TechRadar News

• Researchers spot Medusa ransomware operators deploying smuol.sys
• This driver mimics a legitimate CrowdStrike Falcon driver
• Medusa is actively targeting critical infrastructure organizations

Operators of the Medusa ransomware are engaging in old-fashioned bring-your-own-vulnerable-driver (BYOD) attacks, bypassing endpoint protection, detection and response (EDR) tools while installing the encryptor.

Cybersecurity researchers Elastic Security Labs noted the attacks start as the threat actors drop an unnamed loader, which deploys two things on the target endpoint: the vulnerable driver, and the encryptor.

The driver in question is smuol.sys, and it mimics a legitimate CrowdStrike Falcon driver named CSAgent.sys. It was also said to have been signed by a Chinese vendor the researchers dubbed ABYSSWORKER.

Monitor your credit score with TransUnion starting at $29.95/month

TransUnion is a credit monitoring service that helps you stay on top of your financial health. With real-time alerts, credit score tracking, and identity theft protection, it ensures you never miss important changes. You'll benefit from a customizable online interface with clear insights into your credit profile. Businesses also benefit from TransUnion’s advanced risk assessment tools.

Preferred partner (What does this mean?)View Deal

A growing threat

"This loader was deployed alongside a revoked certificate-signed driver from a Chinese vendor we named ABYSSWORKER, which it installs on the victim machine and then uses to target and silence different EDR vendors," Elastic Security Labs said in its report.

Using outdated and vulnerable drivers to kill antivirus and malware removal tools is nothing new. The practice has been around for years and is being used to deploy malware, steal sensitive information, propagate viruses, and more.

The best way to mitigate potential threats is to keep your software updated.

Medusa ransomware has grown into one of the most prolific Ransomware-as-a-service (RaaS) providers around.

Standing shoulder to shoulder with LockBit, or RansomHub, Medusa has taken responsibility for some of the biggest attacks in recent years, prompting the US government to issue a warning about its activities.

In mid-March 2025, the FBI, CISA, and MS-ISAC said Medusa targeted more than 300 victims from a “variety of critical infrastructure sectors”, by February 2025.

"As of February 2025, Medusa developers and affiliates have impacted over 300 victims from a variety of critical infrastructure sectors with affected industries including medical, education, legal, insurance, technology, and manufacturing," the report says. "FBI, CISA, and MS-ISAC encourage organizations to implement the recommendations in the Mitigations section of this advisory to reduce the likelihood and impact of Medusa ransomware incidents."

Via The Hacker News

You might also like

• US government warns Medusa ransomware has hit hundreds of critical infrastructure targets
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

On the list of high-priority concerns for chief information security officers (CISOs) these days, privacy is occupying an increasingly prominent position.

Privacy has always been important for businesses and other organizations handling sensitive customer and stakeholder information, but recent changes in privacy regulations are putting greater responsibility on CISOs, who are required to perform thorough risk assessments while making the results available on demand. Failure to conduct assessments or correct deficiencies can incur significant fines and, in some cases, even jail time.

CISOs need to take a proactive approach to meeting the challenge, particularly by performing rapid data security assessments to identify vulnerabilities and high-priority risks—including those involving third parties—and implement mitigations to protect data, the organization and its customers.

State regulators want risk assessments—now

A number of states have recently upped the ante for CISOs, requiring detailed risk assessments that must be produced upon request. State regulations that have added new requirements include the California Privacy Protection Agency (CPPA) draft regulations, Texas Data Privacy and Security Act (TDPSA), Virginia Consumer Data Protection Act (VCDPA), Colorado Privacy Act (CPA), and the New York SHIELD Act.

CISOs need to be on the same page with these new regulations, even if not all of them use the same playbook. While Virginia and Colorado, for example, set clear guidelines for assessments, California’s rules are more vague, requiring assessments without offering details.

The responsibility for implementing and documenting privacy controls and policies falls primarily on the shoulders of the CISO, who must ensure that the organization’s procedures for managing information protects privacy data and meets regulatory requirements. Performing risk assessments that identify weaknesses and demonstrate that they are being addressed is a crucial step in the process, even more so now that they must be ready to produce risk assessments whenever regulatory bodies request them.

As if CISOs needed an added incentive, regulators at the state and federal levels have been trending toward targeting organization management, particularly CISOs, in the wake of costly breaches. The consequences include hefty fines for organizations and, in worst-case scenarios, even jail sentences for CISOs.

Responsibility for privacy protections also extends to third-party risks. Organizations can’t afford to rely solely on promises made by third-party providers because regulators and state attorneys generally can hold an organization responsible for a breach, even if the exploited vulnerability belonged to a provider. Organizations need to implement a framework for third-party risk management that includes performing due diligence on the security postures of third parties.

Rapid risk assessments boost both security and compliance

Teams should follow several best practices in performing rapid assessments.

Automated scanning tools can identify a range of vulnerabilities, such as weak or non-existent authentication processes, unpatched and/or outdated software and hardware, and misconfigurations in the network. An assessment of internal risks can be combined with an analysis of common external threats. Security teams can then prioritize risks based on their threat levels and establish a plan for remediation.

Penetration testing serves as another critical tool that can help security teams quickly evaluate and assess potential threats to their infrastructure. By simulating a real-world attack, it is designed to show how an organization's layered controls worked together (or did not work) to defend against a hacker. As a result, organizations have a better understanding of their security posture and vulnerabilities that may attract bad actors.

In today’s threat landscape, with the constant drumbeat of sophisticated attacks, the process needs to be done efficiently, making rapid data security assessments an essential part of any risk analysis framework. Rapid assessments enable teams to quickly identify, prioritize and remediate the greatest risks while laying plans for further remediations. It allows them to determine the appropriate steps in each case, such as whether they need to implement encryption, access controls, intrusion prevention systems, firewalls or other measures.

After remediations have been applied, teams need to test the affected system to verify that the fixes have taken hold, and then conduct broader testing to ensure that systems are functioning as expected.

As a final step, organizations should implement a well-seasoned partner in Managed Services that can watch their security environments and remediate any vulnerabilities that occur. Out of office on vacation? Cybercriminals won’t hesitate to exploit your weaknesses over Thanksgiving dinner. Maintaining a security posture requires 24/7 dedication in today’s evolving and turbulent cyber landscape, which is where managed service providers can offer support.

Performing rapid assessments on a regular basis, such as every six months, is a good practice on its own for protecting against expensive and damaging data breaches, but it is also imperative to being able to comply with increasingly stringent privacy regulations. Documenting the steps they’ve taken, as well as those they plan to take, will enable CISOs to deliver the risk assessments regulators are looking for.

Conclusion

Security is a never-ending process, as CISOs are well aware, but so is compliance. Depending on the field they work in, businesses can face an array of compliance requirements that are frequently changing, as in the case of the recent updates to state privacy laws. In all, 20 U.S. states (so far) have privacy laws, and although there is no overarching federal privacy laws, many businesses must comply with international laws such as the European Union’s General Data Protection Regulation (GDPR) if they do business or monitor data subjects in the EU.

A proactive approach to security that makes use of automation in performing regular, repeatable rapid data security assessments as part of a robust risk management framework will enable CISOs to enhance data protections while keeping pace with the evolving compliance landscape.

We've featured the best online cybersecurity course.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

• Trusted Signing, a Microsoft certificate-signing service, is being abused by criminals, researchers are saying
• The criminals are signing malware with short-lived, three-day certificates
• Microsoft is actively monitoring for certificate abuse

Cybersecurity experts have warned Trusted Signing, Microsoft’s code-signing platform, is being abused to grant malware certificates and help it bypass endpoint protection and antivirus programs.

Certificates are digital credentials that verify the authenticity, integrity, and security of software. They use cryptographic keys to establish secure communications and prevent tampering or impersonation, and are considered crucial for encrypting sensitive data, ensuring secure transactions, and maintaining user trust. In software development, code-signing certificates validate that an application has not been altered after release.

