Latest Tech News

Stay updated with the latest in technology, AI, cybersecurity, and more

Filtered by: format Clear Filter

Fake accounts drove the DeepSeek AI hype and distorted markets

From AI Darling to Disinformation Playbook: How Fake Accounts Managed a Market Frenzy and What DeepSeek’s Hype Teaches Leaders DeepSeek, a new Chinese AI model, stormed app charts and briefly sent shockwaves through global markets. However, our disinformation research team found that much of this excitement was artificial — driven by thousands of fake profiles working in tandem. Their behavior bore the hallmarks of state-linked bot networks, amplifying hype and distorting market perception.

Credit Bureau TransUnion Hit With Data Breach Affecting 4.4 Million People

The TransUnion data breach exposed the data of up to 4.4 million customers. CNET Major credit bureau TransUnion has suffered a data breach that exposed sensitive information belonging to 4.4 million customers, including names and Social Security numbers. The data breach took place on July 28, meaning the sensitive information could have been sold and distributed throughout the dark web already. TransUnion is notifying people who may have been affected that their information could be exposed an

Incogni deletes your family’s private info from the internet – here’s why that matters

In today’s world, everyone needs to have dozens, if not hundreds of different online accounts to keep up with their everyday life. That’s a lot of potentially sensitive data in the hands of tons of different companies – and it’s easy to not think about that. You and your family’s data is incredibly valuable to data brokers, but it isn’t too difficult to put a stop to it – and protect the online privacy of your family. Protecting your online privacy It can be hard to take control of your online

RFC 8594: The Sunset HTTP Header Field (2019)

Internet Engineering Task Force (IETF) E. Wilde Request for Comments: 8594 May 2019 Category: Informational ISSN: 2070-1721 The Sunset HTTP Header Field Abstract This specification defines the Sunset HTTP response header field, which indicates that a URI is likely to become unresponsive at a specified point in the future. It also defines a sunset link relation type that allows linking to resources providing information about an upcoming resource or service sunset. Status of This Memo This docume

RFC 8594: The Sunset HTTP Header Field

Internet Engineering Task Force (IETF) E. Wilde Request for Comments: 8594 May 2019 Category: Informational ISSN: 2070-1721 The Sunset HTTP Header Field Abstract This specification defines the Sunset HTTP response header field, which indicates that a URI is likely to become unresponsive at a specified point in the future. It also defines a sunset link relation type that allows linking to resources providing information about an upcoming resource or service sunset. Status of This Memo This docume

TransUnion suffers data breach impacting over 4.4 million people

Consumer credit reporting giant TransUnion warns it suffered a data breach exposing the personal information of over 4.4 million people in the United States. TransUnion is one of the three major credit bureaus in the United States, alongside Equifax and Experian. It operates in 30 countries, employs 13,000 staff, and has an annual revenue of $3 billion. It collects and maintains credit information on over 1 billion consumers worldwide, with approximately 200 million of those based in the U.S.

Using information theory to solve Mastermind

How you've just played optimal Mastermind Mastermind is a game all about information. The Code Master selects one of \( 6^4 = 1\,296 \) secret codes. Each incorrect guess gives us information by eliminating some of these; the more codes that are ruled out, the more information that guess has provided. Let's quantify this insight! Suppose a guess gets some response that reduces the number of possible keys from some number \(n\) to a smaller \(n'<n\). The convention in information theory, a branc

Healthcare Services Group data breach impacts 624,000 people

The Healthcare Services Group (HSGI) is alerting more than 600,000 individuals that their personal information was exposed in a security breach last year. The healthcare services provider stated that it detected unauthorized access to its network on October 7, 2024, and subsequently discovered that the intrusion had begun on September 27. The investigation that followed revealed that the intruders had exfiltrated data from the systems they had accessed. “The investigation determined that an u

The Top Diseases We Choose to Stay Ignorant About, According to Scientists

The old adage “ignorance is bliss” feels especially fitting when it comes to healthcare. In fact, new research reveals that one in three people avoids—or is likely to avoid—medical information. In a study published in the Annals of Behavioral Medicine earlier this month, researchers investigated data from 92 studies involving 564,497 participants from 25 countries. Despite the fact that successful treatment often depends on early detection, their results indicate that many people are reluctant

Scientist exposes anti-wind groups as oil-funded. Now they want to silence him

Image: Empire Wind Oil-funded groups are engaging in strategic harassment to stop scientists from revealing the nature of their politically-linked disinformation networks – in what should be a surprise to nobody. A new report came out last week from the Climate & Development Lab (CDL) at Brown University, titled “Legal Entanglements: Mapping Connections of Anti-Offshore Wind Groups and their Lawyers in the Eastern United States.” The study focuses on several examples of law firms with connect

