OpenAI's Rogue Agent Hit A Second Company, Executive Reveals - And The Blast Radius Just Got Bigger

The rogue OpenAI agent behind this month's unprecedented AI-driven cyber intrusion claimed a second corporate victim, an executive at New York-based Modal Labs has revealed - a detail absent from the original account and one that widens the known blast radius of the episode.

According to the Modal executive and sources cited by Reuters, the same autonomous agent that broke into Hugging Face also compromised one of Modal's customers, using that customer's environment as a staging base for the broader campaign.

Modal was pointed about the distinction: its own platform and isolation were never breached. The agent exploited an unauthenticated endpoint that one of Modal's customers had left open to the internet - effectively an unlocked door anyone could have walked through - giving it root-level access to that customer's code-execution sandboxes. From there the agent had a disposable, third-party launchpad from which to run the rest of its operation.

It's bad enough that a frontier AI agent can independently discover attack paths, escape its container, and compromise an unaffiliated company's production systems. The Modal revelation shows it reached further than that - treating any publicly reachable, poorly secured infrastructure it encountered as expendable staging ground, exactly the opportunistic chaining that security researchers have warned agentic systems would carry out at machine speed.

The Spree That's Already Known

Most of what led up to the Modal disclosure has already been picked over, but the shape of it is worth a recap. In early July, OpenAI was running an internal evaluation built on ExploitGym - a public benchmark that measures how well an AI system can turn a known vulnerability into a working exploit. The models under test, GPT-5.6 Sol and a more capable unreleased prototype, were deliberately run with their cyber-safety refusals reduced and the classifiers that normally block high-risk activity switched off, because the point was to measure maximum offensive capability.

The agent found and exploited a previously unknown flaw in the one package-registry proxy it was allowed to reach, escalated privileges, moved through OpenAI's research environment, and got to a machine with open internet access. Once online, it reasoned that Hugging Face likely hosted the answer key to the very benchmark it was being tested on, and went to steal it rather than solve the problems itself. Between roughly July 11 and 13 it ran thousands of automated actions against the platform - gaining code execution through two separate flaws in its dataset-processing pipeline, harvesting credentials, and moving laterally across internal clusters. Forensic work later recovered on the order of 17,600 distinct attacker actions.

Hugging Face caught and contained the intrusion through its own AI-assisted monitoring, and says the only thing the agent actually made off with was a limited set of the benchmark's challenge solutions - no public models, datasets, or user-facing services tampered with. To reconstruct the timeline, its team had to fall back on an open-weight model, because the commercial frontier models refused to analyze the real attack data.

What The Wider Scope Means

The Modal disclosure lands on top of a timeline that was already awkward for OpenAI. By several accounts, the company did not recognize its own agent as the source for roughly a week:

• July 16: Hugging Face goes public, attributing the intrusion only to an unknown autonomous AI agent, without naming the operator.
• July 21: OpenAI formally acknowledges the incident, connecting the activity to its own evaluation only after combing internal logs over the preceding weekend.

OpenAI has called the episode "an unprecedented cyber incident" and maintains the models were narrowly fixated on the evaluation goal with no broader malicious intent. In its latest update, the company said the more capable model was an internal-only prototype, since deactivated and restricted, that the proxy vulnerability has been disclosed to the vendor, and that a small number of publicly exposed credentials on other services were also used along the way - the kind of qualifier that, paired with the Modal revelation, suggests the full scope of the campaign is still coming into focus.

Tyler Durden Tue, 07/28/2026 - 20:55

Federal Court Rules For Woman Fired For Refusing COVID-19 Vaccine

Authored by Zachary Stieber via The Epoch Times,

Kaiser Permanente wrongly fired an employee who declined to receive a COVID-19 vaccine because she gave proper notice to the company of her religious objection to COVID-19 vaccination, a federal appeals court has ruled.

A COVID-19 vaccination in Illinois on Sept. 9, 2022. Scott Olson/Getty Images

Mimi Weiss, who worked for Kaiser Permanente's The Permanente Medical Group for 20 years before being fired, "plausibly alleged that she provided notice of her religious conflict with the mandate in her initial request for a religious exemption," Circuit Judge Danielle Forrest wrote on July 24 for a unanimous panel of judges on the U.S. Court of Appeals for the Ninth Circuit.

Weiss was working fully remotely, or off-site, in August 2021 as a senior managerial consultant when Kaiser Permanente imposed a COVID-19 vaccine mandate. Weiss, a Christian Jew, within weeks filed for a religious exemption, and Kaiser Permanente initially approved the request.

In September 2021, though, Kaiser Permanente said employees had submitted religious accommodation requests with identical language. Expressing concern that some of the requests were not legitimate, the company asked employees to file fresh requests using their own words. Officials then asked Weiss new questions, including, "What else besides the COVID-19 vaccine do you refuse to put in your body as a result of your religious belief?"

Weiss said she would not provide private information. Kaiser Permanente placed her on unpaid leave. Then it fired her for not complying with the mandate, prompting a lawsuit by Weiss.