Microsoft describes Trusted Signing as a, “fully managed, end-to-end signing solution that simplifies the certificate signing process and helps partner developers more easily build and distribute applications.”

Lumma Stealer and others

However, BleepingComputer reports multiple researchers observing threat actors using Trusted Signing to sign their malware with “short-lived, three-day code-signing certificates”.

Software signed this way will remain valid until the certificate is revoked, which suggests that the malware could successfully bypass security solutions for a lot longer.

The malware samples they analyzed were signed by "Microsoft ID Verified CS EOC CA 01,” it was said.

Among the campaigns abusing Microsoft are Crazy Evil Traffers’ crypto heist, and Lumma Stealer.

One of the ways Microsoft seems to be tackling this issue is to only allow certificates to be issued under the name of a company that’s been operational for at least three years.

However, individuals can sign up and get faster approval, if the certificate is issued under their name.

Microsoft says it is constantly monitoring the landscape and revoking certificates that were found to have been abused.

"When we detect threats we immediately mitigate with actions such as broad certificate revocation and account suspension. The malware samples you shared are detected by our antimalware products and we have already taken action to revoke the certificates and prevent further account abuse,” the company noted.

You might also like

• This VPN is being abused to spread malware
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• Indiana Jones and the Great Circle will come to PS5 on April 17, 2025
• Bethesda has confirmed this today after rumors pointed to an imminent reveal
• The game was exclusive to Xbox and PC upon its initial release last November

Indiana Jones and the Great Circle will be released for PlayStation 5 on April 27 this year, Bethesda has revealed.

Shared in a post across its social media profiles, the publisher of the game has revealed the much-anticipated PS5 release date with a video featuring a look at the upcoming box art, and presented by Troy Baker in discussion with familiar face Nolan North.

See what happens when @nolan_north sends @TroyBakerVA an urgent text...#IndianaJones and the Great Circle launches on PlayStation 5 on April 17! pic.twitter.com/zY3g608wWRMarch 24, 2025

This is superb news for PS5 players who have been waiting patiently to play as the whip-wielding hero, and its release on another platform also marks the next step in Xbox's multiplatform strategy, taking games to competing platforms to reach more people.

Bethesda had originally teased that something was going to be announced today with the below social media post featuring Troy Baker.

keep eyes here tomorrow pic.twitter.com/WjeA1NSe3rMarch 23, 2025

We first reviewed Indiana Jones and the Great Circle when it initially launched on Xbox Series X, Xbox Series S, PC, and Xbox Game Pass last November, and thoroughly enjoyed it.

Hardware writer Dashiell Wood said that it's "sure to go down as one of the greatest Indy adaptations in history." He went on to say that the game is "a globe-trotting adventure that will delight even those with just a passing knowledge of it. It’s also just a sheer joy to play thanks to some incredibly well-realized locations that are brimming with enough secrets to occupy you for hours at a time."

Fans of the movies who play on PS5, like myself, will now not have to wait long to dive in with pre-orders open right now. There's no news on specific PS5 features or DualSense Wireless Controller functionality yet, or indeed any graphical modes for PS5 Pro, so we'll ahve to wait and see if those details are announced in the run-up to the release date.

You might also like...

• The PS5 is currently selling faster than the PS4 did in the US, but I'm surprised to discover that the Xbox Series X and S are trailing behind Xbox One
• Sony patents PlayStation controller that you can charge by leaving in sunlight
• Minecraft developer rejects generative AI, 'it's important that it makes us feel happy to create as humans'

As AI shapes our world even further, we find ourselves at a defining intersecting moment for innovation and regulation. The EU AI Act officially going into effect and the emergence of tools like DeepSeek brings into focus the ethical implications of AI and the importance of responsible use.

Understanding and integrating AI responsibly doesn’t just stem from awareness, it requires a commitment to education, ethical practices and accountability. As technology propels us forward, businesses must uphold ethical standards to circumvent harm and bias and mitigate risk posed by AI.

By integrating a responsible approach to AI, businesses can continue to innovate while preventing misuse & misappropriation and fostering transparency. Employee training plays a pivotal role in this issue, ensuring a thorough understanding of AI ethics and compliance in practice. In a global context with varying regulatory frameworks, the challenge for companies is to build their own AI frameworks, creating a balance between regulations, ethics, and innovation. But how, and where, do we begin?

Start with compliance and employee training

As AI becomes a critical component of decision-making and daily operations, the importance of ethical AI training cannot be overstated. Organizations must recognize that implementing AI responsibly isn't just a technical challenge, but a people challenge too. A robust training program should cover key areas such as data privacy, misappropriation, transparency, accountability, and fairness, ensuring that AI use aligns with societal values and ethical principles. Neglecting this can lead to serious risks, including the misuse of AI tools, a loss of trust, damage to brand reputation, and even legal liabilities stemming from non-compliance.

To build an effective training strategy, it’s crucial to first assess the AI knowledge and skills of your workforce. Conducting baseline evaluations helps identify existing capability gaps, enabling leaders to design a training program that directly addresses those needs. Tracking progress over time ensures that employees continue to develop their skills and remain competent as AI technologies evolve.

Tailored development plans, which include regular feedback and guidance, empower employees to grow in their roles while fostering confidence in their ability to work with AI. It is crucial to understand how your organization intends to use AI, like specific use cases, and compare these needs with the skills of their workforce.

Role-specific risks must also be carefully considered. Not all employees interact with AI in the same way, so training should be customized to reflect their responsibilities. For example, employees handling sensitive data need advanced expertise in privacy protection and cybersecurity to minimize risks of data breaches.

Meanwhile, decision-makers must understand how to identify and address algorithmic bias to ensure fairness and equity in AI-driven outcomes. By creating role-based learning paths, organizations can prioritize the most relevant skills for each team member, optimizing the impact of training efforts.

Equally important is cultivating a culture of continuous learning. AI and its associated risks are constantly evolving, and regular risk assessments are essential to identify emerging knowledge gaps. Proactively updating training materials and programs helps employees stay prepared for new challenges and ensures they remain equipped to use AI responsibly over time. Additionally, incorporating practical, hands-on exercises, like simulated scenarios or ethical decision-making workshops, can reinforce learning and improve retention.

Implement ethical practices in your organization

The effective integration of AI into business processes requires organizations to adopt ethical practices that prioritize privacy, fairness, transparency, and sustainability. As AI becomes more embedded in decision-making and operations, it’s essential to ensure that its use aligns with both legal standards and ethical principles. This begins with establishing clear and prescriptive policies that outline what is and is not acceptable behavior when it comes to AI applications. These policies should provide guidance on data usage, decision-making, and accountability to prevent misuse or harm.

Compliance with global privacy regulations, such as GDPR and CCPA, is paramount. Organizations must ensure that data collection, storage, and usage practices align with these legal requirements, safeguarding consumer trust and protecting sensitive information. In addition, ethical frameworks need to account for current and emerging policies, like the EU AI Act, which aims to ensure that AI technologies are inclusive, transparent, and safe for users. This requires a proactive approach to understanding and implementing these standards.

To achieve this, organizations should establish a compliant governance structure that includes explicit policies and procedures for ethical AI use. This might involve regular auditing, rigorous testing of AI systems, and continuous monitoring to identify and mitigate potential risks. When businesses take these steps, they not only meet regulatory requirements but also build trust with their stakeholders and contribute to the sustainable and fair use of AI technology.

Investing in AI ethics training and adopting robust ethical practices are essential steps toward responsible and sustainable AI development. These efforts go beyond mere safeguards—they represent a strategic advantage. By integrating ethical principles into their AI workflows, backed with continuous and effective training, organizations can foster innovation responsibly, build enduring trust, and position themselves as leaders in shaping a future where AI serves the greater good.