LiteLLM (YC W23) is hiring a back end engineer

TLDR LiteLLM is an open-source LLM Gateway with 27K+ stars on GitHub and trusted by companies like NASA, Rocket Money, Samsara, Lemonade, and Adobe. We’re rapidly expanding and seeking a founding full-stack engineer to help scale the platform. We’re based in San Francisco. What is LiteLLM LiteLLM provides an open source Python SDK and Python FastAPI Server that allows calling 100+ LLM APIs (Bedrock, Azure, OpenAI, VertexAI, Cohere, Anthropic) in the OpenAI format We have raised a $1.6M seed

Whistleblower claims DOGE uploaded Social Security data to unsecure cloud server

(Wesley Lapointe for The Washington Post via Getty Images) The Social Security Administration’s (SSA) chief data officer, Charles Borges, has filed a whistleblower complaint alleging that members of the Department of Government Efficiency (DOGE) uploaded a copy of a key Social Security database to an unsecured cloud environment in June, the New York Times reported. This may have exposed the personal information of hundreds of millions of Americans. The complaint alleges that under the authority

DOGE uploaded live copy of Social Security database to ‘vulnerable’ cloud server, says whistleblower

A top Social Security Administration official turned whistleblower says members of the Trump administration’s Department of Government Efficiency (DOGE) uploaded hundreds of millions of Social Security records to a vulnerable cloud server, putting the personal information of most Americans at risk of compromise. Charles Borges, the Social Security Administration’s chief data officer, said in a newly released whistleblower complaint published Tuesday that other top agency officials signed off on

How to stop AI agents going rogue

How to stop AI agents going rogue 1 hour ago Share Save Sean McManus Technology Reporter Share Save Getty Images Anthropic tested a range of leading AI models for potential risky behaviour Disturbing results emerged earlier this year, when AI developer Anthropic tested leading AI models to see if they engaged in risky behaviour when using sensitive information. Anthropic's own AI, Claude, was among those tested. When given access to an email account it discovered that a company executive was

Senator castigates federal judiciary for ignoring “basic cybersecurity”

US Senator Ron Wyden accused the federal judiciary of “negligence and incompetence” following a recent hack, reportedly by hackers with ties to the Russian government, that exposed confidential court documents. The breach of the judiciary’s electronic case filing system first came to light in a report by Politico three weeks ago, which went on to say that the vulnerabilities exploited in the hack were known since 2020. The New York Times, citing people familiar with the intrusion, said that Rus

The two versions of Parquet

A few days ago, the creators of DuckDB wrote the article: Query Engines: Gatekeepers of the Parquet File Format, which explained how the engines that process Parquet files as SQL tables are blocking the evolution of the format. This is because those engines are not fully supporting the latest specification, and without this support, the rest of the ecosystem has no incentive to adopt it. In my experience, this issue is not limited to Query Engines but extends to the tools within the ecosystem.

How to Fix Your Context

Mitigating & Avoiding Context Failures Following up on our earlier post, “How Long Contexts Fail”, let’s run through the ways we can mitigate or avoid these failures entirely. But before we do, let’s briefly recap some of the ways long contexts can fail: Context Poisoning: When a hallucination or other error makes it into the context, where it is repeatedly referenced. When a hallucination or other error makes it into the context, where it is repeatedly referenced. Context Distraction: When

Top Secret: Automatically filter sensitive information

We’ve written about how to prevent logging sensitive information when making network requests, but that approach only works if you’re dealing with parameters. What happens when you’re dealing with free text? Filtering the entire string may not be an option if an external API needs to process the value. Think chatbots or LLMs. You could use a regex to filter sensitive information (such as credit card numbers or emails), but that won’t capture everything, since not all sensitive information can

Apple drags ex-Apple Watch engineer to court over Oppo trade secret leak

Every now and then, we hear about Apple going after ex-employees who allegedly try to take trade secrets with them when they get hired by the competition. That’s the case with Chen Shi, a former Sensor System Architect for the Apple Watch team. Here are the details. According to Apple’s claim, Shi worked at the company from January 2020 until a couple of months ago, when he left to go work for Oppo, a Chinese company that also makes phones and wearables, like the Oppo Watch. Apple says that in

