Feed aggregator

Save on top services at LegalZoom, like LLC registration, incorporation, estate plans, and more with coupons and deals from WIRED.

Here are the answers for The New York Times Mini Crossword for Feb. 25.

• Indian loan company Vivifi has reportedly suffered a data breach
• 36 million files were left exposed
• These consisted primarily of personally identifiable information (PII)

A leading digital lending app has apparently exposed sensitive customer data after a misconfigured Amazon AWS S3 was bucket left unsecured without authentication.

Cybernews researchers discovered loan provider Vivifi left 36 million files of Know Your Customer (KYC) documents open online. The primary risk after a data breach is that criminals will use your information to apply for credit cards, loans, or bank accounts in identity theft or fraud schemes - so a loan company having customer information compromised would make it almost too easy for cybercriminals.

Included in the leak were passports, ID cards, driving licenses, utility bills, bank statements, and loan agreement letters, among other things - here’s what we know so far.

Ongoing investigation

Researchers discovered the leak on November 28, 2024, and the bucket wasn’t closed until January 16, 2025, meaning criminals had over a month to find and access the data - although there’s no evidence to suggest any did - only an internal forensic audit would determine this.

Know Your Customer (KYC) documents are used by financial institutions to ensure they comply with regulations and laws with regards to proof of identity, address, and income. Unfortunately though, this is all a cybercriminal would need to take out a loan in a victim’s name, or to craft particularly compelling social engineering attacks.

“For instance, attackers could use leaked loan agreement details or bank information to request urgent payments or account verification,” Cybernews researchers said.

“In some cases, these personal details can be aggregated and sold on the dark web, further escalating the danger and complicating efforts for victims to protect their privacy and secure their identities,” the team added.

Data breaches are all too common, and fintech firms aren’t immune. Earlier in 2025, Mexican FinTech firm Miio suffered a similar data breach which exposed millions of files of sensitive data - although significantly fewer than the Vivifi leak.

Serious risk for customers

This data breach is, unfortunately, the perfect opportunity for an attacker. The KYC documents are exactly what cybercriminals need to facilitate identity theft and fraud. With the identifying documents and personally identifiable information (PII), attackers can take out a loan, credit card, or create new bank accounts in your name.

To stay safe from this, the key is staying alert and monitoring your accounts. There are identity theft protection plans for individuals and for families, which essentially do the monitoring for you, and often provide $1 million or more in insurance plans, as well as dark web monitoring and anti-malware software - which can be very tricky to set up on your own.

If you want to do the monitoring yourself, perhaps you haven’t been directly impacted by a breach but want to stay protected - then here are the things to keep an eye out for.

First, is your bank statements, accounts, and transactions - if you see any suspicious activity, alert your bank immediately and freeze or pause your card if you can.

Next, create a strong and secure password for each individual account, or at least for the ones which hold financial, health, or sensitive information - and if a service you use is involved in a breach or cyberattack, make sure you change the password straight away.

Although it’s a pain, enabling multi-factor authentication or MFA is a great added layer of protection against intruders, so for those accounts with sensitive information - it's vital.

When PII is leaked, there’s always an added danger of social engineering attacks like phishing, which will use the data from the breach to determine which services you use regularly, what your interests are, or even your friends and family.

From there, attackers will send an email impersonating one of the above, and will trick you into clicking a malicious link, scanning a QR code, or handing over your details to them.

Be on the lookout for any unexpected communications, and look closely at the sender of emails - if you’re not sure, then don’t press any links, and search up what the legitimate email address would be - or contact the company directly through their website.

Remember, your bank will not ask you for your account details over the phone or through email - and they won’t ask you to transfer your funds to a different account.

You might also like

• Check out our list of the best firewall software around today
• Top IVF firm says hackers accessed private data during cyber incident
• We've also rounded up the best malware removal software on offer right now

If you use ChatGPT much, you've likely experimented with DALL-E 3, OpenAI’s latest iteration of an AI image generator. DALL-E 3 can be shockingly good at bringing your ideas to life but sometimes shockingly bad at understanding certain details or just shocking in what it chooses to emphasize. Still, if you want to see a specific scene move from your head to your screen, DALL-E 3 is usually pretty helpful, it can even make hands write.

But here’s the thing, DALL-E 3 is still an AI model, not a mind reader. If you want images that actually look like what you’re imagining, you need to learn how to speak its language.

