Category: Science & Tech

Hotel and resort giant Wyndham has acknowledged a cybersecurity incident after a good-faith security researcher discovered a bug that allowed anyone to access customers’ personal information.

Security researcher Robert Wiggins at What Security exclusively told TechCrunch that he found a password inside the source code of Wyndham’s Android app, which granted access to an internal company database, hosted by tech company Salesforce.

Wiggins told TechCrunch that the credential exposed personal information including customer names, email addresses, and Wyndham member number. The database also stores customer login tokens (called JWTs) that keep customers logged-in to their apps without having to repeatedly enter their own passwords. The database also contained Wyndham files and documents.

According to Wiggins, the database had millions of entries, but he was unable to determine an accurate number of exposed customer records. Wyndham touts some 129 million enrolled rewards members.

The researcher said Wyndham has no way to report security flaws, and asked for TechCrunch’s help in alerting the company. Soon after TechCrunch made contact with Wyndham’s cybersecurity staff, the exposed credential was revoked, preventing any further access to the database. 

“Our team has taken steps to address the issue,” a Wyndham spokesperson told TechCrunch. “Our security teams have no evidence of any unauthorized access,” the statement added, without providing any further details about the claim. 

Wyndham’s spokesperson would not say, when asked, whether the company had the technical means, such as server logs, to determine if anyone had access to the customer database prior to Wiggins’ discovery.

TechCrunch also asked Wyndham whether it planned to notify customers whose information was exposed, or disclose the incident to U.S. regulators, such as filing notice with the U.S. Securities and Exchange Commission, as public companies are required to do for material cybersecurity incidents. Wyndham would not say.

Wyndham’s spokesperson also would not say for how long the password was embedded in its customer app, or how Wyndham plans to ship a long-term fix for the app.

App and software developers generally avoid including passwords or credentials in publicly released software and apps — a process known as “hardcoding” — as anyone with a copy of the software can extract the password and use it to access its associated server or database. Given the simplicity with which credentials can be abused by anyone, many cases of hardcoded credentials are automatically scored with a maximum vulnerability severity rating.

For Wiggins, this is his second public discovery of hardcoded credentials in two weeks, after finding a near-identical vulnerability in Applebee’s app that exposed customer data. Wiggins relied on AI tools for both discoveries, which make it easier to identify simple security weaknesses, such as hardcoded credentials, by scouring thousands of lines of decompiled app code every second.

When asked about Wyndham’s response, Wiggins told TechCrunch that he was glad the company fixed the issue, but noted that without visibility into any past activity, someone with malicious intent could have leaked large amounts of customers’ personal information.

The company would not say whether it planned to set up a vulnerability disclosure program to allow outside security researchers and hackers to privately report security vulnerabilities. 

Michael Francess, Wyndham’s cybersecurity director who responded to the incident, told TechCrunch that researchers can alert [email protected] with details of potential security issues. The company’s spokesperson would not say say whether it plans to change its website to allow researchers to alert the company. 

This is the latest known security incident at Wyndham in recent years. In 2024, a security researcher found several Wyndham hotel check-in desks were running remote-access spyware that was leaking screenshots of guest records and booking information to the open web. 

Soon after our reporting, the spyware app found on the hotel computers, called pcTattletale, was hacked and its banks of customer data stolen, forcing the company’s owner, Bryan Fleming, to shut down the company. Fleming, a former resident of Michigan, was later arrested by federal authorities and convicted of making, selling, and advertising unlawful spyware.

Scientists have made a surprising discovery that could slash the amount of energy used in oil refining by eliminating part of the high-temperature heating process required to separate the fractions of crude oil.

The scientific breakthrough has found a simple and inexpensive way to separate crude oil molecules through a polymer membrane, radically different from previous attempts to fractionate crude oil via membranes.

In previous studies of membrane-based fractionations, scientists believed a selective layer should be coated onto the surface to help separate molecules.

But these additional-layer membranes are expensive to make and potentially defective when produced at scale.

So the researchers from the Korea Advanced Institute of Science and Technology (KAIST) and the Georgia Institute of Technology took a radically different approach. Instead of relying on an ultra-thin specialized coating, they passed crude oil directly through a bare porous polyacrylonitrile (PAN) membrane. This membrane is a chemically stable and inexpensive polymer commonly used as a support material in industrial membranes.

When the scientists had crude oil pass through the membrane, at room temperature, the membrane turned into a kind of clogged filter as heavy hydrocarbons selectively stuck to the pore walls. The accumulation of heavy hydrocarbons gradually narrowed the pores and created separation channels smaller than 2 nanometers, according to the study published in Nature this year.

The researchers showed that the bare PAN membrane delivered crude oil permeation rates approximately 23 times higher than those of previously reported state-of-the-art crude oil membranes while maintaining stable performance for 28 consecutive days.