We've featured the best productivity tool.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro

Looking for a different day?

A new NYT Strands puzzle appears at midnight each day for your time zone – which means that some people are always playing 'today's game' while others are playing 'yesterday's'. If you're looking for Monday's puzzle instead then click here: NYT Strands hints and answers for Monday, March 24 (game #386).

Strands is the NYT's latest word game after the likes of Wordle, Spelling Bee and Connections – and it's great fun. It can be difficult, though, so read on for my Strands hints.

Want more word-based fun? Then check out my NYT Connections today and Quordle today pages for hints and answers for those games, and Marc's Wordle today page for the original viral word game.

SPOILER WARNING: Information about NYT Strands today is below, so don't read on if you don't want to know the answers.

NYT Strands today (game #387) - hint #1 - today's theme What is the theme of today's NYT Strands?

• Today's NYT Strands theme is… For starters

NYT Strands today (game #387) - hint #2 - clue words

Play any of these words to unlock the in-game hints system.

• PROP
• SAME
• SAWING
• CHOP
• CHAP
• POEM

NYT Strands today (game #387) - hint #3 - spangram letters How many letters are in today's spangram?

• Spangram has 9 letters

NYT Strands today (game #387) - hint #4 - spangram position What are two sides of the board that today's spangram touches?

First side: left, 4th row

Last side: right, 5th row

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

NYT Strands today (game #387) - the answers

(Image credit: New York Times)

The answers to today's Strands, game #387, are…

• SLIDERS
• POPPERS
• WINGS
• FONDUE
• NACHOS
• CALAMARI
• SPANGRAM: APPETIZER

• My rating: Easy
• My score: Perfect

The WINGS phenomenon is fascinating. Once the part of a chicken no one wanted to eat, they are now so popular that there’s a few weeks every year where people panic on social media over a shortage when the supermarket shelves are temporarily bare and word spreads that Domino's is running short.

It’s no wonder there are supply fluctuations. According to the National Chicken Council, Americans consumed 1.5 billion chicken wings over Super Bowl weekend this year – that’s up 20 million on the previous year.

How are there enough birds to cope?

That said, although there are no chains exclusively selling squid – Squid Stop, anyone? – it’s amazing that there’s enough of them caught to cope with the demand for CALAMARI, truly the king APPETIZERs.

Although quality can vary between sublime and rubber band, I can’t look anywhere else if ever I see them on a menu.

How did you do today? Let me know in the comments below.

Yesterday's NYT Strands answers (Monday, 24 March, game #386)

• EAGLES
• KISS
• RUSH
• GENESIS
• JOURNEY
• HEART
• BLONDIE
• SPANGRAM: ROCK BAND

What is NYT Strands?

Strands is the NYT's not-so-new-any-more word game, following Wordle and Connections. It's now a fully fledged member of the NYT's games stable that has been running for a year and which can be played on the NYT Games site on desktop or mobile.

I've got a full guide to how to play NYT Strands, complete with tips for solving it, so check that out if you're struggling to beat it each day.

Looking for a different day?

A new Quordle puzzle appears at midnight each day for your time zone – which means that some people are always playing 'today's game' while others are playing 'yesterday's'. If you're looking for Monday's puzzle instead then click here: Quordle hints and answers for Monday, March 24 (game #1155).

Quordle was one of the original Wordle alternatives and is still going strong now more than 1,100 games later. It offers a genuine challenge, though, so read on if you need some Quordle hints today – or scroll down further for the answers.

Enjoy playing word games? You can also check out my NYT Connections today and NYT Strands today pages for hints and answers for those puzzles, while Marc's Wordle today column covers the original viral word game.

SPOILER WARNING: Information about Quordle today is below, so don't read on if you don't want to know the answers.

Quordle today (game #1156) - hint #1 - Vowels How many different vowels are in Quordle today?

• The number of different vowels in Quordle today is 4*.

* Note that by vowel we mean the five standard vowels (A, E, I, O, U), not Y (which is sometimes counted as a vowel too).

Quordle today (game #1156) - hint #2 - repeated letters Do any of today's Quordle answers contain repeated letters?

• The number of Quordle answers containing a repeated letter today is 1.

Quordle today (game #1156) - hint #3 - uncommon letters Do the letters Q, Z, X or J appear in Quordle today?

• No. None of Q, Z, X or J appear among today's Quordle answers.

Quordle today (game #1156) - hint #4 - starting letters (1) Do any of today's Quordle puzzles start with the same letter?

• The number of today's Quordle answers starting with the same letter is 2.

If you just want to know the answers at this stage, simply scroll down. If you're not ready yet then here's one more clue to make things a lot easier:

Quordle today (game #1156) - hint #5 - starting letters (2) What letters do today's Quordle answers start with?

• S

• W

• F

• S

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

Quordle today (game #1156) - the answers

(Image credit: Merriam-Webster)

The answers to today's Quordle, game #1156, are…

• SNEER
• WATCH
• FLORA
• STAIN

My winning streak continues, thanks in large part to getting plenty of green and yellow hits with my start words.

With three options on my final word (either batch, patch or WATCH) I managed to pick the correct one.

I’d like to pretend there was some science to this choice, but it was pure luck. That said, I did ask myself, “what is the most Quordley of all these words?”, which is usually the word you are least likely to think is correct.

How did you do today? Let me know in the comments below.

Daily Sequence today (game #1156) - the answers

(Image credit: Merriam-Webster)

The answers to today's Quordle Daily Sequence, game #1156, are…

• CROWN
• ANGRY
• RATIO
• CLONE

Quordle answers: The past 20

• Quordle #1155, Monday 24 March: FLOUT, SMASH, RAZOR, SMALL
• Quordle #1154, Sunday 23 March: ALLEY, GAMUT, GEESE, RADII
• Quordle #1153, Saturday 22 March: PICKY, BRACE, BOOZE, SURLY
• Quordle #1152, Friday 21 March: DITTY, VALOR, TOWER, AMISS
• Quordle #1151, Thursday 20 March: RIPER, STILL, WAFER, NOTCH
• Quordle #1150, Wednesday 19 March: FLAME, CARAT, SPAWN, TUNIC
• Quordle #1149, Tuesday 18 March: NAVEL, TREAT, COPSE, SLINK
• Quordle #1148, Monday 17 March: INFER, FLINT, CHUMP, SHEER
• Quordle #1147, Sunday 16 March: INPUT, RUMBA, ELUDE, BLANK
• Quordle #1146, Saturday 15 March: SPEED, CLASH, RISKY, GRATE
• Quordle #1145, Friday 14 March: BELIE, SPEAR, GROWL, ADOBE
• Quordle #1144, Thursday 13 March: AMISS, METAL, TEARY, SHORT
• Quordle #1143, Wednesday 12 March: AWASH, PUSHY, RALPH, PROOF
• Quordle #1142, Tuesday 11 March: LINGO, JUICE, SHARP, REVEL
• Quordle #1141, Monday 10 March: AVAIL, THESE, TITAN, GAUGE
• Quordle #1140, Sunday 9 March: SKILL, IDEAL, WHERE, REHAB
• Quordle #1139, Saturday 8 March: WATCH, LAPSE, WREST, HEFTY
• Quordle #1138, Friday 7 March: GLEAN, SHINY, DECRY, MANGA
• Quordle #1137, Thursday 6 March: TRICE, EIGHT, BELCH, RIPER
• Quordle #1136, Wednesday 5 March: PLEAD, TWANG, MAJOR, RISKY

Looking for a different day?

A new NYT Connections puzzle appears at midnight each day for your time zone – which means that some people are always playing 'today's game' while others are playing 'yesterday's'. If you're looking for Monday's puzzle instead then click here: NYT Connections hints and answers for Monday, March 24 (game #652).

Good morning! Let's play Connections, the NYT's clever word game that challenges you to group answers in various categories. It can be tough, so read on if you need Connections hints.

