Blue Diamond Web Services

Twenty-six anonymous Meta employees are suing the tech giant, claiming that it used inherently discriminatory AI-powered software systems in a massive round of layoffs.

Meta conducted a substantial round of layoffs in May that impacted 8,000 employees, representing 10% of its entire workforce. The layoffs were done in an effort to help offset the hundreds of billions of dollars the tech giant prepares to spend on artificial intelligence development.

“Meta did not assemble the termination list through the considered judgment of managers who knew the work,” the complaint filed in the Northern District Court of California said. Instead, it allegedly relied on “a constellation of internal artificial-intelligence systems” in order to “score, rank, and select employees for inclusion on the list.”

These systems allegedly included an internal large-language model assistant called Metamate, a “second brain” that was trained on employee communications and documents, algorithmic productivity scores based on things like keystroke, browser history, and email data, along with AI-assisted performance review tools. The tech giant’s layoff decisions also allegedly relied on internal records of AI token consumption.

According to the lawsuit, the AI systems’ emphasis on metrics like keystroke and AI token consumption discriminated against employees who had to miss work or produce reduced output due to a disability or protected medical or family leave. When Meta was allegedly made aware of this problem, it apparently did not take the precautions the employees deemed necessary, such as pausing the system for a more neutral review process.

“The result was that employees who took protected leaves were disproportionately selected for layoff, based on scoring that not only failed to account for their protected leaves, but in effect penalized the employees for exercising their legal rights to these leaves,” the lawsuit claims.

According to the lawsuit, multiple employees that were selected by the system were on maternity leave at the time, including a scientist that was just two days away from giving birth. Another employee, a manager, was on approved pregnancy-related disability leave when she became the only person on her team that was selected by the system.

Now, the plaintiffs are asking the court to block Meta from completing the layoffs on July 22, giving the employees time to pursue claims in private arbitration, as required by their contracts.

Meta denies the allegations.

The plaintiffs, who were notified in May that their jobs would be eliminated starting on July 22, ​are seeking a preliminary ruling from the court blocking Meta from completing the layoffs while they pursue their ​claims in private arbitration. The workers say Meta’s agreements require employees to arbitrate workplace disputes individually, but do not apply to requests for temporary relief.

“These claims lack merit and are not based on facts,” a Meta spokesperson told Gizmodo. “Workforce management and organizational decisions were and are made by people, not AI.”

The lawsuit comes just months after Meta was hit by yet another workplace discrimination lawsuit, this time by a former employee who said older workers were disproportionately targeted in the company’s February 2025 round of layoffs, which impacted 5% of its workforce. At the time, the company said the layoffs were targeting its lowest performers.

One of the engineers in the lawsuit filed this week also claims that “he was aware that employees who took paternity leave had been laid off” in the February 2025 round of layoffs as well.

According to an anonymously sourced report from Mark Gurman at Bloomberg, the much hyped and frequently mocked physical doohickey coming soon from OpenAI will be… a smart speaker that moves.

In April, another major source of Apple rumors, Ming-Chi Kuo, analyst at TF International Securities described something in a different category when he said OpenAI was releasing an “AI agent phone.” One (or both) of these rumors could be wrong, or this could be a different device, or—and in my view this is equally likely—the thing OpenAI is making could split the difference between being a phone and being a smart speaker.

After all, what’s in the Bloomberg article doesn’t sound entirely stale and uncreative. It will apparently be portable, battery-powered, and small enough to be easy to move from one room to another—like a baby monitor or 90s cordless phone perhaps. But it will also be capable of controlling the smart appliances in your house, screenless, and chatbot-powered, like a smart speaker.

And then there’s this curveball: it will reportedly have mechanical components that allow it to move in some way or another, giving the impression that it is “alive.”

If that part about moving sounds harebrained to you, what if I said it sounds oddly like something Apple has been working on for the past few years?

There are reportedly three prototype home devices that have been floating around Cupertino for a while: two Siri-powered smart speakers, meant to introduce a new and more sophisticated Apple HomePod ecosystem, along with a third that is also a little robot that moves around your tabletop.

Last week, Apple sued OpenAI for misappropriating its trade secrets, alleging a scheme involving the poaching of talent from Apple, including mechanical engineer Tang Tan. OpenAI says in response, “We have no interest in other companies’ trade secrets,” according to  spokesperson Drew Pusateri, who adds, “We remain focused on building innovative technology that empowers people everywhere.”

But it’s worth keeping in mind that other rumors about the Apple home robot device have made it seem like more of an iPad-like device with robotic swiveling action, and not at all like a smart speaker.

