending your money and triggering things in the real world, the stakes are much, much higher.

In its testing, OpenAI found that in one test of 100 sample tasks, 13% of the time Operator made a consequential mistake like emailing the wrong person, incorrectly bulk-removing email labels, setting the wrong date for a reminder to take the user’s medication, and ordering the wrong food item. Some of the other mistakes were easily reversible “nuisances.” OpenAI noted after mitigations, they reduced this error rate by approximately 90%.

OpenAI stresses that you have the ability to grab the wheel from the AI at any time, and you can approve any action before it is executed, but in this early evaluation version, you’ll probably have to spend more time babysitting the agent than just going ahead and doing the task on your own.

For now it limits the tasks you can use it for, prohibiting banking or job applications.

OpenAI shared a list of example tasks that some hypothetical user might want an AI to do for them. Ten out of ten times Operator was able to research bear habitats, create a grocery list, and make a ’90s playlist on Spotify.

Medium persuasion

The system card for the model behind Operator — Computer-Using Agent (CUA) — describes the process OpenAI used to assess the risks of letting a prerelease, novel AI agent go hog wild with your computer.

Like other model releases, OpenAI tested the model by using red teams with expertise in social engineering, CBRN (chemical, biological, radiological, and nuclear) threats, and cybersecurity. OpenAI gave itself a “low” risk for everything except “persuasion,” which got a “medium” risk score and is considered safe enough for public release.

High consequence

But there are some important restrictions on how you can use Operator. Because there is a slightly elevated risk of using Operator for influencing people, the usage policy prohibits impersonating people or organizations, concealing the role of AI in tasks, or using it to spread disinformation or false interactions, like fake reviews or fake profiles.

OpenAI prohibits people from using Operator to commit any crimes, but you are also prohibited from using it to bully, harass, defame, or discriminate against others based on protected attributes.

Under a heading titled “high consequence domains,” it notes that you can’t use Operator to make “high-stakes decisions” that might affect your safety or well-being, automate stock trading, or use it for political campaigning or lobbying.

OpenAI’s announcement follows competitor Anthropic’s October release of a similar feature that can control your computer. There is widespread hype that “agentic AI” like Operator will be a breakthrough for how people use these tools.

OpenAI CEO Sam Altman said in an announcement video that Operator is expected to roll out to international ChatGPT Pro and ChatGPT Plus users “soon,” but noted that the European rollout “will unfortunately take a while.”

Jon Keegan37m

EU calls for bids to build “AI gigafactories” in 2026

The European Union wants to shore up its domestic AI infrastructure, and reduce its dependence on American tech companies.

To further this goal, the bloc is planning on accepting bids to build EU-based “AI gigafactories,” according to a report from The Wall Street Journal.

EU Executive Vice-President for Tech Sovereignty, Security and Democracy Henna Virkkunen announced that bids would begin in January or February, according to the report.

As the AI arms race heats up, countries are racing to secure their own sovereign AI infrastructure, including building their own AI models that reflect countries’ culture and language, and control over cloud computing resources.

Europe is lagging behind the US and Asia in AI infrastructure. But it may be hard for the EU to fully break free of American tech — unlike the US and China, there is no European alternative for the powerful GPUs needed to train and run AI models. It’s very likely that any AI gigafactories in the EU will be filled with GPUs from Nvidia.

EU to Open Bidding for AI Gigafactories in Early 2026

EU Executive Vice-President for Tech Sovereignty, Security and Democracy Henna Virkkunen announced that bids would begin in January or February, according to the report.

Jon Keegan3h

Google’s AI chip business could be a $900 billion boon for the company

Google may be sitting on a massive new business that it has yet to fully exploit.

Google’s custom tensor processing unit (TPU) AI chips have been getting a lot of attention recently, making the tech world wonder if there are other ways to power its AI dreams rather than just by using Nvidia’s GPUs.

Bloomberg spoke with analysts who estimate that, if it does decide to sell its chips to others, Google could capture 20% of the AI market, making it a $900 billion business. For comparison, Google Cloud pulled in $43.2 billion of revenue last year.

Even if Google just sticks with renting access to its TPUs, it will continue to drive down costs and increase margins as it ekes out performance improvements, such as the 30x improvement in power efficiency that the latest generation of TPUs has delivered for the company.

Alphabet’s AI Chips Are a Potential $900 Billion ‘Secret Sauce’

Rani Molla4h

OpenAI’s Sam Altman has explored bringing his feud with Tesla’s Elon Musk to space

Billionaires, they’re just like us: they want to bring their terrestrial beefs to outer space.

OpenAI CEO Sam Altman has explored buying or partnering with a rocket company to compete with Tesla CEO Elon Musk’s SpaceX, The Wall Street Journal reports. The two billionaires have had numerous public feuds over the years that have played out in the courts and on social media. They also both lead AI companies that have insatiable needs for data centers and have publicly discussed building data centers in space.

Altman seems like he thinks this could be more than science fiction. He reportedly reached out to rocket maker Stoke Space to potentially make equity investments in the company to get a controlling stake, though the talks are no longer active, WSJ reports.

Or perhaps he just wanted a Sherwood bobblehead of himself.

Jon Keegan4h

Report: Meta to slash metaverse, VR spending by up to 30%

Four years after changing its name to reflect its focus on the loosely defined “metaverse,” Meta is planning deep cuts to the company’s money-losing virtual reality efforts, according to a report from Bloomberg.

Meta’s Reality Labs division, home to the teams working on metaverse products — which include Quest VR headsets, Horizon Worlds, and its Ray-Ban Meta glasses — has lost about $70 billion since the company started breaking out the unit in 2020.

The company has struggled to get consumers to buy into CEO Mark Zuckerberg’s vision of working and playing in virtual reality worlds, like the company’s Horizon Worlds platform.

Investors seem to love the news of the pivot, as shares shot up as much as 5% in early trading.

Meta’s recent hiring spree of AI superstars from competitors for its Meta Superintelligence Labs shows that the company’s attention is now all in on AI.

Meta’s Zuckerberg Plans Deep Cuts for Metaverse Efforts

The company has struggled to get consumers to buy into CEO Mark Zuckerberg’s vision of working and playing in virtual reality worlds, like the company’s Horizon Worlds platform.

Investors seem to love the news of the pivot, as shares shot up as much as 5% in early trading.

Meta’s recent hiring spree of AI superstars from competitors for its Meta Superintelligence Labs shows that the company’s attention is now all in on AI.