What should you do once you've finished? Why, play some more word games of course. I've also got daily Strands hints and answers and Quordle hints and answers articles if you need help for those too, while Marc's Wordle today page covers the original viral word game.

SPOILER WARNING: Information about NYT Connections today is below, so don't read on if you don't want to know the answers.

NYT Connections today (game #653) - today's words

(Image credit: New York Times)

Today's NYT Connections words are…

• BALE
• CRYSTAL
• PANDA
• MUD
• TEDDY
• TUXEDO
• ABLE
• DAY
• ABEL
• CAL
• A BELL
• OREO
• ABE
• CROSSWORD
• BELA
• DICK

NYT Connections today (game #653) - hint #1 - group hints

What are some clues for today's NYT Connections groups?

• YELLOW: Monochrome
• GREEN: Letter jumble 
• BLUE: Head of state
• PURPLE: Sayings, apparently

Need more clues?

We're firmly in spoiler territory now, but read on if you want to know what the four theme answers are for today's NYT Connections puzzles…

NYT Connections today (game #653) - hint #2 - group answers

What are the answers for today's NYT Connections groups?

• YELLOW: BLACK-AND-WHITE THINGS
• GREEN: ANAGRAMS 
• BLUE: U.S. PRESIDENTIAL NICKNAMES 
• PURPLE: CLEAR AS ___ 

Right, the answers are below, so DO NOT SCROLL ANY FURTHER IF YOU DON'T WANT TO SEE THEM.

NYT Connections today (game #653) - the answers

(Image credit: New York Times)

The answers to today's Connections, game #653, are…

• YELLOW: BLACK-AND-WHITE THINGS CROSSWORD, OREO, PANDA, TUXEDO
• GREEN: ANAGRAMS ABEL, ABLE, BALE, BELA
• BLUE: U.S. PRESIDENTIAL NICKNAMES ABE, CAL, DICK, TEDDY
• PURPLE: CLEAR AS ___ A BELL, CRYSTAL, DAY, MUD

• My rating: Moderate
• My score: 1 mistake

Today’s puzzle was CLEAR AS ___ MUD to begin with and I foolishly wasted time thinking that one group was made up of famous actors CRYSTAL (Billy), BALE (Christian) and, stretching the limits of logic, DAY (Daniel Day-Lewis), and er DICK (Van Dyke).

Next, I pondered on A BELL being an abbreviation of Alexander Graham Bell, the telephone inventor.

And then, well it became as clear as CRYSTAL.

I wasn’t sure if OREO qualified but I was confident that CROSSWORD, PANDA, and TUXEDO are BLACK-AND-WHITE THINGS.

The unexpected joy of snagging a purple foursome followed, before I finally clocked that ABEL, ABLE, BALE and BELA contained the same four letters and were ANAGRAMS of each other.

All a blessing, as I would never have gotten U.S. PRESIDENTIAL NICKNAMES in a month of Super Tuesdays.

How did you do today? Let me know in the comments below.

Yesterday's NYT Connections answers (Monday, 24 March, game #652)

• YELLOW: BAD-SMELLING FOUL, RANK, RIPE, SOUR
• GREEN: UNFLUCTUATING CONSTANT, LEVEL, STABLE, UNIFORM
• BLUE: WHO VIDEO GAMES ARE FOR, PER ESRB RATINGS ADULTS ONLY, EVERYONE, MATURE, TEEN
• PURPLE: THINGS WITH LAYERS EARTH, HENHOUSE, ONION, PHOTOSHOP

What is NYT Connections?

NYT Connections is one of several increasingly popular word games made by the New York Times. It challenges you to find groups of four items that share something in common, and each group has a different difficulty level: green is easy, yellow a little harder, blue often quite tough and purple usually very difficult.

On the plus side, you don't technically need to solve the final one, as you'll be able to answer that one by a process of elimination. What's more, you can make up to four mistakes, which gives you a little bit of breathing room.

It's a little more involved than something like Wordle, however, and there are plenty of opportunities for the game to trip you up with tricks. For instance, watch out for homophones and other word games that could disguise the answers.

It's playable for free via the NYT Games site on desktop or mobile.

• Apple is rolling out lossless audio and ultra-low-latency support to AirPods Max with USB-C
• You'll still need a cable though: Apple's introduced a new USB-C to 3.5mm audio option
• It's the first major software update for Apple's over-ears, which were refreshed with USB-C in September 2024

While iOS 18.4 won't introduce the AI-infused Siri, it will expand the feature set and capabilities of Apple’s AirPods Max. The Cupertino giant's flagship over-ear headphones last received an update in September 2024, with a USB-C port and new colors, but no changes to the audio performance.

That will change with iOS 18.4, though, as when Apple ships the software update in April, AirPods Max will get support for lossless and ultra-low latency audio when used with a cable. Yes, note the catch: it won’t be available wirelessly.

While you might be thinking you’re limited to a USB-C to USB-C cable, ie. the one that comes in the box with AirPods Max, Apple is also – finally – bringing back a USB-C to 3.5mm audio cable to let you unlock listening in better quality with more devices. Yes, you'll need to get one of those, but if you wanted to listen to Apple Music's Lossless output on some of the best hi-res audio players it is now an option.

(Image credit: Apple)

Now, Apple Music's Lossless audio on AirPods Max with a cable will be supported up to 24-bit/48 kHz, a healthy improvement over than anything you'll get over Bluetooth. That is also an improvement over AirPods Max with Lightning, as when used with a cable, you could hit 24-bit/48 kHz audio, but it went through some additional processing.

It’s worth noting, though, that this is below Apple Music's maximum Hi-Res Lossless output – that is, 24-bit/192kHz – so you’d need one of the best headphone DACs to unlock the streaming service's full potential.

Even so, this should improve the listening experience, and lossless audio here will also be compatible with ‘Personalized Spatial Audio’ to improve immersion, and dynamic head-tracking will still be enabled here. Apple also notes that this makes the AirPods Max the first headphones to let you create, mix, and listen in the spatial audio standard with head-tracking as well.

If you’re after the lowest latency possible, maybe for the first listen of a new album or to give you a leg up when playing a AAA title like Resident Evil 3 or Assassin's Creed Shadows, plugging in via the cable is likely worth it.

Apple says the ultra-low-latency experience with AirPods Max is on par with the speakers built into the iPhone, iPad, and Mac. This means that the AirPods Max should offer a meaningful improvement in lowering lag, making them a better fit for gaming or even live streaming your playing.

Two features arriving in the form of a free update is excellent news, but Apple also releasing a USB-C to 3.5 mm audio cable is fantastic news. It was a weird omission not to offer this alongside the original launch of AirPods Max with USB-C, and it is a very welcome addition to Apple’s cable lineup.

(Image credit: Apple)

The cable is up for order now at $39.99 / £39 / AU$65 in one size – 1.2 meters – and only comes in the color white. I do hope Apple might eventually drop it in color-matched options to the AirPods Max with USB-C – purple, blue, orange, Midnight, or Starlight – though.

We also know that iOS 18.4 – as well as iPadOS 18.4 and macOS Sequoia 15.4 – will arrive in April of 2024, and you’ll need to install that on your iPhone, iPad, or Mac to get your AirPods Max with USB-C to be updated. In fact, just a few weeks back, Apple finally shed some light on how to update your AirPods.

There is also no change to pricing for AirPods Max with USB-C, in any color these over-ear cans are $549 / £499 / AU$899. Though right now, in the United States, on Amazon, they are just $479.99.

Much like Adaptive Audio or the hearing health features arriving for AirPods Pro, Apple continues to have an excellent pattern here – and that’s offering up new features months down the line as a free upgrade for its AirPods customers.

It stretches the value significantly, especially for a product like AirPods Max, which doesn’t get as frequently refreshed as AirPods Pro or AirPods – it’s nice to see feature drops like this arrive.

