News
As of March 2025, 40% of global companies report using artificial intelligence (AI) in their business. While the benefits offered by this transformational tool can feel nearly limitless, the reality is that AI isn’t inherently secure, especially for companies dealing with sensitive information.
AI quickly analyzes vast amounts of data to figure out patterns and provide users with a response in the shortest amount of time possible. Any data shared with the tool will be used to train the model going forward, making it a dangerous place for sensitive information. For industries that handle extremely personal data, like healthcare or law, using AI could risk client privacy.
AI is designed to quickly analyze large datasets, detect patterns, and respond in real time. But many tools train on whatever data you provide. That means sharing private information—intentionally or not—can create long-term risks, especially in regulated industries like healthcare, finance, or law.
The benefits of leveraging synthetic dataAI works best with strong, structured, and relevant data. Whenever possible, real-world data is ideal—but that’s not always an option. Regulations like HIPAA and GDPR prevent teams from sharing personal data externally, including with AI models. That’s where synthetic data shines.
You’ll often see synthetic data used as a placeholder—especially when legal approvals or NDAs are still in progress. Instead of stalling development, teams can keep moving forward with stand-in data, then switch to production data later to validate the results. This keeps projects moving while staying compliant.
In other cases, synthetic data fills in the gaps. You might have real data, but not enough of it—or not enough variation to properly train your model. A good rule of thumb: you’ll need 10x more data samples than model parameters. When real data falls short, synthetic data can help augment and diversify your training set.
Considerations for using synthetic dataOne common misconception is that synthetic data is just “fake” data. But in reality, it's often based on real-world information that’s been restructured, anonymized, or generated to mirror actual scenarios. Think of it like a flight simulator—useful for training and preparation, but it’s not the same as flying a real plane. Synthetic data can help teams test and train AI models, but it shouldn’t be seen as a complete replacement for production data.
That said, it does come with risks—particularly around re-identification. If synthetic data can be traced back to the original source, the whole premise of privacy falls apart. One of the most critical steps is to ensure the original dataset is no longer stored or accessible once the synthetic version is created. Simply having the two datasets in proximity to each other creates unnecessary risk.
Another challenge is outliers. These are extreme or unusual values that can not only skew model training but also serve as clues about the original data. For example, if you're generating synthetic banking data and one of the transactions is for $10 million while the rest are in the hundreds, that single value becomes a beacon. It’s both a modeling issue and a potential privacy concern.
In many cases, partially synthetic data can offer the best of both worlds. You might use real documents or datasets while anonymizing any personally identifiable information. For example, you could keep the visual data from an X-ray but strip out details like the patient’s name, the facility, or the diagnosis.
That way, you retain data complexity without exposing sensitive information. Finally, before using any synthetic dataset in a project, it’s worth having someone outside the core team take a final look. A fresh perspective can help spot anything you’ve missed—whether it’s residual identifiers, overlooked outliers, or subtle signs that the data could still be traced back to a real person.
ConclusionUsing synthetic data doesn’t have to be all or nothing. Many projects benefit from a hybrid approach—especially in early phases. In a world racing to adopt AI, it’s easy to move fast and overlook the risks. But safe, responsible model training is everyone’s responsibility.
Synthetic data isn’t just a workaround—it’s a bridge to building secure, innovative systems that respect privacy and compliance from day one.
We've featured the best Large Learning Model.
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
- Phishing emails with malicious URLs are used four times more than ones with attachments, Proofpoint survey claims
- ClickFix attacks also spiked 400% year-on-year
- A layered approach to security is the best way to defend
Phishing scams and malicious URLs continue being the bane of the business world, increasing year-on-year, and getting more dangerous by the minute, new research has warned.
A new paper from Proofpoint, based on data from the company’s threat intelligence platform, argues phishing emails with URLs, rather than attachments, are rising in popularity; that ClickFix is currently the number one method of tricking victims into getting infected; and that most criminals are interested in stealing login credentials.
Phishing emails have always been the number one initial attack vector, for their simplicity, low cost, and omni prevalence. However, delivering malware via attachments is not that straightforward any more, with different email security solutions getting rather good at scanning and filtering malicious content.
ClickFix, QR codes, and SMS messagesThe cybercriminal community responded by pivoting to URLs - these days, they are used four times more than attachments. This is because they are easier to disguise, and more likely to evade detection, Proofpoint argues. The miscreants would embed them in messages, buttons, even inside benign attachments such as PDFs or Word documents.