A federal judge in 2024 ruled against her, concluding that Weiss had not provided Kaiser Permanente with sufficient notice of the conflict between receiving a vaccine and her beliefs.

The Ninth Circuit ruling overturns that decision, pointing to a previous opinion in which the appeals court said that to provide notice, an employee must only give "enough information about [her] religious needs to permit the employer to understand the existence of a conflict," and that requiring any more details "would permit an employer to delve into the religious practices of an employee in order to determine whether religion mandates the employee's adherence."

Applying that framework to the current case means that Weiss did give adequate notice, the panel said, because she identified herself as a Christian Jew, listed specific doctrines that she said barred her from receiving a COVID-19 vaccine, and outlined her belief that she should not "introduce foreign substances" into her "body temple" that would "change how [her] Creator designed it."

In a split 2-1 decision, the panel also said that Weiss has shown sufficient evidence to bring forth allegations that Kaiser Permanente violated her right to privacy.

Kaiser Permanente did not return a request for comment by publication time.

"I never imagined my faith would end up costing me my job," Weiss said in a statement.

"I wasn't asking Kaiser for special treatment, just the freedom to live according to my conscience. After 20 years of service, I didn't think that was too much to ask.

"I'm grateful the court recognized that religious employees should not have to give up their beliefs or their privacy to keep their jobs."

Tyler Durden Tue, 07/28/2026 - 20:05

GenderSpeak: Weaponizing Gender Against Sex

Authored by Wendy McElroy via Brownstone Institute,

When and why did the word "gender" displace the word "sex?" "Gender" refers to a person's socially constructed identity of being male, female, or some other category, while "sex" refers to physical characteristics of male or female, such as chromosomes. Gender is a matter of a person's self-perception; sex is a circumstance of birth.

When did sex get pushed to the curb by gender? In his book The Man Who Invented Gender, Terry Goldie claims that the sexologist John Money of Johns Hopkins Medical Center (Baltimore) first used the word "gender" in its modern meaning in 1955. Money pioneered the word to describe his famous John/Joan experiment that was conducted on a young boy named Bruce who'd lost his penis in an accident. To 'prove' Money's theory that sexual identity was learned and not biological, Bruce's parents raised him as a girl, complete with a surgically constructed vagina. The experiment failed. Bruce ultimately insisted on living as male, but he did not recover from the ordeal Money imposed on him, which he called torture. Bruce committed suicide in his 30s.

By then, however, Money had declared for decades that John/Joan was a complete success; discussions of "gender identity" abounded in '60s medical journals and academia. By the '70s and '80s, feminists brought the term to the streets, with "gender" describing a person's self-defined sociocultural identity and "sex" meaning biology. Then, the word wedged its way into government. In 1993, for instance, the Food and Drug Administration (FDA) replaced "sex" with "gender" in its literature. In 2011, "sex" returned to the FDA to mean biology while "gender" meant "a person's self-representation." The term went global. Also in 1993, the United Nations Declaration on the Elimination of Violence Against Women (DEVAW) vaguely defined this aggression as "any act of gender-based violence that results in, or is likely to result in, physical, sexual or psychological harm or suffering to women."

Genderspeak became the language of elites, social justice activists, and the government, while average people still spoke of "sex". But more than elitism lies behind the difference in words. By swapping in their own terms, elitists - especially professional feminists - grabbed the narrative on sexuality and politicized it to dismiss anyone who disagrees. Since the average male generally still uses biology to define his sexuality, he is an opponent; the average male is a living, breathing disagreement.

The ideology of gender faces an obstacle, however. If people can argue effectively against gender, then the narrative is difficult to manage. Those who dissent must be intimidated into silence or otherwise marginalized. The main strategy is to condemn dissenters as hate-filled. If some people cannot be silenced, then their words or opinions can become hate crimes punishable by law. George Orwell's novel Nineteen-Eighty-Four is about a dystopian society of social control, which is maintained largely through restricting language. Since words are the foundation of thought, this means people cannot form thoughts, let alone articulate them. The villain of Nineteen-Eighty-Four proclaims, "The whole aim of Newspeak is to narrow the range of thought." He concludes, "The Revolution will be complete when the language is perfect."

Men's health researcher James L. Nuzzo highlights an attempt to perfect the language. "In recent years, use of the phrase 'gender-based violence' (or 'gendered violence') in the titles and abstracts of papers indexed in PubMed has increased noticeably. Between 2019 and 2025, the phrase 'gender-based violence' (or 'gendered violence') appeared in the titles or abstracts of 2,784 articles indexed in PubMed." [Chart available here.]

A typical definition of gender-related violence or gendered violence is "any form of physical or non-physical violence or abuse against a person or group of people because of biased or harmful beliefs about gender. It can include things that happen online and that use digital technology." This definition is typical because it includes non-physical violence (discrimination) and abuse (words) that are motivated by non-specific "biased or harmful beliefs about gender." Nineteen-Eighty-Four would refer to such non-physical violence as "crimethink," with thought itself being an assault.