After some trial and error (and a few unintentional horrors), I've become fairly adept at speaking its language. So here are five key tips that will help you do the same.

Polish in HD

DALL-E 3 has a quality mode called ‘hd,’ which makes images look sharper and more detailed. Think of it like adding a high-end camera lens to your AI-generated art – textures look richer, lighting is more refined, and things like fabric, fur, and reflections have more depth.

Check out how it looks in this prompt: "A rendering of a close-up shot of a butterfly resting on a sunflower. Quality: hd."

(Image credit: Image created with OpenAI's DALL-E 3) Play with aspect ratio

Not every image has to be a square. DALL-E 3 lets you set the aspect ratios, which may seem minor but can be huge if you want to make something look more cinematic or like a portrait.

Square is fine, but why limit yourself when you can make something epic? This is especially useful for social media graphics and posters, like the one below, which had the prompt: "A vertical poster of a vintage travel advertisement for Paris, size 1024x1792 pixels."

(Image credit: Image created with OpenAI's DALL-E 3) Think like a film director

To get an image capable of evoking emotion, sometimes it helps to think like you're a photographer in the real world. Think about camera angle or composition techniques; look them up if necessary. The result can dramatically change how an image looks.

Instead of a flat, dead-on image, you can request angles like close-up, bird’s-eye view, or over-the-shoulder. The same goes for composition styles and terms like ‘symmetrical composition’ or ‘depth of field.’

That's how you can get the following image from this prompt: "A dramatic over-the-shoulder shot of a lone cowboy standing on a rocky cliff, gazing at the vast desert landscape below. The sun sets in the distance, casting long shadows across the canyon. The cowboy's silhouette is sharp against the golden sky, evoking a cinematic Western feel."

(Image credit: Image created with OpenAI's DALL-E 3) Iterate, iterate, iterate

One of DALL-E 3’s lesser-known but highly effective tricks is telling it what not to include. This helps avoid unwanted elements in your image. That might mean specifying negative elements like colors, objects, or styles you don't want or refining the style and mood by what you don't want it to feel like.

That's how I got the image below, using the prompt: "A peaceful park in autumn with a young woman sitting on a wooden bench, reading a book. Golden leaves cover the ground, and a soft breeze rustles the trees. No other people, no litter, just a quiet, serene moment in nature."

(Image credit: Image created with OpenAI's DALL-E 3) Be overly specific

Think of DALL-E 3 as a very literal genie: it gives you exactly what you ask for, no more, no less. So if you type in “a dog,” don’t be surprised when it spits out a random dog of indeterminate breed, vibe, or moral alignment. The more details you include – like breed, color, setting, mood, or even art style – the better the results.

As an example, You might start with: “A wizard casting a spell," but you'd be better off submitting: “An elderly wizard with a long, braided white beard, dressed in emerald-green robes embroidered with gold runes, conjuring a swirling vortex of blue lightning from his fingertips in a stormy mountain landscape.” You can see both below.

(Image credit: Image created with OpenAI's DALL-E 3)

(Image credit: Image created with OpenAI's DALL-E 3) You might also like

• OpenAI confirms 400 million weekly ChatGPT users - here's 5 great ways to use the world’s most popular AI chatbot
• ChatGPT’s free tier just got a very handy image-making upgrade
• Try out Google's DALL-E defeater in Imagen 3, Gemini's new AI image generator

"No one knows what we are supposed to do," said one federal employee amid conflicting and shifting directives on whether to comply with Elon Musk's directive to list five accomplishments.

(Image credit: Tasos Katopodis)

• Netflix’s American Murder: Gabby Petito has upset some people for using an AI-generated voice to narrate Petito’s journal entries.
• Despite permission from Petito’s family, critics argue the AI voice raises ethical concerns.
• This isn't the first occurrence of such debate, and it will likely keep happening as the technology improves.

Netflix’s latest true-crime docuseries, American Murder: Gabby Petito, has stirred up a heated debate over how to deploy AI to mimic the voices of people who have passed away. The filmmakers employed AI to recreate Petito's voice and have it narrate excerpts from her personal writings, which has reportedly made many viewers feel uncomfortable and raised ethical concerns about using AI to give voice to the deceased.