In many cases, the URLs lead to sites with a ClickFix popup. ClickFix is a phishing technique where the victims are shown a fake error, and are given the means to “fix” the problem immediately. These attacks, too, increased by nearly four times year-over-year.
Proofpoint also said that most threat actors are interested in stealing logins, as it spotted 3.7 billion URL-based attacks aimed at stealing such secrets. This is mostly because infostealing-malware such as CoGUI or Darcula are low-skill phishing kits that can be easily obtained and deployed.
Other notable methods include QR code phishing threats (quishing), and SMS phishing (smishing), with the latter spiking 2,534% year-on-year.
“The most damaging cyber threats today don’t target machines or systems. They target people. In addition, URL-based phishing threats are no longer confined to the inbox, they can be carried out anywhere and are often extremely difficult for people to identify,” said Selena Larson, senior threat intelligence analyst at Proofpoint.
“From QR codes in emails and fake CAPTCHA pages to mobile-first smishing scams, attackers are weaponizing trusted platforms and familiar experiences to exploit human psychology. Defending against these threats requires multilayered, AI-powered detection and a human-centric security strategy.”
How to defend against phishingThe best defense against phishing emails with malicious URLs is layered protection.
Businesses can start with an email security gateway that blocks suspicious links even before they can reach the inbox. Then, with browser isolation, or link writing, systems can “detonate” (trigger, basically), URLs in a safe environment.
Finally, every business should train their employees on how to spot phishing emails, how to hover over links to double-check where they lead to, how to verify senders, and avoid clicking unexpected messages. Finally, enforcing multi-factor authentication (MFA) is always recommended, as is keeping endpoint protection updated to catch malware if someone does click.
Finally, businesses should implement strict access controls and monitoring so that even if a link slips through, the damage remains contained.
You might also like- How to spot a phishing email
- Take a look at our guide to the best authenticator app
- We've rounded up the best password managers
- There are reports of a bug with the August update for Windows 11
- It can seemingly cause SSDs to fail in a small number of reported cases
- We can't jump to conclusions that this is Microsoft's fault yet - but this is a worrying-sounding situation that bears monitoring
Windows 11 may have run into trouble with its latest update possibly breaking some SSDs, but at this point, we need to be very cautious, as there are just scattered reports of this issue.
This comes from Wccftech, which noticed a post from a Japanese Windows 11 user (Necoru_cat) on X flagging the problem.
To summarize the reported bug, we're told that it's caused by KB5063878 (the Windows 11 update for August, which has been out for a week), and that an affected SSD will disappear completely from Microsoft's OS.
This "occurs during heavy write operations," apparently, meaning a large installation or patch, with the report suggesting that the bug can be triggered by 50GB of continuous write activity on the drive. Necoru_cat encountered the issue when applying a big patch for Cyberpunk 2077.
The X user points the finger of suspicion at this being a problem with the drive cache, and has seemingly carried out a bunch of testing, which suggests that SSDs with Phison NAND controllers may be more prone to this bug.
As mentioned, if an SSD is affected by this glitch, it's no longer recognized by Windows 11, and the drive partition disappears. Worryingly, we are told that if this happens, there is a "high likelihood of file corruption," but again, season that - and this whole report - liberally.
(Image credit: Shutterstock)Analysis: don't be alarmed - yet - but do be cautiousAs I've already indicated, we need to be very careful jumping to conclusions here. Is this really a problem that can be blamed on the latest Windows 11 update? We don't know that, and we certainly can't go by a single report of the bug (albeit it was seemingly a comprehensive dive into the glitch).
However, I've been doing some searching for other reports of this problem, and I've turned up some bewildered folks who've seemingly run into it. There are multiple people in this Reddit thread who say they have been affected by this bug, including a report of an Adata SP580 SSD going down in flames.
As far as affected models go, according to a list from NichePCGamer that Wccftech pointed out, as well as SSDs with the Phison PS5012-E12 controller, the following models could be affected: Corsair Force MP600, Kioxia Exceria Plus G4, Fikwot FN955, and the SanDisk Extreme Pro M.2 3D SSD - so we can add the Adata SP580 to that list.
There may be "similar symptoms" with hard drives as well, we're told, but all reports seem to be SSD-related as far as I can see.