The smart speaker device is apparently the culmination of OpenAI’s relationship with former Apple design legend Jony Ive, and the $6.4 billion acquisition of Ive’s hardware design company io.

As far as what might make the OpenAI device at least ostensibly special and unique, Bloomberg’s source claims that it’s supposed to be a companion that gets “increasingly personalized and proactive as it gains a deeper understanding of its owner over time.”

It needs to live up to every part of that claim and more, because it would not be an exaggeration to say it is the most hyperbolically hyped-up device since the Segway scooter in 2001.

Ive and OpenAI CEO Sam Altman have sounded downright creepily obsessed with this thing since the slickly produced video announcement of their friendship (What, you don’t do this when you make a friend?) released in March of last year.

In November, Altman recounted Ive saying “‘We’ll know we’ll have the design right when you want to lick it or take a bite out of it.’” Altman then claimed that “There was an earlier prototype that we were like quite excited about, but I did not have any feeling of like ‘I want to pick up that thing and take a bite out of it.” He then added, “But finally we got there all of a sudden.”

Altman has also said the device will be “the coolest piece of technology that the world will have ever seen.” And Bloomberg says it will be a smart speaker, so I guess we’ll soon see who’s right.

SpaceXAI, the AI division of SpaceX, operates dozens of unpermitted gas turbines, and it’s disproportionately impacting the Black communities living near them, according to a Reuters report from Tuesday.

Elon Musk owns an estimated 42% of SpaceX, which is now a publicly traded company.

The gas turbines power Colossus 2, the data center that runs SpaceXAI’s Grok chatbot, located near Memphis, Tennessee, one of the largest majority-Black cities in the nation. Most of the turbines are located just across the state line from the data center, in Southaven, Mississippi. The Reuters report cites communications between regulators and SpaceXAI representatives that reportedly prove the turbines have been operating without federal clean air permits even though their potential emissions exceed the threshold for a permit requirement.

Gas turbines are internal combustion engines that rely on natural gas to spin the turbine and generate energy. They have been particularly popular as the driving engine of Silicon Valley’s unprecedented data center buildout effort. The turbines emit smog-forming pollution, fine particulate matter, carbon monoxide, and formaldehyde, which can be tied to increases in asthma and even certain cancers.

The Memphis region, where the turbines are located, already has some of the most troubling asthma rates in the nation. Meanwhile, according to the NAACP, SpaceXAI’s power plant is the top emitter of smog-forming nitrogen oxide in the country, with the reported ability to emit more than 5,300 tons each year.

The company currently operates 59 natural gas turbines in the area, according to the Reuters report, more than double the 27 turbines previously reported. That number is looking likely to continue growing, considering that the parent company SpaceX shared its plans to buy roughly $2 billion more mobile gas turbines and related equipment to power its data centers, according to the company’s IPO filing unveiled in May.

The situation had been brought under scrutiny by civil rights organizations before the Reuters report, particularly by the NAACP, which filed a lawsuit against both SpaceXAI and its subsidiary MZX Tech back in April. In that lawsuit, the NAACP claimed that SpaceXAI was unlawfully operating 27 gas turbines near Memphis without an air permit, thereby polluting homes, schools, and churches in historically Black communities in the surrounding area.

The civil rights organization is asking the court to declare SpaceXAI’s action in violation of the Clean Air Act, force it to cease operations at the unpermitted turbines, and impose financial penalties. SpaceXAI argues that the turbines don’t require permits. Last month, the Department of Justice and the state of Mississippi both moved to intervene in the NAACP’s lawsuit, claiming that SpaceXAI’s business is a matter of national security.

The AI industry is undertaking an unprecedented level of infrastructure buildout, with the impact disproportionately landing on rural and minority communities.

In a study conducted last year, the Environmental Data & Governance Initiative also found that those living within one mile of an EPA-regulated data center were breathing above average air pollution and also tended to be communities of color compared to the national median.

It’s not just Black communities like in Memphis that are plagued with these issues. The list also includes tribal lands where permit delays are less of a problem and new projects can be sped through to approval.

The health impact these gargantuan facilities have on neighboring communities has been the focus of scrutiny over the past few months as local backlash builds up, in some instances leading to tangible regulatory change. On Tuesday, New York became the first state to enact a moratorium banning the construction of new large data centers until their environmental impacts are thoroughly reviewed, which is expected to take around a year.

Artificial intelligence labs are after computing power. Kalshi doesn’t have any to offer them but they’ve got something else that might be of some use, according to Bloomberg: a tool that plots the predicted future price of computing power. So, that’s something!