“One of the key challenges facing membrane systems for crude oil separation was the low productivities of the membrane units – the PAN membranes with their surprising separation mechanism – dramatically increase the productivity of the membrane unit, to the point where industry should seriously consider adopting the technology,” said Professor Ryan Lively from Georgia Tech.

The new technology can be adopted by refineries with simple retrofits by adding a modular filtration unit that would not require massive investments, the researchers say.

These inexpensive membranes could materially lower the energy use and emissions from refining operations. The so-called “clogged filters” cannot replace distillation proper, but can be used at an early stage to separate fractions without the need to heat crude oil above 350 °C (662 F) from the very beginning, thus avoiding part of the energy use and emissions in the refining process.

Process simulations in the study showed that using the membrane as a pretreatment step before conventional distillation could reduce energy consumption by 31.6%, carbon dioxide emissions by 37.6%, cooling water usage by 20.7%, and operating costs by 36%.

The membrane could also be applied in other chemical separation processes in the petrochemicals and waste plastics recycling industries, according to the scientists. 

“Our goal is to precisely control this spontaneous pore-constriction phenomenon and develop it into a membrane platform applicable to the entire refining process,” said Dr. Jihoon Choi and Dr. Hyeokjun Seo of KAIST, the study’s co-first authors.

“We also aim to expand the technology to plastic recycling, biofuel purification, and other sustainable chemical processes that support carbon neutrality.”

This research is not the first time scientists have developed membranes that could reduce energy use in oil refining.

For example, engineers at MIT last year said they had developed a new approach to crude oil fractionation that could slash the amount of energy needed for processing the crude into different fuels by as much as 90% in a breakthrough that could revolutionize fuel production.

In a study published in the journal Science, the team at MIT revealed that it had developed a new filtration membrane that can efficiently separate heavy and light components from oil. The membrane is resistant to the swelling that tends to occur with other types of oil separation membranes.

MIT engineers modified polymers that have been used in water desalination by changing the bonds and introducing monomers.

“The key innovation is that the filters we developed can separate very small molecules at an atomistic length scale,” said Zachary P. Smith, an associate professor of chemical engineering at MIT and the senior author of the study.

By Tsvetana Paraskova for Oilprice.com

More Top Reads From Oilprice.com

Last month, hackers who were likely working on behalf of North Korea absconded with billions in stolen cryptocurrency after breaking into Binance, the world’s largest crypto exchange. Now, according to the crypto research firm Elliptic, Pyongyang is using meme coins to launder the hijacked funds.

The Binance heist is the biggest in crypto history, with hackers making off with $6.3 billion in crypto coins in just 20 minutes after months of complex planning.

As outlined in a blog post, Elliptic found that the isolated authoritarian nation is now minting meme coins — speculative digital assets often based on viral memes, celebrities, or niche internet jokes — and using them to quietly move stolen coins around liquidity pools. The coins are designed to be single-use and are fundamentally worthless, yet their creators can still set their price, making them what Elliptic calls a perfect “container” to pass pilfered crypto assets around. It’s also cost-effective, as meme coins are cheap to make.

“Most people think of meme tokens as joke coins whose prices rise and fall on hype,” Elliptic cofounder and chief scientist Tom Robinson wrote in the firm’s blog post. “North Korea’s launderers have found a different use for them, creating tokens that nobody else wants to buy and using them to disguise the movement of stolen funds.”

Elliptic found at least six meme coins believed to be connected to the scheme, each of which was created shortly after the $6.3 billion heist occurred. The hackers have reportedly been able to launder around $430,000 of their multibillion-dollar heist using this method.

To North Korea’s credit, the coins do look extremely dumb and worthless. One of them carries an image of sometimes-trillionaire Elon Musk under an “Elon 2028 President” banner, per Elliptic. (There’s no evidence that Musk is involved in the scheme.) Others feature a monkey with a pink hair bow and a weird duck in front of an American flag. One is simply called “FUK.”

“Meme coins are launched and traded in their thousands, and most fade within days, so one more short-lived token with a cartoon logo attracts little attention,” Robinson wrote. “A wallet whose history shows trades in a new meme coin looks like a speculator rather than a launderer, unless someone checks where its money came from.”

This isn’t the first time that North Korea has used a similar strategy. Last year, after the same group of North Korean hackers stole $1.5 billion from the crypto exchange Bybit, Elliptic found that the cybercriminals moved at least $24 million from the billion-dollar haul using meme coins.

More on the Binance hack: It Turns Out Binance’s $6 Billion Hack Happened Because Its Employees Were Vibe Coding

TL;DR: Get a lifetime Office 2019 license for one Windows PC for $15.98 with code DEAL20 through Nov. 15, including the full app suite with no recurring fees.

This is your chance to get a lifetime Office 2019 license for one Windows PC for $15.98. Simply use the code DEAL20 through Nov. 15 to qualify.