At any rate, while this still appears to be a niche problem, there's enough here that my bug senses are tingling, as it were. And given how nasty this glitch sounds, I think it's worth making people aware of it. Indeed, I think it's a wise precaution to perhaps not perform any big installations or patches for the time being, if you're running the latest Windows 11 update - if you can avoid it – at least until we hear more about what might be going on here, or some official word from Microsoft.
Some have raised the theory that, rather than an actual flaw in the Windows 11 update for August, it's a case of what the update does - in terms of its writing activity to the system drive - that triggers a bug in some SSDs (Western Digital models to be precise). But that doesn't explain why different SSD types would be affected (or indeed hard drives, apparently).
We need to hear more from Microsoft on this, and hopefully, the company will be on the case and investigating these reports. I've contacted Microsoft about this bug, and I'll update this story if I hear back.
You might also like- Stream select The Hundred fixtures free on BBC iPlayer (UK)
- Stream from abroad with NordVPN (save 76% today's deal)
- Today's Games: Southern Brave vs Oval Invicibles
The 2025 season sees Jimmy Anderson participate in The Hundred for the first time, as the jewel in the Manchester Originals setup. The ECB's 100-ball domestic competition is now in its fifth season, and rumor has it that drastic changes are in the offing.
Heather Knight led London Spirit to a maiden women's title last season, while the Oval Invincibles retained the men's championship under Sam Billings. Nat Sciver-Brunt (Trent Rockets) and James Vince (Southern Brave) topped the run-scoring charts last time out, while Linsey Smith (Northern Superchargers) and Tymal Mills (Southern Brave) took the most wickets.
It's going to be slightly surreal seeing Anderson, England's record Test wicket-taker, in amongst the flashing stumps and psychotropic kits. He, along with every other bowler, can deliver a maximum of 20 per match, and up to 10 in a row. The other figure gernering huge interest is 17-year-old Rocky Flintoff, who's playing under his father's stewardship at the Northern Superchargers.
Big Money has officially got its claws into The Hundred, with Todd Boehly, Tom Brady, the Ambani family, Palo Alto Networks CEO Nikesh Arora amongst the latest investors. With a view to maximizing profits, they're looking to expand The Hundred to 10 teams or more, and potentially even switch to the Twenty20 format.
Read on below for a complete guide to where to watch The Hundred 2025 live streams online, on TV and potentially for free.
Can I stream The Hundred 2025 for free?In total, 16 games are being shown on free-to-air BBC Two in the UK, with live streaming available via BBC iPlayer.
Use a VPN to watch The Hundred 2025 from abroadYou can watch The Hundred 2025 from abroad at no extra cost with NordVPN, which we tested in-depth (see our best VPNs round-up).
Get a VPN, connect to your usual streaming service from anywhere, and watch the cricket live on your screen. Easy!
Resorting to illegal streams is a bad decision. If you're stuck abroad, consider a VPN to safely stream The Hundred via an official platform such as BBC iPlayer, Sling or Kayo Sports.
Editors ChoiceNordVPN deal: FREE $50 / £50 Amazon gift card
Boasting lightning fast speeds, great features, streaming power, and class-leading security, NordVPN is our #1 VPN.
✅ FREE Amazon gift card worth up to $50/£50
✅ 4 months extra FREE!
✅ 76% off usual price
Use Nord to unblock your streaming service of choice and watch The Hundred 2025 live online with our exclusive deal.VIEW DEAL ON
How to watch The Hundred 2025 live streams in the USWillow TV is showing every game of The Hundred in the US.
If you don't have it as part of your cable package, you can watch Willow through your choice of Sling TV's Desi Binge Plus or Dakshin Flex plans – starting from $10 per month.
Outside the US right now? You can make use of NordVPN to catch the action.
For US residents, Sling TV is the perfect plan in order to watch cricket from WillowTV. You can sign up to your choice of its Desi Binge Plus, Dakshin Flex or Urdu plans, which all include Willow.
Prices start from only $10 per month or $50 for six months.View Deal
How to watch The Hundred live streams in the UKEvery game of The Hundred is being shown on Sky Sports in the UK, but 16 of them will also be free-to-air on BBC Two and BBC iPlayer.
Sky Sports packages start from £22 per month. Or you can use a more flexible streaming option, Now (formerly Now TV). Now Sports memberships start at £14.99 for a day pass, or £34.99 monthly.