Let’s hope there will be more in store at Apple’s WWDC 2025, which should take place in just a few short months in June of 2025 if that tradition continues.

You might also like

• The best headphones 2025 tested by our experts, for all budgets
• iOS 18.4: 5 new features to expect, including Ambient Music and Photos filtering
• AirPods with cameras for Visual Intelligence could be one of the best personal safety features Apple has ever planned – here's why
• ‘We only build technology in the interest of storytelling’ – Disney’s associate lab director of Robotics on the Star Wars BDX Droids and what lies ahead

@techradar

♬ original sound - TechRadar

ChatGPT was down for many on Monday (March 24), especially users in the US, where prompts were returning error messages across multiple models, including 4o and 03-mini.

We first noticed the issue on a number of our own ChatGPT queries and then confirmed that others are reporting similar issues with the artificial intelligence platform on Downdetector and on Reddit. Dwondetector, which tracks outages across the web, noted outage reports starting at roughly 9AM ET and growing since then until then started declining around 11AM ET.

ChatGPT and Sora most recently suffered major outages late last year and in January but have been fairly stable since then.

Today's issues, which were acknowledged by OpenAI on its status page, lasted roughly three hours. However, the worst of them was only about 45 minutes (we couldn't process prompts and Sora was unavailable).

OpenAI has not published any further details on the outage beyond the acknowledgment and mitigation measures.

No answers

(Image credit: Future)

This is what the outage looks like when you're trying to get a prompt response from ChatGPT 4o.

What's notable here is that the platform is not down but ChatGPT's ability to answer after ingesting a prompt appears compromised.

ChatGPT has yet to acknowledge any issues on its X (formerly Twitter) feed but we'll keep an eye on it for updates.

Up not down

(Image credit: Future)

Downdetector's report tracker for OpenAI services has been rising since the morning (9AM ET) and steadily rising since then.

It's worth noting that the service is tracking all OpenAI services and not just ChatGPT. However, most reports we're seeing elsewhere only point to ChatGPT as the primary culprit.

The outage may not be global since our counterparts in the UK report no issues with processing prompts.

We're seeing them on both the desktop and iOS app.

Reddit gets it

(Image credit: Future)

While not everyone is reporting issues, the outage has risen to what we like to call "Reddit Level". That's when word of it props up on Reddit, a place where someone can stand up and say (post), "YoC chatGPT is not working for me. Whatsup?!" Reddit's 1.6B users will then be on the case.

In this case, the user asked if it was down and then noted how his chat history appeared to have disappeared. Thankfully, our chat history is still in place.

Sora block

(Image credit: Future)

While we don't have many individual outage reports for Sora, we did find that we can no longer log into the generative video AI platform. That's a bit worse than ChatGPT where we can at least log in and enter a prompt that currently goes nowhere.

Downdetector doesn't break out individual OpenAI service platforms, but, since this is an OpenAI outage, we can safely assume that it may be affecting more than just ChatGPT.

Signs of life

(Image credit: Future)

While we successfully entered a prompt and received a response from ChatGPT, the results spit out slowly, like a second per word. So there is still sluggishness here and we're not clear if OpenAI and ChatGPT are out of the woods, yet.

We've also contacted OpenAI for comment on the service issue and we'll update our live coverage with its response.

OpenAI confirms

(Image credit: Future)

While we have yet to hear back from OpenAI, its own status system reveals ongoing issues.

"We're currently experiencing issues," notes the page and specifically points to ChatGPT and "increased error rates."

Even though OpenAI doesn't call put Sora, we can see from the chart that it too has had a few blips this morning.

Sora coming back

(Image credit: Future)

We were able to successfully log into Sora, enter a prompt, and generate a video. That's good news for all you fans of anthropomorphic animals (our own personal favorite generative video subject).

Still, if OpenAI's servers are having trouble, you may want to wait a bit for the platform to clear the decks before you hand it any mission-critical work.

The recovery continues

(Image credit: Future)

All of ChatGPT's key models are now working again – at least in our tests – but the results still flow in very slowly. While a bit frustrating, it's preferable to the models being completely unavailable.

Similarly, Downdetector reports for OpenAI have been on the decline, and that's a good sign from the wider audience.

A return to normal

(Image credit: Future)

As you can see from the above chart, OpenAI, and ChatGPT in particular, is returning to normal operations.

OpenAI offered a small update on its status page regarding the outage, noting that it's identified the issue and is working on "implementing mitigation." Whatever OpenAI is doing, it appears to be working.

(Image credit: Future) Regular speed

(Image credit: Future)

Tests of multiple ChatGPT models show the generative AI platform is not only back, it's essentially at full speed. 4o, o1, and 03-mini are all returning answers almost instantly (the reasoning ones are designed to take a beat before spitting out a response).

Tough day for many AI

(Image credit: Future)

When ChatGPT outage reports first surfaced, we noticed that Perplexity was, according to Downdetector. It quickly recovered but now we're seeing Anthropic's Claude users reporting service issues.

Granted, Claude AI is not nearly as popular or widely used as ChatGPT but one has to wonder if these outages are in any way connected.

It is possible that some systems might be relying on the same web service provider (say they were all using AWS). Not likely, but the coincidence is interesting and, for now, unexplained.

Meanwhile, ChatGPT is almost back to normal.

It could've been worse

(Image credit: Future)

A closer look at OpenAI's status page reveals an over 99% uptime for Sora and ChatGPT, a great record by any measure. Plus, this outage, which appears nearly resolved, did not spread globally.

Multiple check-ins with our UK counterparts found a responsive ChatGPT service capable of instantly answering any prompt.

In fact, this outage didn't even warrant a red bar. There were outages in December 2024 and January of this year which were much more severe.

(Image credit: Future) It's over

(Image credit: Future)

According to OpenAI's status page, there are no longer any issues. TechRadar can confirm that ChatGPT is now functioning normally. Now it's time to get back to work with these penguins.

(Image credit: Sora)

• Researchers claim primary target of a recent cascading supply chain attack was Coinbase
• The cryptocurrency exchange was not compromised, but hundreds of other projects might suffer
• The attack went through a GitHub Action tool

The endgame of the recent cascading supply chain attack on GitHub was to breach Coinbase, one of the world’s most popular centralized cryptocurrency exchanges, experts have claimed.

Cybersecurity researchers Unit 42 (Palo Alto), and Wiz, revealed the attack, noting although Coinbase successfully defended itself, it is difficult to deem the attack a failure, since hundreds of other projects suffered as collateral damage.

Coinbase claims no damage was done - however, 218 other repositories are thought to have been impacted as a result of this attack.

No damage to Coinbase

A cascading supply chain attack is a cyberattack where compromising one component, such as a software dependency or tool, triggers a chain reaction that spreads the breach to multiple connected systems or projects.

In this case, cybercriminals tampered with a small tool, a GitHub Action called reviewdog/action-setup@v1. It is a popular tool that helps automate tasks in software projects. How they breached this Action wasn’t revealed, but the attackers managed to get the tool to leak certain access codes into publicly visible logs.

They then used these codes to inject more malicious code into another widely used tool, called tj-actions/changed-files. This tool is part of Coinbase's development process, and by doing so, they tried to move into the exchange’s code repository, gain deeper access, and wreak more havoc.

"The attacker obtained a GitHub token with write permissions to the coinbase/agentkit repository on March 14, 2025, 15:10 UTC, less than two hours before the larger attack was initiated against tj-actions/changed-files," Palo Alto Unit 42 said.

"We followed up by sharing more details of our findings with Coinbase, which stated that the attack was unsuccessful at causing any damage to the agentkit project, or any other Coinbase asset," the researchers added.

Once the threat actors realized their attack against Coinbase was unsuccessful, they pivoted to other projects, the researchers said. We don’t know if any other attacks were more fruitful for the criminals.