Here’s how the deal breaks down:

• Lifetime license to Word, Excel, PowerPoint, Outlook, OneNote, Publisher, and Access

• One-time purchase for one Windows 10 or 11 PC — device-based activation

• Improved inking, new Excel analysis tools, upgraded PowerPoint features

• Outlook refinements for email and contact management

• Instant email with download link and license key — redeem within 7 days

• All languages supported, updates included for the 2019 version

If your to-do list needs serious tools without a serious bill, this is the easy win — install once, get to work, and keep moving.

Grab a lifetime license to Microsoft Office Professional Plus 2019 for Windows for $15.98 with coupon code DEAL20 (retails at $229) until Nov. 15.

StackSocial prices subject to change.

Save $556: Whether you’re using AI to write, code, brainstorm or research, having access to multiple models can make a big difference. Right now, ChatPlayground AI is offering lifetime access to more than 20 leading AI models for just $63 through a limited-time StackSocial promotion that you can grab now, well before Black Friday kicks off. It is now available for just $63, and all you need to do is use code DEAL20.

It’s one of the biggest AI software discounts we’ve seen recently, and we’ll walk you through everything included before the deal expires.

ChatPlayground AI

$63 with code DEAL20 (Original price: $619)

ChatPlayground AI offers multiple AI models in a single app, so you can send prompts and requests and then compare the results that you receive. This takes the guesswork out of choosing the best model for a specific task and lets you select the best response from all 20-plus models.

As you might expect, you can upload images and PDFs, and your chat history is saved for easy reference and for ongoing projects. You can also choose to permanently delete chats if you prefer.

This unlimited tier removes all token limitations you would normally have to worry about, allowing you to send as many prompts as you like. This plan also gets you priority access to new features and future models as they become available.

Alternatives you might like

If this ChatPlayground AI deal doesn’t hit the spot, check out a similar StackSocial offering for 1min.AI. Its advanced business plan lifetime subscription is on sale for $100, a nice 81% discount and also brings multiple AI models under one roof.

Oliver Haslam has been writing about phones, computers, games, and anything else that takes a battery or plugs in for more than he’d like to admit.

With a focus on mobile and laptops, Oliver is never too far away from whatever social network is trending today and is never short of an opinion to share.

See full bio

A week after a humiliating midterms, Elon Musk is back to his old habits: flirting with Nazism.

Perhaps Musk really didn’t intend to perform back-to-back salutes that suspiciously looked a lot like a Sieg Hiel at a Trump inauguration party nearly two years ago, but you have to admit that he has quite the track record when it comes to espousing white supremacist conspiracy theories and interacting with people who wouldn’t hesitate to describe themselves as outright Nazis.

This time, his racist run-in was even called out by his own chatbot, Grok. On Sunday night, Musk responded to a supremacist meme posted by an account called @VladTheInflator railing against “white guilt” by downplaying the Atlantic slave trade in a way reminiscent of Holocaust denialism, while also perpetrating racist myths about slavery in Africa. Musk tagged Grok to ask if the meme’s claims were true, which the AI assessed to be “mostly accurate.”

This questionable interaction didn’t get overlooked by other users on the site, though, one of whom turned Grok back on its creator by asking the “maximum truth-seeking” AI to confirm if Musk was replying to a Nazi account.

“Yes,” Grok responded, seemingly atoning for all the times it called itself “MechaHitler.” “VladTheInflator has posted that being called a Nazi ‘feels like a compliment,’ that ‘Nazis are sounding better by the day,’ and that ‘Everything Hitler said about Zionists was true,’ plus explicit racial hierarchy claims.”

Grok explained in a helpful followup that Musk and the Nazi account go back. “Elon has replied to and retweeted VladTheInflator multiple times before this,” it said.

It’s not just that one account, either. Musk has an extensive history with promoting racist talking points and endorsing those posted by other far right figures, including a debunked conspiracy theory about “white genocide” in his homeland of South Africa.

His Grok-edited Wikipedia alternative, Grokipedia, has been caught citing actual neo-Nazi forums as sources of information. He also hosted an article writing contest on X and awarded the first place prize to a user who described himself as, and we quote, a “Nazi.” He even appeared at a rally of the German far-right party, AfD, telling the country that it needs to get over its guilt for the Holocaust.

More on Elon Musk: The Death of the Cybertruck Proves That Elon Musk Can’t Actually Predict the Future

Claude maker Anthropic has signed an unprecedented $3.4 billion deal with a Texas-based energy company called OnePWR Solutions.

As the Financial Times reports, the deal will provide gas-fired engines to provide AI data centers with off-grid electricity. The AI industry is racing to secure enough power and related equipment for its incredibly resource-intensive — and massively hated — facilities, and the Dario Amodei-helmed company is looking to cut out the middle man.