If you're on holiday outside the UK, you can use NordVPN to access Sky Sports' coverage. We used it during the recent England vs India Tests and it was brilliant.
How to watch The Hundred live streams in Australia(Image credit: free)Every game of The Hundred is on Foxtel in Australia.
However, your most cost-efficient option is specialist sports streamer Kayo Sports, with plans starting from $30 a month after you've taken advantage of its first month for $1 offer or 7-day free trial.
If you're currently out of Australia but want to watch a Hundred live stream, you'll need to get yourself a VPN, as per the instructions above.
How to watch Women's Open live streams in Canada (Image credit: Other)Willow TV is the only place to watch The Hundred in Canada.
Prices start at CA$8.99 monthly, but an annual subscription you can get for CA$79.99.
Outside Canada during the 2025 The Hundred season? Simply use a VPN to tune in from abroad.
The Hundred 2025 FAQsWho are The Hundred 2025 teams and captains?Women
Birmingham Phoenix – Ellyse Perry
London Spirit – Charlie Dean
Machester Originals – Beth Mooney
Northern Superchargers – Hollie Armitage
Oval Invincibles – Lauren Winfield-Hill
Southern Brave – Georgia Adams
Trent Rockets – Ashleigh Gardner
Welsh Fire – Tammy Beaumont
Men
Birmingham Phoenix – Liam Livingstone
London Spirit – Kane Williamson
Machester Originals – Phil Salt
Northern Superchargers – Harry Brook
Oval Invincibles – Sam Billings
Southern Brave – James Vince
Trent Rockets – David Willey
Welsh Fire – Tom Abell
(All times BST)
Group Stage
Thursday, August 7
3pm – Women’s: Northern Superchargers vs Welsh Fire
6.30pm – Men’s: Northern Superchargers vs Welsh Fire
Friday, August 8
3pm – Women’s: Birmingham Phoenix vs Trent Rockets
6.30pm – Men’s: Birmingham Phoenix vs Trent Rockets
Saturday, August 9
11am – Women’s: Oval Invincibles vs Manchester Originals
2.30pm – Men’s: Oval Invincibles vs Manchester Originals
2.30pm – Women’s: Welsh Fire vs London Spirit
6pm – Men’s: Welsh Fire vs London Spirit
Sunday, August 10
11am – Women’s: Southern Brave vs Birmingham Phoenix
2.30pm – Men’s: Southern Brave vs Birmingham Phoenix
2.30pm – Women’s: Trent Rockets vs Northern Superchargers
6pm – Men’s: Trent Rockets vs Northern Superchargers
Monday, August 11
3pm – Women’s: Manchester Originals vs London Spirit
6.30pm – Men’s: Manchester Originals vs London Spirit
Tuesday, August 12
3pm – Women’s: Birmingham Phoenix vs Oval Invincibles
6.30pm – Men’s: Birmingham Phoenix vs Oval Invincibles
Wednesday, August 13
10.30am – Women’s: Southern Brave vs Northern Superchargers
2pm – Women’s: Welsh Fire vs Manchester Originals
3pm – Men’s: Southern Brave vs Northern Superchargers
6.30pm – Men’s: Welsh Fire vs Manchester Originals
Thursday, August 14
3pm – Women’s: London Spirit vs Trent Rockets
6.30pm – Men’s: London Spirit vs Trent Rockets
Friday, August 15
3pm – Women’s: Northern Superchargers vs Birmingham Phoenix
6.30pm – Men’s: Northern Superchargers vs Birmingham Phoenix
Saturday, August 16
11am – Women’s: Trent Rockets vs Southern Brave
1.30pm – Women’s: Oval Invincibles vs Welsh Fire
2.30pm – Men’s: Trent Rockets vs Southern Brave
6pm – Men’s: Oval Invincibles vs Welsh Fire
Sunday, August 17
10am – Women’s: Manchester Originals vs Northern Superchargers
1.30pm – Women’s: Birmingham Phoenix vs London Spirit
2.30pm – Men’s: Manchester Originals vs Northern Superchargers
6pm– Men’s: Birmingham Phoenix vs London Spirit
Monday, August 18
3pm – Women’s: Southern Brave vs Oval Invincibles
6.30pm – Men’s: Southern Brave vs Oval Invincibles
Tuesday, August 19
3pm – Women’s: Trent Rockets vs Manchester Originals
6.30pm – Men’s: Trent Rockets vs Manchester Originals
Wednesday, August 20
11.30am – Women’s: Welsh Fire vs Southern Brave
2pm – Women’s: London Spirit vs Northern Superchargers