The three-part series chronicles the 2021 murder of 22-year-old Petito at the hands of her fiancé, Brian Laundrie. It pieces together her final months through interviews, personal videos, and social media posts, evoking how the tragedy happened in real-time on the internet. True crime aficionados famously dissected every frame of Petito’s travel vlogs before authorities found her remains in Wyoming.

At the start of the series, a disclaimer appears: “Gabby’s journal entries and text messages are brought to life in this series in her own voice, using voice recreation technology.” That means the voice narrating parts of the documentary isn’t actually Petito’s but a synthetic recreation made with an AI model. Netflix has said the filmmakers received permission from Petito’s family to do so. That hasn’t stopped some people from vocalizing how eerie the AI-generated voice feels. Social media content creators have racked up hundreds of thousands of views discussing it.

AI ghosts

This isn't the first controversy over AI-generated voices. Roadrunner: A Film About Anthony Bourdain faced similar criticism when its director revealed that parts of the documentary featured AI-generated narration of Bourdain’s own words. That movie didn't indicate which bits were narrated by the AI or by Bourdain, which led many to feel that the technique was deceptive.

Filmmaker Michael Gasparro defended the decision in an interview with Us Weekly, saying the team wanted to tell the story as much “through Gabby’s voice as possible.” They had access to a wealth of her journals, notes, and online posts and thought AI narration would bring them to life in a more powerful way. “At the end of the day, it’s her story.”

Technology has always shaped the way we tell stories, but AI presents a new challenge, especially when it comes to memorializing people who can no longer speak for themselves. Robert Downey Jr. has vowed that AI will never replicate him on screen, while James Earl Jones secured a deal with Disney before passing away, allowing them to use his voice for Darth Vader under certain circumstances.

Meanwhile, ElevenLabs has inked deals with the estates of James Dean, Burt Reynolds, Judy Garland, and Sir Laurence Olivier to let it add AI versions of their voices to its Reader app. As deepfake technology and voice cloning become more sophisticated, filmmakers and media companies will have to reckon with how (and if) these tools should be used to tell real-life stories.

You might also like

• Robert Downey Jr. vows there will never be a digital AI replica of him on-screen
• How you can get (AI versions of) Judy Garland or Burt Reynolds to read to you
• Why the passing of James Earl Jones doesn't mean you won't hear Darth Vader's voice again

The Trump administration may continue — for now — to keep the AP from covering key events. A federal judge declined to issue a temporary restraining order on Monday.

(Image credit: JIM WATSON/AFP via Getty Images)

On Friday, Donald Trump fired Chairman of the Joint Chief's of Staff CQ Brown, along with several other top Pentagon officials.

Now, Senator Jack Reed, a Democrat from Rhode Island, has a question for the man tapped to succeed him, Retired Air Force Lieutenant General Dan Caine.

Quote — "will he have the ability to speak truth to power?"
Senator Reed is the top democrat on the Armed Services Committee.

The Trump administration says it wants a military built on meritocracy. Critics say it's building one governed by political loyalty.

For sponsor-free episodes of Consider This, sign up for Consider This+ via Apple Podcasts or at plus.npr.org.

Email us at considerthis@npr.org.

(Image credit: Chip Somodevilla)

The man poised to become Germany's new Chancellor says the US administration doesn't care about the fate of Europe. And, on the third anniversary of Russia's full-scale invasion of Ukraine what has been the impact of the largest conflict in Europe since the second world war?

(Image credit: Marcus Brandt)

More than 2,000 people could be displaced by the construction of the Río Indio dam. The Panama Canal Authority says the dam solves a long-term water shortage problem.

(Image credit: Tomas Ayuso for NPR)

Researchers say the Trump administration's plan to slash payments for indirect costs will hamper new medical science. One example? A lab studying respiratory viruses faces losing half its staff.

(Image credit: Rob Stein)

• Hainan’s underwater computing cluster matches the power of 30,000 gaming PCs operating simultaneously
• Seawater-cooled AI centers reduce energy costs and enhance security
• Microsoft abandoned underwater data centers after early trials

China has deployed 400 high-performance servers beneath the sea as part of an expanding underwater data center in Lingshui, Hainan.

China Media Group says this infrastructure will complete a year’s worth of computations for an average PC in just one second and can enable DeepSeek’s AI assistant to handle 7,000 conversations per second.

The newly-installed module, measuring 18 meters long and 3.6 meters in diameter, is linked to an existing facility to form a computing cluster designed to support AI-driven applications — providing computing power equivalent to 30,000 high-end gaming PCs operating simultaneously.