Via BleepingComputer

You might also like

• This VPN is being abused to spread malware
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• The Microsoft Surface 7 has been labeled "frequently returned" on Amazon
• No reason was officially given, but user reviews cite Arm-based incompatibilities
• It's believed this could be happening due to more competition being available

The Qualcomm Snapdragon X Elite-powered Microsoft Surface Laptop 7 has been deemed "frequently returned" on Amazon.

Specificially, it's an Amazon listing for the Microsoft Surface Laptop (2024) featuring a 15-inch touchscreen, Snapdragon X Elite SoC, 32GB RAM, and 1TB of storage.

Despite having nearly 400 user reviews, with an overall rating of 4.2 / 5 stars, the laptop appears to be returned often according to the retailer, with the disclaimer telling prospective buyers to "check the product detais and customer reviews to learn more about this item".

Detailed top reviews on the laptop from verified buyers have rated the Microsoft Surface 7 with five stars, with particular praise for the battery life. However, a common complaint is that "a lot of programs didn't work with Arm" which is something we've said in our own coverage.

In our five-star review, we praised the incredible performance, excellent battery life, great display, and solid pricing, with the only minor negatives being a lacking port selection, lack of an OLED variant and slight issues with the touchpad's sensitivity. With that said, why are users returning it so frequently? Amazon has not issued any reason why.

Another reason the Microsoft Surface 7 could see so many returns is that people could trade the device in for a computer with a competitor's chip. That's because AMD has made waves with its Ryzen AI processors, such as the coveted Ryzen AI Max+ 395, which has just arrived in some of the best laptops and best ultrabooks on the market. Those wanting superior performance for general computing, AI workloads, and greater compatibility may return to x86 architecture.

We can also look at the pricing of the laptop compared to its competition. This listing is available on Amazon with an MSRP of $2,099.99. However, the machine has been discounted by 26% to $1,544.83 for some time. It's possible that those who purchased it for the sticker price could have buyer's remorse, especially when weighing up against similarly priced options like the Arm-based MacBook Pro M4 (starting at $1,599 for the 14-inch model) which came out after the Surface.

Consumers still don't know what they're getting themselves into

If you've kept a keen eye on the laptops over the past 12 months, you'll no doubt have heard the buzzword "AI PCs" when it comes to Snapdragon X Elite, Ryzen AI, and Lunar Lake laptops in 2024 and 2025. There has been a wealth of promise in advertising around all the new tools you can use through having a dedicated NPU, and how many TOPs of performance you can expect from Qualcomm and the competition in general.

However, to the average everyday computing consumer, there's been no solid throughline in what this tech does, and what it can offer you, especially in the jump from traditional x86 architecture and over to Arm-based silicon. Yes, the latter can have superior battery life and snappier performance in isolation, but it comes at the cost of compatibility.

Windows on Arm has come a long way in the last year as more devices (like the Microsoft Surface 7) adopt it over x86-based alternatives, but you're still not going to get to do everything that you could on that architecture. A lot of popular software is yet to offer an Arm version, and it's not been communicated to hardware manufacturers what's available and what isn't.

As such, users could buy the Microsoft Surface 7 expecting it to do everything their old laptop could, and then be disappointed when they run into problems. If you're asking people to spend $2,000, they will want everything to be perfect out of the box, and a common understanding of differing architectures (and compatibility overall) just isn't going to happen overnight.

You may also like...

• My days of forgetting to drink water are over thanks to this new app
• The dawn of PCIe 7.0 could mean faster SSDs for everyone... eventually
• Samsung's rumored smart specs could be coming out before the end of the year

• HP says it's now considering making a handheld gaming PC thanks to Valve's SteamOS
• SteamOS will soon be available on other handhelds like the Lenovo Legion Go S
• The reason why HP hasn't made one yet is because of Microsoft's Windows 11

Sales of handheld gaming PCs continue to grow year by year, with a steady stream of new releases from Lenovo, Asus, and MSI, but one PC manufacturer has been missing in action - and the reason why may not come as a huge surprise.

As reported by Notebookcheck (via an interview with XDA), HP is now considering making a handheld gaming PC powered by Valve's SteamOS, as its Senior Vice President and Division President of Gaming Solutions, Josephine Tan, confirmed. And if you were wondering why it took so long, it's because of Microsoft's Windows 11 operating system.

Tan suggested that she has struggled with using Windows 11 and further hinted at the operating system's lack of simplicity, especially for handheld owners when resuming games—a feature that devices like the Nintendo Switch and Steam Deck take advantage of. This meant there were no plans from HP for a handheld.

However, now that Valve's SteamOS will be available for other handhelds besides the Steam Deck (like the Lenovo Legion Go S), Tan made it clear she is open to the idea of making a handheld gaming PC. While SteamOS still has some disadvantages, notably its inability to launch games with anti-cheat software baked in, and no official support for features such as frame generation, the user experience is still arguably much better than when using Windows 11 on a handheld.

(Image credit: Future) Microsoft should take this as a heads up...

If it wasn't clear enough already, Windows 11 hasn't been one of Microsoft's best releases in recent years: I'd say it's one of the worst I've seen in terms of bugs and seemingly neverending issues (specifically with its 24H2 update), and for a handheld gaming PC enthusiast, it's just not up to the job.

Aside from 24H2's frustrating bugs, Windows 11 pales in comparison to SteamOS for handhelds - and while it's understandable that there’s no 'portable' mode for Windows 11, as Microsoft's focus is on desktop PCs and laptops, it's another reason why many are moving away from the OS.

It's gotten so bad that I've considered installing Bazzite (a SteamOS clone) on my main desktop PC - and testing it out on my Asus ROG Ally for months suggests that it might not be a bad idea. Ultimately, this should be a wake-up call for Microsoft. Windows 11, besides its current 24H2 issues, just doesn’t work for handhelds, and with the popularity of handheld gaming PCs showing no sign of slowing down, now would be the perfect time for Microsoft to look into making its operating system work better on these devices - otherwise, it could really start to miss out.

You may also like...

• Playing Assassin's Creed Shadows at 8K with the Nvidia RTX 5090 was a nightmare - until I turned on Multi Frame Gen
• Finally, the more powerful Lenovo Legion Go S model has a release date - but the price is a gut punch
• Intel makes its XeSS 2 upscaling tech easier to add to games - here's why that's good news for PC gamers with tight budgets

• A threat actor is offering a database for sale, alleging it came from Oracle
• The archive contains encrypted SSO passwords and more
• Oracle denied being breached or losing data

Oracle has denied suffering a cyberattack and a data breach, following claims from a hacker to have stolen millions of records from company servers.

In mid-March 2025, a threat actor with the alias rose87168 released 6 million data records, claiming they were seized from Oracle’s Cloud federated SSO login servers. The archive posted on the dark web included a sample database, LDAP information, and a list of companies.

Perhaps unsurprisingly, Oracle was having none of it, issuing a statement declaring, "There has been no breach of Oracle Cloud. The published credentials are not for the Oracle Cloud. No Oracle Cloud customers experienced a breach or lost any data."

Encrypted SSO passwords

In the meantime, rose87168 took the archive for sale, in exchange for either an undisclosed sum of money, or zero-day exploits.

The threat actor claims the data includes encrypted SSO passwords, Java keystore (JKS) files, key files, enterprise manager JPS keys, and more.

"The SSO passwords are encrypted, they can be decrypted with the available files. also LDAP hashed password can be cracked," rose87168 said.

"I'll list the domains of all the companies in this leak. Companies can pay a specific amount to remove their employees' information from the list before it's sold."

Before listing the stolen archive for sale, the threat actor apparently asked Oracle for 100,000 XMR (the Monero cryptocurrency), but the company also demanded “all information needed for fix and patch,” and since rose87168 did not provide, the negotiations broke down..