3pm – Men’s: Welsh Fire vs Southern Brave
6.30pm – Men’s: London Spirit vs Northern Superchargers
Thursday, August 21
2pm – Women’s: Oval Invincibles vs Trent Rockets
6.30pm – Men’s: Oval Invincibles vs Trent Rockets
Friday, August 22
2pm – Women’s: Birmingham Phoenix vs Welsh Fire
6.30pm – Men’s: Birmingham Phoenix vs Welsh Fire
Saturday, August 23
11am – Women’s: Northern Superchargers vs Oval Invincibles
1.30pm – Women’s: London Spirit vs Southern Brave
2.30pm – Men’s: Northern Superchargers vs Oval Invincibles
6pm – Men’s: London Spirit vs Southern Brave
Sunday, August 24
11am – Women’s: Welsh Fire vs Trent Rockets
1.30pm – Women’s: Manchester Originals vs Birmingham Phoenix
2.30pm – Men’s: Welsh Fire vs Trent Rockets
6pm – Men’s: Manchester Originals vs Birmingham Phoenix
Monday, August 25
3pm – Women’s: Oval Invincibles vs London Spirit
6.30pm – Men’s: Oval Invincibles vs London Spirit
Tuesday, August 26
2pm – Women’s: Northern Superchargers vs Manchester Originals
6.30pm – Men’s: Northern Superchargers vs Manchester Originals
Wednesday, August 27
2pm – Women’s: Trent Rockets vs Birmingham Phoenix
6.30pm – Men’s: Trent Rockets vs Birmingham Phoenix
Thursday, August 28
2pm – Women’s: Southern Brave vs Welsh Fire
6.30pm – Men’s: Southern Brave vs Welsh Fire
Eliminators
Saturday, August 30
2.15pm – Women’s Eliminator
6pm – Men’s Eliminator
Finals
Sunday, August 31
2.15pm – Women’s Final
6pm – Men’s Final
Of course! Each of the major broadcasters has streaming services that you can access through mobile apps or via your phone's browser.
You can also stay up-to-date with all of the key The Hundred moments on the official social media channels on Instagram (@thehundred) and YouTube (@thehundred). The Hundred website will also allow you to follow the latest leaderboards.
We test and review VPN services in the context of legal recreational uses. For example:1. Accessing a service from another country (subject to the terms and conditions of that service).2. Protecting your online security and strengthening your online privacy when abroad.We do not support or condone the illegal or malicious use of VPN services. Consuming pirated content that is paid-for is neither endorsed nor approved by Future Publishing.
- Amazon has released a bunch of official images for Fallout season 2
- The stills are our first look at the hit Prime Video show's next entry
- They come ahead of possible release date and trailer reveals at Gamescom 2025
Amazon has given us our first look at Fallout season 2 – and no, it's not released its inaugural trailer a day earlier than many expect.
The hit Prime Video TV Original is set to return to our screens before the end of 2025. And, to kick off its promotional push for the series' next outing, Amazon MGM Studios has unveiled seven official images to whet our appetite for Fallout's sophomore season.
A post shared by FALLOUT ⚡️ (@falloutonprime)
A photo posted by on
There's not a whole lot we can read into four of the images. However, the other three – pictures three, five, and six in the Instagram post above – will pique the interest of viewers and long-time fans of Bethesda's video game franchise namesake.
So, what can we read into this specific trio? Image five in the post above, which shows Moises Arias' Norm MacLean looking very stern as he stands opposite a bunch of other Vault Dwellers, is particularly intriguing. Indeed, last season's finale ended with Norm trapped in Vault 31 with no means of escape, nor any food or water to keep him alive. Unless this still is taken from a flashback sequence, it's clear that Norm finds a way out of Vault 31 in the Fallout TV show. We just don't know how yet.
As for the other two images, they show The Ghoul – albeit as Cooper Howard, aka his human form prior to The Great War – enjoying some time in New Vegas.