Anthropic’s deal, per the FT, is the “largest known single purchase of power-generation equipment by an AI lab.” It isn’t just symptomatic of a desperate race to secure the necessary utilities: it’s also one of the clearest signs yet that the AI industry has turned its back on the growing climate crisis in favor of pollution-spewing gas generators it’s installing across the globe.

Researchers have found that the effects of AI on the environment are nothing short of devastating. And the prognosis for the future is grim. Last month, environmental nonprofit Sierra Club estimated that upcoming AI data centers could release 2.1 billion metric tons of carbon dioxide by 2030, wiping out nearly 15 years worth of greenhouse gas reduction efforts.

After OnePWR’s attempts to acquire Caterpillar gas generators stalled thanks to a lack of funds, Anthropic decided to upgrade its client relationship to take over the Caterpillar orders itself as part of the latest deal.

The company’s competitors, including Elon Musk’s SpaceX and Google parent company Alphabet, have also bought energy developers and power equipment directly.

While analysts say Anthropic could be protecting itself from future energy equipment price swings as it prepares to go public, its deal with OnePWR once again highlights just how deep the relationship runs between dirty fossil fuels and frontier AI labs.

That reality is unlikely to win over the American public, which voted defiantly against data centers a mere week ago. Out of 41 data center-related ballot measures, only a single one swung in favor of the AI industry.

Beyond the staggering environmental damage, the facilities are also known to emit earsplitting noise pollution and raise local electricity prices, making them staggeringly unpopular.

More on data centers: New AI Data Center Is So Huge That It’ll Generate the Heat of 23 Atomic Bombs Every Single Day

Apple is reporting strong iPhone sales while its new CEO privately presses his software chief to move faster on artificial intelligence. Those two facts sitting side by side reveal the bind Ternus stepped into on September 1, 2026. The company’s investment in AI chips signals how seriously Apple takes the competitive pressure.

The New CEO’s First Real Test

Ternus came up through hardware. His most urgent problem is software.

Ternus succeeded Tim Cook, who moved into an executive-chairman role after leading Apple. Ternus built his reputation running hardware engineering, overseeing the chips and enclosures that define Apple’s products. Fortune has characterized his early tenure by noting the honeymoon is already over.

The Push Behind Closed Doors

Bloomberg reported a private meeting that signals how seriously Ternus takes the AI deficit.

According to Bloomberg, Ternus held a private meeting with software chief Craig Federighi and pressed Apple to accelerate its AI roadmap. The reported concern is specific: improving Siri alone was reportedly viewed as insufficient to close the gap with competitors. Bloomberg’s account is based on anonymous sources, and Apple has not publicly confirmed the meeting or its contents.

Why Siri Is Not the Whole Answer

Turning a voice assistant into an autonomous agent is harder than shipping a new chip.

Making Siri genuinely capable runs into real obstacles: privacy constraints, permissions architecture, and reliable execution at scale. Apple built its advantage around deep hardware-software integration, while competitors have pursued rapid AI-model iteration. Reconciling those two philosophies takes time, and time is a resource under pressure.

An Organization Built for a Different Clock

A reported mismatch between hardware timelines and software cycles is now a strategic liability.

Reports describe internal tension between hardware development cycles that run several years out and software planning windows of roughly 12 to 18 months, according to sources cited by Ynet News. Any AI push spanning devices, operating systems, and services simultaneously collides directly with that gap. Executive turnover in corporate development and App Store management, reported by The Verge, adds further difficulty in establishing a unified strategic direction.

Margin Pressure Arrives at the Wrong Moment

Rising memory costs and slowing services growth are tightening the window for a successful AI pivot.

Rising memory-chip prices are expected to squeeze Apple’s margins and could push product prices higher, according to Reuters. Analysts cited by Reuters also expect services-business growth to slow compared with earlier periods, raising the stakes for AI-powered offerings to generate new momentum.

Regulators Add Another Constraint

European rules on App Store access could limit how aggressively Apple moves in key markets.

Ongoing European Union scrutiny over App Store rules and platform access remains unresolved, according to CNN. Requirements around third-party distribution and AI assistants could restrict Apple’s options in markets where the company has significant exposure.

The Business Still Works. The Next Chapter Does Not Yet Exist.

Apple’s risk is not today’s balance sheet; it is the absence of a clear next growth engine.

Fortune reported record iPhone sales in a recent quarter, showing the hardware ecosystem has not lost its hold on consumers. The harder question is whether the next major growth engine exists before the current one plateaus. Ternus must build that engine without dismantling what already works.

Three questions will define his tenure: Can Apple close the AI gap without compromising its privacy identity? Can improved AI actually drive new revenue rather than simply defend existing ground? And can a hardware-trained CEO accelerate software decisions fast enough to matter? Apple has not disclosed its full AI roadmap. Bloomberg reports, however, that the internal urgency is real.