To prove the stolen files were legitimate, the threat actor gave BleepingComputer a URL for Internet Archive, which shows that they uploaded a .txt file containing their email address to login.us2.oraclecloud.com server.

The publication reached out to Oracle for an explanation - we have also contacted the company for comment.

You might also like

• Several ServiceNow flaws are reportedly being linked together to attack companies and organizations
• We've rounded up the best password managers
• Take a look at our guide to the best authenticator app

• A fresh-faced Elon Musk revealed the new Roadster back in 2017
• Tesla’s website still has a page dedicated to Roadster reservations
• Does the promise of rockets and an ability to fly still stand?

Despite an original unveil in 2017 and suggested on-sale date of 2020, Tesla is yet to deliver any further information on its long overdue Roadster 2 model - one that was slated to obliterate the 0-60mph sprint in 1.9 seconds and offer a 620-mile range from a single charge.

Soon after its glitzy reveal eight years ago, the Tesla website began taking “fully refundable” reservations for a $200,000 ‘regular’ model or a hefty $250,000 deposit for one of 1,000 limited edition ‘Founder Series’ variant, as it did with the Cybertruck.

There is no official word on how much Tesla has made so far on Roadster 2 deposits, but it will continue to take your money, despite no update from Elon Musk on a production date, full specification, final design or a final price, as pointed out by CarScoops.

If you’re mad enough, you can head to the Tesla US website, where you will be asked to hand over $5,000 of your hard earned cash to simply to join the reservation list, where a further $45,000 will be required within 10 days to secure a build slot.

Remember, there’s no word on what the final cost will be if the new Roadster ever does materialize.

But back in 2017, Tesla fanatics went crazy for Musk’s wild promises of fantastical acceleration times, frankly ridiculous range targets and a 250+mph top speed.

Serial Tesla Tweeter, Sawyer Merritt, even suggested the marque was targeting a sub-1 second 0-60mph sprint time, with Elon Musk apparently backing up those claims on X - despite not explaining how the physics would actually work.

What’s more, the divisive CEO then went on to explain that the Roadster 2 would feature 10 small rocket thrusters to help it achieve those eye-watering acceleration times and even make it fly… yes, fly.

Analysis: Tesla customers are willing to pay for promises

(Image credit: Tesla)

We all know that Tesla is the undisputed king of wild promises, but Roadster 2 might go down in history as one of the most far-fetched. Eight years later and we are still no clearer on whether or not it was all just a lucid dream.

But for some Tesla fans, it has proven to be a nightmare, with message boards containing desperate pleas for help regarding potential legal procedures against the company, as they hold out to be among the first Roadster 2 customers.

No matter your stance on Tesla as a business, eight years is a frankly ridiculous amount of time to be waiting for a product, especially one that has cost some folk $250,000.

As one Reddit commentator pointed out, if those same customers ploughed their Roadster deposits into Tesla stock at its $21 value back in 2017, it would now be worth around $3 million. A far worthier investment.

But this isn’t the only issue that has irked Tesla customers, as Tesla’s Full Self-Driving package was originally sold as genuine “hands-off, eyes-off” autonomous technology, despite no legal framework being place to allow such a thing.

Tesla quickly had to adapt the wording and refer to its system as Full Self-Driving (Supervised), despite continuing to make customers pay for the technology. Its website makes it very clear that it won’t be offering refunds.

It begs the question, how much longer are Tesla's biggest advocates willing to bankroll the company's promised innovations? Especially if it continues to fail to deliver on them.

You might also like

• Tesla says it will refresh the Model S and Model X this year – here are 5 things I want to see
• I’ve spent hours with the new Tesla Model Y – here are 5 updates I like (and 3 things I don't)
• I’ve driven the Tesla Model 3, but Elon Musk is making it hard for us all to love the brand

• Potential 24 April 2025 release date for DJI's new flagship consumer drone
• Features to include digital ND filters and 52-minute battery life
• Pricing may start from $2,250

We’ve been eagerly awaiting official news on the DJI Mavic 4 Pro for a while but, if a new leak is true, we won’t have to hold our breath for much longer. The leak also revealed some exciting news about the premium drone’s feature list, not to mention pricing.

The news comes from serial leaker Jasper Ellens, who runs the site DroneXL and is usually highly accurate when it comes to reveals on DJI. In a story posted on the site, he claims to have been given the information from two separate sources, both of which have never been wrong in the past – but as always we suggest you treat these rumors with a pinch of salt, just as with any other pre-release leaks on tech gear.

Ellens claims the DJI Mavic 4 Pro will be launched on Thursday 24 April 2025, just over four weeks from now, and that DJI would be teasing the announcement via an official video on 17 April.

Today two little birds told me the official #launch date, Confirmed 7 key #features and on top of that one even told me the official #prices of the #DJIMavic4PRO. Let's dive into them! Cheers Jasperhttps://t.co/a8ccT81J9yMarch 19, 2025

Are these the Mavic 4 Pro's top features?

Ellens’ sources also spilled the beans on the Mavic 4 Pro’s feature list. They claim the drone will, like its predecessor the Mavic 3 Pro, come with three cameras, with focal lengths equivalent to 28mm, 70mm and 168mm. The sensor size will be larger than on the Mavic 3 Pro, however, offering video recording of up to 6K resolution.

The camera gimbal has been fully redesigned, allowing for much greater range of movement, while ND filters will be digital and built into the camera rather than physical attachments sold as optional extras. These are two features that seem likely to be very popular with filmmakers, as they greatly enhance creative options.

Battery life will be 52 minutes, as previously rumored, while 240W fast charging will allow a user to fully recharge three batteries in just 90 minutes. Finally, DJI will launch an all-new remote controller alongside the drone. Called the RC Pro 2, this controller will come with a larger 7-inch touchscreen that can be tilted.

Pricing revealed – but where's LiDAR?

In terms of pricing, it appears there will be three separate bundles to buy. Ellens’ sources claim the DJI Mavic 4 Pro (with a RC 2 controller) will cost $2,250 while a DJI Mavic 4 Pro Fly More Combo (which will likely feature two extra batteries, a charging hub, an RC 2 controller and a carrying bag) will cost $3,200. Finally, he says there’ll be a DJI Mavic 4 Pro 512GB Creator Combo, which will seemingly include all of the above but swap out the RC 2 for the RC Pro 2 controller, and cost $4,400.

All of the above sounds completely plausible to us – although we do think it's interesting that there's no mention of the much-rumored LiDAR module – and if Ellens is correct we’ll doubtless start to see a drip feed of further leaks (and potential official teases from DJI) in the run-up to that 24 April date. So we’ll likely have more rumors on DJI’s upcoming flagship camera drone to share with you very soon.

You might also like

• Camera rumors for 2025: the new gear we're expecting soon from DJI, Sony, Canon and more
• Zero Zero Robotics unveils world’s first sub-250g bi-copter drone with 'lightning fast acceleration' and 'unmatched agility'
• The best drone 2025: top flying cameras for all budgets

• De'Longhi has launched a new manual espresso machine, the Linea Classic
• It has everything you'd expect from an entry-level machine at a budget price
• It's out now in the US; we're still waiting for international prices and dates

De'Longhi has launched a new no-frills espresso machine that's compact, good-looking, and has a remarkably affordable price tag.

The De'Longhi Linea Classic is available now from retailers including Amazon and Crate & Barrel for an early-bird price of $179.95 (about £140 / AU$290). The regular price will be $199.95 (about £150 / AU$320), but De'Longhi has yet to announce international pricing and availability.

Most of the best coffee makers I've tested this year have been automated bean-to-cup machines with extras like automatic milk frothing and cold brewing – and price tags that many coffee-lovers will find prohibitive. It's therefore refreshing to see an espresso machine that just focuses on doing the basics well, at a very reasonable price.