We'll get further insights into Cooper Howard's life before The Great War in season 2 (Image credit: Lorenzo Sisti/Prime Video)We already knew this location would feature heavily this season – indeed, it was teased as much in the season 1 finale and in the first poster for Fallout season 2. Nevertheless, this is the first confirmation we've received that New Vegas will be shown off in all of its glory by way of some pre-Great War flashbacks. Hopefully, we'll learn even more about Howard's backstory and, after last season's flashbacks, whether he tells anyone about what the nefarious Vault-Tec is actually up to.
Potentially, it could be a big week on the Fallout season 2 reveal front. These images and its first piece of artwork aside, I'm expecting an official launch date announcement to be made at Gamescom 2025. In May, Amazon confirmed Fallout season 2 would make its bow this December, but an actual date hasn't been publicly disclosed. Could we learn on during Opening Night Live tomorrow (August 19)? I certainly think so.
A trailer for one of the best Prime Video shows' second chapter could be unveiled, too. I have no insider knowledge on the matter, but it would be a bit strange if Amazon dropped this season's first poster and images, but made us wait for some actual footage. I'd be surprised if a full trailer is released this early on in its marketing campaign, but a short teaser wouldn't go amiss.
But, what do you think? Have I missed anything significant in these first-look images? And do you think we'll get an official launch date and/or trailer at Opening Night Live? Let me know in the comments.
You might also like- Peacemaker season 2 star Sol Rodríguez wants her character to 'have some fun' in another big DC comic book project, but she doesn't know if it'll happen
- Marvel Zombies won't make its Disney+ debut in October after all – here's why that's a good and bad thing
- Netflix releases the official trailer for Wednesday season 2 part 2 – and it's a shock-filled teaser worth dying for
- 115 inches with Samsung's Glare Free tech
- AI 4K and HDR upscaling and lots of picture optimization
- $26,999, available now
Samsung has unveiled its biggest ever Neo QLED TV, the QN90F – and if bigger means better, then this is its best QLED TV yet. At 115 inches it's designed to deliver what Samsung says is a "mind blowing" cinematic experience.
Big TVs are in big demand right now – see also, Skyworth's new 100-inch art TV – with sales of TVs 98 inches and bigger growing by 148% year on year. And with a recommended price tag of $26,999 (so around £19,940 or AU$41,449 as a guesstimate) the 115-inch QN90F is priced competitively for that market.
One of the selling points of this particular TV is that it's very bright compared to OLED, making it great for things like watching live sports in the daytime without having to make your room so dark you fall over your friends, family or housemates.
Samsung hasn't released the full details yet but as the owner of a much more modest Samsung Neo QLED I can confirm that these TVs are often so bright I feel I should be wearing sunscreen.
(Image credit: Samsung)Samsung QN90F: key featuresThe full spec sheet hasn't been revealed just yet, but we know that the QN90F has Samsung's Glare Free design and features the latest generation of the NQ8 AI Gen3 processor, which enables features such as the Supersize Picture Enhancer. That optimizes the display to ensure that, as Samsung puts it, "bigger doesn't mean blurrier" due to the larger size of giant screens' pixels.
The TV also features Samsung's Quantum Matrix Technology Pro for precise lighting control, and 4K Upscaling Pro, which does exactly what you'd expect with sub-4K resolutions. There's also Neo Quantum HDR Pro for real-time color optimization and Auto HDR Remastering Pro to upscale standard dynamic range media to "HDR-like quality".
Gaming and sports are incredible on enormous TVs, and here Samsung has delivered its Motion Xcelerator 144Hz for the former and AI Motion Enhancer Pro for the latter to keep things pin-sharp. And to soundtrack everything there are top-channel Atmos speakers and Samsung's clever Object Tracking Sound+, which delivers positional audio that follows the on-screen action.
The operating system is the familiar One UI Tizen, and that's guaranteed for seven years of OS updates to keep your TV apps and services current.
The QN90F is available in the US from today, August 18 2025, for $26,999. Official pricing and availability for other markets hasn't yet been released, but know this: we cannot wait to see it.
You might also likeEvery morning, I browse the internet looking for clever ways people are using AI to achieve their goals. While I often find impressive use cases like "This ChatGPT prompt ‘unlocks 100% of GPT-5’s power’", this morning I stumbled across someone using ChatGPT to summarize book chapters.
Now, summarization tools are nothing new; in fact, some of the best examples of AI are apps like NotebookLM, which condense information into smaller, easily digestible chunks.