Expanding China's AI infrastructure

Xu Tan, vice president of Highlander, a Shenzhen-listed maritime technology firm and a key player in China's underwater data center expansion, highlighted the advantages of submerged data facilities.

"With the advent of the 5G and even 6G era, data will increase exponentially, and the construction of undersea data centers could meet the future demands of this growth," he said.

The Hainan underwater data center, first launched in 2023, was the first commercial project of its kind. According to CGTN, the long-term plan entails deploying up to 100 data cabins in multiple phases.

The project aligns with China’s strategy to enhance its AI capabilities, with 219 intelligent computing centers launched across 81 cities from 2022 to 2024, according to IDC (originally in Chinese).

The firm also projects China's intelligent computing capacity will grow at an annual rate of 33.9% until 2027, playing a crucial role in AI research, data processing, and automation across various industries.

Almost 10 companies have signed agreements to use the computing power of underwater data centers for AI model training, industrial simulations, game development, and marine research.

However, this project will require thousands of submerged servers, and the specific hardware to be used remains unknown. It is likely that high-performance AI accelerators, such as Nvidia’s H100 or Huawei’s Ascend 910, will be considered.

China has been pushing for domestic alternatives, so advanced AI chips from companies like Biren Technology or Loongson are also possibilities.

While China moves forward with plans to deploy up to 40,000 underwater servers in the coming years, American companies have taken a more cautious approach. Microsoft previously experimented with an underwater data center off the coast of the UK in 2018, but after retrieving it in 2020, the company ditched its National Park underwater data center plans.

Via Globaltimes

You may also like

• We've rounded up the best portable monitors available now
• Microsoft's deep-sea experiment revealed some very shocking server data
• Scientists say fiber optic cables could save lives by detecting tsunamis

In case you haven't been paying attention, Intel has been getting banged up lately in both the press and on Wall Street. Over the last 12 months, it's lost just under 43% of its share value as well as its CEO, Pat Gelsinger. It's watched Nvidia and AMD make major gains in the data center market at its expense, and its most recent client processors, the Intel Core Ultra 200S series, were rather underwhelming, especially for gaming.

Probably the biggest concern for Intel was its Q3 2024 earnings, which saw the company post a staggering $16.6 billion loss, the largest the company had ever seen. And while this was generated almost entirely by accelerated depreciation charges and the restructuring required after laying off 15,000 workers, rather than products sitting unsold on the store shelves, it still went off like a bomb on Wall Street and in Washington, where talk about how Intel could be "saved" remains rampant.

A lot of this talk has centered on Intel Foundry, the division of the company that actually manufactures semiconductors. Under Gelsinger, Intel invested billions of dollars into developing leading-edge process technology that could compete with Taiwan Semiconductor Manufacturing Company (TSMC), which manufactures chips for AMD, Apple, Nvidia, and even Intel, among many others.

So far, those efforts haven't borne fruit, and there have been many, many calls for Intel to dump its foundry business and focus on designing processors and outsourcing the actual manufacturing to TSMC.

All of this makes the newest process technology from Intel Foundry, Intel 18A, a make-or-break proposition for Team Blue, and no matter where you stand on the AMD vs Intel vs Nvidia, you should hope that Intel pulls off a miracle here.

What is a process node and what makes Intel 18A so special?

(Image credit: Intel)

Process node technology is very in-the-weeds stuff, and for someone looking to buy the best laptop or best processor for their PC, the question of which process node a chip was fabricated on rarely enters the discussion, even among tech enthusiasts.

A process node is the set of specific manufacturing processes used to fabricate a silicon wafer containing the chips that go into computers, as well as different innovations and features these chips include.

Traditionally, these nodes are referred to by a specific nanometer designation, i.e., a 3nm node, that is used to represent the physical length of the transistor gate. Over the last decade, though, these designations have become more marketing than anything, and the nanometer designation isn't directly tied to the physical size of the transistors.

Rather these designations say that a chip performs as well as a specific nanometer chip would if it were possible to produce transistors at such small scales, which nowadays is largely impossible (welcome to the end of Moore's Law). Now, these designations largely represent generational leaps in technology, and TSMC's process technology is currently the best there is.