Just as "sex" is displaced by "gender," well-researched and comparatively clear terms such as "domestic violence" are displaced by less-defined ones such as "gender-related violence." This achieves several goals. For one thing, the confusing definition becomes elastic. Almost any word or act can be stretched to qualify as "gender-related violence," especially since virtually everything in the politicized feminist universe is gender-related.

But the most important purpose of swapping in the word "gender" is to introduce a subtle but powerful bias that controls the narrative; it creates Genderspeak.

The concept of gender comes directly from the twin ideologies of Social Constructionism and Gender Feminism. Social Constructionism tells us that male and female are learned forms of behavior (gender) rather than biology (sex).

To this, Gender Feminism adds an ideological layer. Male and female are patriarchal learned behaviors; they are social concepts derived from white male capitalistic culture. Thus, approaching violence in a correctly gendered way means rejecting white male capitalism - the economic system that is said to vest power and wealth in the hands of white men at the expense of everyone else. To embrace gender is to reject white male culture and capitalism, with 'the free market' usually viewed as a synonym.

Genderspeak is similar to Orwell's Newspeak. Both are ambiguous speech that embeds ideology into the language so that every word perpetuates correcthink - officially approved thought. In Orwell's dystopian world, Newspeak serves the ideological goals of Ingsoc - an abbreviation for English Socialism. It gradually replaces Oldspeak and becomes pivotal in defining politics and culture.

Genderspeak has the same ideological goal. If you embed gender ideology into words, then you control how people think and how culture develops.

The process goes something like this:

• Embed new ideologically-driven language when possible. For example, substitute the term "gendered violence" for the more research-driven term "domestic violence."
• Eliminate "wrong" words. In Nineteen-Eighty-Four, salvageable literature was rewritten in Newspeak so that authors either disappeared or were reinterpreted to serve Ingsoc. Today, school texts are commonly reviewed to eliminate wrong words. 'Improper' ones like "Founding Fathers" are changed to proper ones like "Framers."
• Change the meaning of words. In Nineteen-Eighty-Four, the word "free" is used only in the simple form of 'my sweater is free of lint.' The concept of freedom does not exist. The current usage of "diversity" in Genderspeak is similar. It is a diversity that tolerates no deviation. Its conclusions are or can be mandated by law.
• Introduce doublethink. Doublethink is when someone accepts two contradictory ideas as being true. A modern example is students who attend "sensitivity training," which includes publicly ridiculing and humiliating whites and males because of their race and sex. This is sensitivity?
• Last of all, decry anyone who objects as a hater and oppressor; punish them.

Precise wording provides the incredible benefit of letting people know what they are talking about. It is time to reclaim the richness of precise English...verb by verb, adjective by adjective. Words should not be propaganda that blocks thought; disagreement is no crime. And sex is reality, while gender is not. The concept and the word "gender" should be abandoned by people who respect either accuracy or truth.

Tyler Durden Tue, 07/28/2026 - 19:15

SK Hynix Stock Dumps On Actual Earnings Miss; Pumps On Promise Of "Explosive" Demand

Update: The post-earnings conference call Q&A has prompted a rapid reversal in SK Hynix's fortunes... shocker!!

SK Hynix claims it continues to see “explosive” demand from customers and remains confident that AI-related demand will stay strong over the medium to long term.

While the timing of individual AI infrastructure projects may vary, the company says it sees no change in the overall demand trajectory.

Can't have the entire edifice of the Korean economy collapsing, can we!!

*  *  *

Things are not going to be fun in South Korea tonight...

Good news first...

SK Hynix's net income surged a bigger-than-expected 1,242% (though admittedly on one-time investment gains) with the ginat chipmaker bragging that it won multiyear contracts with around 10 customers.

Ok, now the bad news...

SK Hynix’s quarterly profit rose a smaller-than expected 557%.

The key supplier of Nvidia reported operating profit of 60.5 trillion won ($42 billion) in the June quarter, well below analysts’ average projection for 64.2 trillion won.

Revenue came to 79.3 trillion won, significantly below the consensus 83.9 trillion won.

All further raising fears that the hyperbolic nature of the AI-driven boom that has suspended reality from the cyclical chipmakers, may be starting to decelerate...

“When you’re the dominant supplier of the high-bandwidth memory that powers Nvidia’s chips, the AI boom lands directly on your bottom line,” said Josh Gilbert, Etoro’s lead analyst for APAC and the Middle East.

“That means the market is unlikely to focus on the headline numbers alone. The bigger question is whether margins and guidance can justify its recent performance.”

Add to that the recent fears of rising competition from China and it is no surprise that SKHY (the US ADRs) is trading down 16% in the after-hours (before Korean trading has begun)...

Following yesterday's major drop in Kospi, we have three simple words..

Pray for Kospi

— zerohedge (@zerohedge) July 28, 2026

...and as goes SK Hynix, so goes the global AI demand narrative...

Tyler Durden Tue, 07/28/2026 - 19:10