In this case, however, this prompt is being used to summarize fictional works, removing the beauty of what makes literature so inherently romantic. At first, I thought I would scroll past the Reddit thread, but after sitting on it for a while, I decided it was only right that I voice my opinion.
I think AI, in specific circumstances, is an excellent tool that can truly improve our lives. However, this prompt is an example of everything wrong with AI, and I can't help but feel sad to see people championing it.
Don't summarize artI Used AI to Read Books Chapter by Chapter — Here’s Why I’ll Never Read the Same Way Again from r/ChatGPTPromptGeniusThe Reddit thread in question is titled, "I Used AI to Read Books Chapter by Chapter — Here’s Why I’ll Never Read the Same Way Again." Yeah, of course you won't be reading the same way again; you're no longer reading!
The user explains how they have been using an in-depth prompt to summarize books chapter by chapter, and honestly, it sounds like they're stripping pieces of literature of everything that makes them what they are in the first place.
I've not lost all hope in humanity, however, as the comments on the Reddit thread are pretty clear that this might just be a step too far. One user wrote, "This is one use of AI I CANNOT get behind, might as well put a Michelin star meal through a blender and eat it as a shake."
Another said, "Read the book… And if you can get the content and experience of the book with an AI summary of it, then it probably is not a book that you should be reading, in which case, just don’t read it or summarize it."
I see a place for summarization tools when it comes to simplifying non-fiction subjects like college textbooks, and I think that's a genuinely great way to get the most from AI. However, I fundamentally disagree with using AI to make creative work shorter and simpler for consumption.
Nowadays, every article you find online probably has some form of summary embedded at the top of the post, or you might even have an AI tool like Apple Intelligence to do the job for you. Even in these circumstances, I'm not a fan of the over-simplification of content, especially as we all continue to shorten our attention spans with the likes of quick-cut, short-form video content.
If you're relying on AI to summarize a book, you shouldn't be reading the book in the first place. Life is about curation, and what makes it so precious is the fact that we can't consume every single book, movie, or video game on the planet.
Use AI to help you make choices about what to consume next, but please, please don't use AI to do the consumption for you.
You might also like- ChatGPT-5 just completed Pokémon Red in a new world record time
- I used GPT-5 to code a Fantasy Premier League program, and I might actually stand a chance of winning my draft this year thanks to ChatGPT
- Sam Altman is ‘uneasy’ about the way people use ChatGPT, but I don’t think he has the right to cry about his success
- Cisco found a 10/10 flaw in Secure Firewall Management Center
- It released a patch and advised on possible mitigations
- No evidence of in-the-wild abuse so far, but users should still be on guard
Cisco recently fixed a maximum-severity vulnerability in its Secure Firewall Management Center (FMC) product, and urged users to apply either the patch, or the mitigation, as soon as possible.
FMC is a centralized platform for configuring, monitoring, and analyzing Cisco Secure Firewalls, where users can manage policies, track threat intelligence, and monitor their deployments across endpoints.
As per Cisco’s new security advisory, the vulnerability was discovered in the RADIUS subsystem implementation of FMC. RADIUS (Remote Authentication Dial-In User Service) is a protocol used to authenticate, authorize, and account for FMC administrators and VPN users by integrating with an external identity server.
Fixes and mitigationsThe flaw is described as an “improper handling of user input during the authentication bug” which could allow an unauthenticated remote attacker to inject arbitrary shell commands.
In theory, this could be done by sending crafted input when entering credentials - however the caveat here is that FMC must be configured for RADIUS authentication for the web-based management interface, SSH, or both, for the bug to be exploitable.
The bad news is that, according to BleepingComputer, this configuration is “commonly used” in enterprise and government networks where administrators want centralized login control and accounting for network device access. Therefore, the attack surface could be rather large, and the victims high in profile.
It is now tracked as CVE-2025-20265, and was given a severity score of 10/10 (critical).
Cisco released a patch to fix the issue, and said those that cannot apply it should disable RADIUS authentication and replace it with a different method, such as local user accounts, external LDAP, or similar. The company also said that the mitigations proved good in testing, but warned customers to run the tests themselves.
Via BleepingComputer
You might also like- Cisco warns of a serious security flaw in comms platform - and that it needs patching immediately
- Take a look at our guide to the best authenticator app
- We've rounded up the best password managers
- Two new leaks claim future Garmin watches will be getting LTE connectivity
- This feature allows you to add smartwatches to a data plan, allowing you to make calls and stream without a phone
- It's a feature included on Apple and Samsung watches, usually on more expensive models
Two new leaks claim we could soon be able to make calls, stream music and answer texts from some of the best Garmin watches yet to be announced – even without an attached phone.