DaVita says ransomware gang stole data of nearly 2.7 million people

Kidney dialysis firm DaVita has confirmed that a ransomware gang that breached its network stole the personal and health information of nearly 2.7 million individuals. DaVita serves over 265,400 patients across 3,113 outpatient dialysis centers, 2,660 in the United States, and 453 centers in 13 other countries worldwide. The company reported revenues of over $12 billion in 2024 and of $3.3 billion for the second quarter of 2025. In April, the healthcare provider revealed in a filing with the U

Code formatting comes to uv experimentally

August 21, 2025 The latest uv release (0.8.13) quietly introduced an experimental new command that Python developers have been waiting for: uv format . This addition brings code formatting directly into uv’s toolkit, eliminating the need to juggle multiple tools for basic Python development workflows. What is uv format? The uv format command provides Python code formatting through uv’s interface. Under the hood, it calls Ruff’s formatter to automatically style your code according to consisten

Uv format: Code Formatting Comes to uv (experimentally)

August 21, 2025 The latest uv release (0.8.13) quietly introduced an experimental new command that Python developers have been waiting for: uv format . This addition brings code formatting directly into uv’s toolkit, eliminating the need to juggle multiple tools for basic Python development workflows. What is uv format? The uv format command provides Python code formatting through uv’s interface. Under the hood, it calls Ruff’s formatter to automatically style your code according to consisten

Project to formalise a proof of Fermat’s Last Theorem in the Lean theorem prover

Fermat’s Last Theorem An ongoing multi-author open source project to formalise a proof of Fermat’s Last Theorem in the Lean theorem prover. Information about the project The project is currently being led by Kevin Buzzard. It is funded by grant EP/Y022904/1, awarded by the UK’s Engineering and Physical Sciences Research Council. The project is hosted at Imperial College London. Kevin would like to extend many many thanks to both of these institutions for their ongoing support of this nonstand

Do Large Language Models Dream of AI Agents?

During sleep, the human brain sorts through different memories, consolidating important ones while discarding those that don’t matter. What if AI could do the same? Bilt, a company that offers local shopping and restaurant deals to renters, recently deployed several million agents with the hopes of doing just that. Bilt uses technology from a startup called Letta that allows agents to learn from previous conversations and share memories with one another. Using a process called “sleeptime compu

Magic Cue on Pixel 10: What is it and how will it make your life simpler?

TL;DR With the Pixel 10 series, Google has revealed a new AI feature called “Magic Cue.” Magic Cue watches your screen and offers suggestions by pulling relevant information from multiple apps. It currently supports a host of Google apps, with support for third-party apps promised to arrive soon. Despite nearly a decade of selling its Pixel phones (and several years with the Nexus devices before that), Google still strongly relies on unique software experiences to make its phones more compell

NY Business Council discloses data breach affecting 47,000 people

The Business Council of New York State (BCNYS) has revealed that attackers who breached its network in February stole the personal, financial, and health information of over 47,000 individuals. As the state's largest statewide employer association, BCNYS represents over 3,000 member organizations, including chambers of commerce, professional and trade associations, and other local and regional business organizations, as well as some of the largest corporations worldwide, which employ more than

T-Mobile claimed selling location data without consent is legal—judges disagree

A federal appeals court rejected T-Mobile's attempt to overturn $92 million in fines for selling customer location information to third-party firms. The Federal Communications Commission last year fined T-Mobile, AT&T, and Verizon, saying the carriers illegally shared access to customers' location information without consent and did not take reasonable measures to protect that sensitive data against unauthorized disclosure. The fines relate to sharing of real-time location data that was reveale

The Cutaway Illustrations of Fred Freeman (2016)

During the two-year research for our book LOOK INSIDE we discovered many amazing illustrations and artists that, for one reason or another, did not make it into the final version of the book. It would be a pity to leave these forgotten on a drawer, so during the next few weeks we will present here some of these masters of the cutaway. A while ago we wrote here about Frank Soltesz, an American illustrator active from the 30’s to the 60’s, and author of a marvelous series of architectural cutaw

HR Giant Workday Got Hacked

Workday, a company that provides human resources technology to over 11,000 corporations and 70 million users worldwide, announced in a classic Friday news dump that it suffered a data breach. The company did not disclose how much information was stolen by the hackers, but did reveal that information—including the names, email addresses, and phone numbers—of some users was compromised. The company said the breach hit some of its third-party customer relationship databases. If any other data was

Workday says hackers used social engineering to access personal data during a breach

Human resources technology company Workday has confirmed that a data breach has affected its third-party CRM platform. In a blog post announcing the breach, the company said that a social engineering campaign had targeted its employees, with threat actors posing as IT or HR in order to trick employees into sharing account access or personal information. The company says that while the threat actors were able to access some information from the CRM, there is no indication of any access to custom