But where and by whom the chip that powers your PC or graphics card is made can affect everything from its price to its availability. Right now, the vast majority of advanced computer chips, whether laptop processors or data center chips are manufactured by TSMC, with its most advanced N3 process node being in the most demand.

(Image credit: Alexander_Safonov / Shutterstock)

TSMC's next-gen process node, N2, is slated for volume production sometime in 2025, meaning it will be making production chips for AMD, Apple, Nvidia, and others later this year. The '2' in N2 is meant to represent 2nm, which puts it well beyond anything rival Samsung or Intel foundries have been able to produce.

Intel Foundry's 18A process, meanwhile, is a 1.8nm process, making it even more advanced than TSMC's. If Intel can deliver on the promises it has made about 18A, it would be the most advanced process technology in the world and would immediately upend the chip fabrication industry.

The fact that these chips would be produced in the United States makes the proposition even more enticing for American companies like Apple who have faced longstanding criticism over their offshore manufacturing and supply chain.

The most important result of Intel 18A's success for consumers, though, would be the long-term decrease in advanced chip prices and a more stable supply across several industries dependent on these advanced chips.

The COVID-19-driven supply chain issues in 2020 and 2021 revealed a real weakness in the globalized production model. While globalization has (mostly) led to lower prices for consumers, COVID showed that it only works when everything is running smoothly. Any disruption in one of the links in the supply chain can be felt for years in terms of higher prices and lower stock of the products we want to buy.

Having nearly all our advanced semiconductor supply tied to Taiwan is a recipe for disaster

(Image credit: Sundry Photography / Shutterstock)

Taiwan is a lovely country and it deserves all of the prosperity its semiconductor industry has brought to its people. But it is also a geopolitical crisis point vis-a-vis China, which believes Taiwan belongs under mainland rule.

This potential conflict over Taiwan means the global supply of the most advanced semiconductors could be threatened.

The existing status quo benefits Taiwan, as its semiconductor industry acts as a 'silicon shield' for the island, but for the rest of us, we're pretty much depending on China and the US to not escalate a conflict that, in even the best case scenario, could cripple the flow of advanced chips that the modern global economy depends on.

Reducing this dependence on TSMC is reason enough to want Intel 18A to succeed. If Intel can provide a genuine alternative for the most in-demand chip technology with the stability that comes from US production, it'll provide valuable insulation for the global economy and inject much-needed competition into the chip fabrication industry, bringing prices down for everyone.

What would 18A mean for Intel?

(Image credit: Intel)

Beyond the geopolitical issues, a successful rollout and adoption of 18A among industry customers like AMD, Nvidia, and possibly even Apple, would be a massive source of revenue for Intel separate from its own processor business, as well as providing an 'in-house' manufacturer of its processors, lowering their cost and giving it a competitive advantage over rival AMD.

Understandably, Intel is very bullish on 18A. It has to be since the future success of the company relies on 18A delivering the kind of performance that industry customers need, especially as AI is placing extraordinary demands on existing hardware.

If Intel 18A fails to establish Intel Foundry as a serious competitor to TSMC, it will likely be spun off as Wall Street is demanding. Without the support of the Intel Client side of the business to help it ramp up, it'll likely never reach the kind of advanced node technology that TSMC has, leaving us all dependent on a single supplier for cutting-edge chips and all the vulnerabilities and added cost that entails.

You might also like...

• The end of an era? TSMC, Broadcom could tear apart Intel's legendary business after 57 years by separating its foundry and chip design
• Hell freezes, pigs fly: Rumor has it that Intel could merge with AMD's former foundry in potential multi-billion deal
• Intel slams Nvidia and AMD, claims chip giants have huge numbers of security flaws

Here are some hints -- and the answers -- for Connections: Sports Edition No. 155 for Feb. 25.

U.S. Special Counsel Hampton Dellinger has asked the Merit Systems Protection Board to temporarily reinstate six federal employees fired from their jobs and is considering ways to seek relief for others.

(Image credit: Alex Wong)

Watch a scary flick while you wait for more White Lotus.

Dan Bongino rose through the ranks of right-wing media thanks to his unflinching loyalty to Donald Trump and willingness to push baseless conspiracies—including about the FBI.

Codes sent via SMS messages will be replaced by QR codes sent to user's phones to verify Gmail accounts.

Here are some hints -- and the answers -- for Strands No. 359 for Feb. 25.

Here are some hints -- and the answers -- for Connections No. 625 for Feb. 25.