The De'Longhi Linea Classic has all the features you'd expect from an entry-level espresso machine, including a 15-bar pump (with a stainless steel pressure gauge so you can watch the profile as your espresso is extracted), adjustable-height drip tray to accommodate different cups, and an adjustable steam wand.

(Image credit: De'Longhi) Everything you need, nothing you don't

It's a compact machine too, measuring just 12 x 8.89 x 11.14 inches / 30.5 x 22.6 x 29cm (H x W x D). That's small enough to tuck away on even a tiny kitchen counter – and because its case is stainless steel rather than plastic, it should look and feel a lot more expensive than it actually is.

It certainly sounds appealing, and if it delivers on its promises, it might even match the superb Breville Bambino as the best compact espresso machine around. It's a tall order – the Bambino is the entry-level machine recommended by many professionals – but I've got high hopes.

We should be getting our hands on the De'Longhi Linea Classic for testing soon, and we'll bring you a full review once we've had a chance to put it through its paces.

You might also like

• I wasn't a fan of syrup in coffee until I tried the Jura J10 espresso machine – here's why
• I've tested a lot of coffee machines, and the Siemens EQ900 Plus makes the best macchiato I've tasted
• I tried Nespresso's energy-boosting coffee capsules, and I have a new favorite morning brew

• Over 12,000 Bella Pro Series and Cooks coffee machines have been recalled
• Several users reported injuries after their machines failed at high pressure
• Owners should stop using the machines immediately and request a refund

Thousands of espresso machines have been recalled after 18 people reported a component being "forcefully ejected" during use. Of those, eight received burns or lacerations as a result.

The affected machines are no longer on sale, but several owners wrote reviews to warn potential customers about the risks before the voluntary recall. Some even included pictures of broken coffee makers in their kitchens.

"The machine blew up on the very first time used," wrote one owner. "Glass and metal flew all over the kitchen. I was getting milk from the refrigerator and that is why I was not hurt."

"This product is dangerous," said another. "It explodes while brewing. The portafilter literally detached itself from the machine during the brewing cycle. I have serious burns [due] to [...] the defect of this machine. Do not buy this machine. It is hazardous."

"The machine blew up"

The problem affects two coffee machines: the Bella Pro Series Steam Espresso Maker and the Cooks Steam Espresso Maker. The machines were imported by Sensio Inc. of New York, and 12,300 were sold at JC Penney and Best Buy – in stores and online – between August 2023 and January 2025.

All the recalled machines are made from black plastic with stainless steel trim and have either “Bella Pro Series” or "Cooks" on the front. Turn your machine over and you should see a sticker bearing the model number. If this is 90195 or 22395, you should stop using it immediately according to guidance from the US Consumer Product Safety Commission (CPSC).

To request a refund, you can contact Sensio toll-free at 855-647-3125 between 8 am and 5 pm ET Monday through Friday, or visit Bella Kitchenware's website. You will need to damage the UL certification number under the machine, cut its cord, and provide photos as evidence that you've made it unusable.

You might also like

• 3 coffee maker mistakes that are ruining your morning brew, according to an expert
• Are you one of the 39% of coffee drinkers who still use instant? Upgrading to one of these three cheap coffee machines will deliver a better-tasting brew
• 5 coffee innovations from CES – from condensing espresso machines to super-fast cold brew

• Samsung promises free repairs for everybody
• You're probably going to have to send your unit away for a while
• Some support techs don't seem to have been informed

As we reported last week, several Samsung soundbars including the flagship Samsung HW-Q990D have been rendered inoperable by a faulty firmware update. The firm promised to investigate, and now there's a free fix – although it's not ideal.

In a statement sent to Digital Trends, Samsung said:

“Samsung Electronics has identified a software update error as the cause of the operational issue affecting certain 2024 soundbar devices. We are taking immediate action to resolve the situation. Samsung is offering free repairs for all affected units—regardless of warranty status.”

Unfortunately that's likely to require you to send your soundbar away for a while – hardly convenient for a hefty soundbar at the best of times, and maybe doubly so if you didn't keep the box. And that's not to mention that your TV audio will suffer during that time…

It also seems that Samsung hasn't told all its support techs yet, based on reports online.

Samsung HW-Q990D: affected users have been promised free repair irrespective of warranty status. (Image credit: Future) What are Samsung support techs saying?

Some Samsung owners have turned to Reddit to discuss the "debacle", with some saying that support techs have been problematic: TigerHawk7 was "offered free parts but [told to] pay shipping and labor and that was after arguing after quite some time."

Over on the Samsung community forums, UK user ApocalypseRogue7122 – whose experience with support so far has been very frustrating, with Samsung initially refusing to take the soundbar back – reports getting a call from Samsung's authorized UK repairer: "The affected soundbars require a 'new part' (motherboard) and due to the number of failures Samsung is out of stock of said part. He told me that they (Mint Group FX) have ordered the part and if they are unable to get it by next Wednesday 25th March, then I will either receive a refund or replacement unit."

Hopefully Samsung is on top of this, because it's bad enough having to send back a soundbar without then having to wait an uncertain amount of time for a repair or replacement.

We've asked Samsung what customers should do if they encounter problems getting a free repair, and we'll update you when the company responds.

We rate Samsung's options highly among the best soundbars, and while it sounds like the update danger has now passed, Samsung's response to the problem might strongly affect how well people trust the company in the future. You only need to look at Sonos' nearly year-long problems to know that trust can be broken quickly and slow to regain.

You might also like

• The best Dolby Atmos soundbars you can buy
• 5 cheap soundbars we highly recommend
• Setting up the Sonos Arc Ultra was an awful app experience, but the sound was so worth it

• Employees of Minecraft developer Mojang have spoken on the role of generative AI in game development at a recent event
• They didn't seem too thrilled with the technology and stated that Minecraft required a human touch
• It comes as other Microsoft studios begin to use generative AI more heavily

It seems like you don't need to worry about generative AI (artificial intelligence) worming its way into the hit open-world survival game Minecraft.

According to IGN, Minecraft Vanilla game director Agnes Larsson discussed the issue of AI in game development at a recent event and wasn't too enthusiastic about the prospect. “Here for us, just like Minecraft is about creativity and creating,” they said.

“I think it's important that it makes us feel happy to create as humans. That's a purpose, [that] makes life look beautiful. So for us, we really want it to be our teams that make our games.”

Minecraft Vanilla executive producer Ingela Garneij added that the "specific touch of: what is Minecraft, how does it look?" is an "extra quality" that "is really tricky to create through AI." They even said that the developer has struggled to work with remote teams and emphasized the importance of human, face-to-face interactions.

"I mean creativity is... you need to meet like this as a person as a human to really truly understand the values and principles and the ecosystem," Garneij continued. "The lore, everything - it's so massive. Minecraft, it's a planet, it's massive."

Minecraft developer Mojang is a part of Xbox Game Studios and is ultimately owned by Microsoft. Other Xbox Game Studios developers don't appear to share the same reservations regarding generative AI. Call of Duty: Black Ops 6, for example, is filled with assets that were created with generative AI tools - a decision that caused some backlash within the game's community.

Earlier this year, Microsoft revealed a "breakthrough" generative AI model that could be used to "create consistent and diverse gameplay." Known as Project Muse, Xbox head Phil Spencer indicated that it could be used to help preserve old games.

Although generative AI seems exciting and is becoming an increasingly large part of our lives, it poses a number of ethical and environmental concerns that seem tricky to resolve. The decision to keep it away from Minecraft seems to go hand-in-hand with the game's focus on exploring a distinctly human kind of creativity.

You might also like...

• McDonald's reveals A Minecraft Movie meal with a bizarre set of collectibles and the most sinister sounding sauce ever
• Competitive shooter Fragpunk wowed me with its game-changing Shard Cards, but I can't stand the aggressive monetization
• I tried 15 hosts for streaming and hosting Minecraft games and these are the best