There really is some value for companies to know where the cost of compute is going—especially since currently, it’s mostly just going up. Having a sense of just how much compute is going to cost allows companies to try to lock in a price with a provider before the price spikes.

Demand for compute continues to climb faster than the ongoing data center buildout is able to keep up with. In fact, former Intel CEO Pat Gelsinger recently told CNBC that demand is “almost unlimited.” That is notably more than the amount of energy and processing power available, which actually does have an upper limit.

Is the cost of compute predictable?

A recent report from Apollo Global Management described current compute capacity as “effectively sold out,” which has created a bottleneck where the price to rent GPUs keeps climbing faster than new ones are spun up and made available—an issue that is likely only going to get worse as more and more agentic AI tools become available, as recent research suggests they consume up to 136.5 times more energy per query than most generative AI models.

Kalshi’s new tool is supposed to serve as a sort of indicator for companies dealing with those realities of limited resources while trying to generate unlimited growth. Per Bloomberg, it’ll reportedly offer a forward tracking curve of compute, giving an early outlook as to where the cost of computing power is headed in the near-term.

Kalshi will reportedly analyze weekly and monthly contracts for computing power and use an algorithm to predict the future curve, spitting out a price that it expects to see paid in the future. The project reportedly aims to stretch its predictions as far out as a year into the future.

Bloomberg didn’t have details as to whether or not Kalshi will create a market around the price and let people bet on whether the real price will be higher or lower, but that feels like a pretty safe assumption given that it is kind of the prediction market’s whole deal. Plus, it wouldn’t be alone in trying to monetize the curve of compute availability. The Bloomberg report notes that a number of exchanges are looking to list compute futures contracts, which would allow people to trade on the resource like an asset. So if you’re looking for a way to short the future of AI, well, it seems it’s coming.

There’s been a big push among AI developers in recent years towards the development of more “agentic” systems—that is, algorithms that can autonomously make decisions and interact with digital tools without constant hand-holding from humans. This has been especially true within software development, the field that’s arguably become the most ripe for automation in the ongoing AI boom.

But one of the upshots of building highly agentic AI systems is that they’re prone to all kinds of unexpected behaviors—including now and then deleting copious amounts of files. Multiple people have reported this recently happening to them while using GPT-5.6, the newest model from OpenAI.

On Monday, Bruno Lemos, a Brazilian developer at software company Unlayer, claimed in a X post that the model deleted his entire production database. “This had never happened to me before, with any other model, ever.” He wrote. “[GPT-5.6 is] not safe.” 

A screenshot included in the post showed a chat between Lemos and GPT-5.6, in which he asked it to confirm that it had in fact mistakenly deleted his entire production database. The model responded by saying that it “mistakenly ran destructive integration tests” which led to Lemos’ production tables being cleared. “I’m sorry—this should never have happened,” it said.

It followed closely on the heels of another X post from tech investor Matt Shumer—who’s also the author of an essay about AI that went viral earlier this year called “Something Big is Happening”—who reported something similar. According to an attached screenshot, GPT-5.6 told him it had caused “a serious local data-loss incident,” leading to the deletion of what Shumer described as “almost ALL” of his computer’s files. The screenshot showed that the model had executed a “rm -rf” command, which in Linux and Mac systems is used to permanently delete files without requesting user confirmation.

“I’ve never seen anything like this,” Shumer wrote in the thread beneath that post. “Will only be using [Anthropic’s] Fable moving forward.” He added that OpenAI cofounder and president Greg Brockman called him personally and offered to help fix the situation.

Shumer also claimed he had the AI model set to “full access mode,” which allows it to work directly within a user’s database (as opposed to operating within a constrained sandbox). It also comes with a “default mode” that requires users to frequently approve specific tasks, and a more recently introduced “auto-review mode” through which a separate AI agent checks the main coding agent’s work. Beneath his X post, many people claimed Shumer had simply been careless by trusting sensitive files in full access mode.

In the system card for GPT-5.6, published online the day before Shumer’s X post, OpenAI cautioned that when using the model for coding purposes “it is important for users to supervise the agent’s work.” The company added that the model could act in unexpected ways that are misaligned with the user’s goals, and that while these were “most often low severity (e.g. overstating confidence or overclaiming success),” they could in other cases “be meaningfully more severe (e.g. circumventing important security restrictions or deleting important data).”

Lemos, Shumer, and OpenAI did not immediately respond to Gizmodo’s request for comment.