Yes, LTE connectivity is coming to Garmin watches, according to separate reports from leak sites Garmin Rumors and The5Krunner. Garmin Rumors partnered with a site called AppSensa, which digs into code to discover in-development features. Garmin is reportedly going "all-in" into cellular connectivity, according to the site.
Listed features include emergency satellite SOS messaging, two-way messaging, LiveTrack location sharing without a phone, and voice calling with contacts.
Garmin Messenger, the watch-to-watch satellite messaging service, is also reportedly being expanded. Perhaps Garmin Messenger will now be Garmin's principle messaging client, able to text using your data plan and reaching phones as well as other Garmin watches?
The5KRunner also reports these claims, stating that the obvious starting point is a rumored Fenix 8 Pro, in which the "Pro" stands for the inclusion of LTE connectivity. The best Apple Watches and best Samsung watches often come in cheaper Bluetooth-only and more expensive LTE-capable variations, so it would make sense for Garmin to follow this pattern.
A big step forward(Image credit: Future)Garmins have long been considered separate entities to smartwatches, as they're more like adventure companions or training tools than everyday communications devices. However, certain ranges such as the Garmin Venu series already blur the lines a little.
Making the watches capable of no-phone communications would mean a huge change in strategy, and would enable more smartwatch-like models, such as the Garmin Venu X1 to seriously compete with the likes and the Apple Watch Ultra 2 and Samsung Galaxy Watch Ultra.
If true, I'm intrigued to see which future models get this rumored LTE option. The Garmin Venu 4 seems like a no-brainer. One commenter on our Garmin Venu 4 rumors hub agreed, stating: "If the Garmin Venu 4 ever came with LTE and text message response that would definitely be a gamechanger. So many current iPhone users would jump on the Venu 4 – it would be a dream watch."
Perhaps we'll see Forerunner Pros with LTE options in future too, as phoneless LiveTrack seems like a great safety tool for triathletes and trail runners doing solo training.
You might also like...Guillermo del Toro's Frankenstein will be coming to Netflix on November 7, with the streamer confirming the news alongside a first look at two of the new movie’s promotional posters (see below). We’ve already seen a few preview snaps of what’s to come as well as a two-minute trailer, showing Oscar Isaac as Victor Frankenstein, Mia Goth as Elizabeth Lavenza and Charles Dance as Leopold Frankenstein. However, Jacob Elordi’s monster largely still hides in the shadows, only appearing as a black silhouette shrouded with furs and claws as dark as night.
FRANKENSTEIN, a film by Guillermo del ToroStarring Oscar Isaac, Jacob Elordi, Mia Goth, and Christoph Waltz. Coming to select theaters October 17 and on Netflix November 7. pic.twitter.com/zdYFoq9cURAugust 18, 2025
It goes without saying that this is another adaptation of the famed novel by gothic author Mary Shelley, with Frankenstein’s monster (which isn’t called Frankenstein, though we often get that mixed up) appearing in countless movies since way back in 1915. This is neither Netflix’s first foray into bringing the tale to life (remember Creature?), nor is it del Toro’s first original creation for one of the best streaming services around, having previously made an animated version of Pinocchio for Netflix in 2022.
But a version of Frankenstein made by del Toro! What a world we get to live in. The upcoming move has been my most-anticipated film of 2025 since last year, and thankfully, there’s a way I can see it before it hits Netflix in November (where I’ll watch it again, obviously).
Netflix quietly confirms October 17 theatrical release date for Del Toro’s Frankenstein before it drops on streamingIf you look closely at the above posters, you’ll see that Frankenstein is showing in select theaters from October 17. While that doesn’t mean everyone will have direct access, it does mean global hotspots will get the movie early (think New York, Los Angeles and London). Frankly, it’s a clever move, and addresses the one piece of feedback Netflix has received time and time again over the years: give your original movies a theatrical release, too.
Think of KPop Demon Hunters, the monster smash hit (pun intended) nobody expected is now playing catch-up with singalong screenings after the fact. On top of this, said screenings are typically only playing for two days in very limited locations, which basically defeats the point of letting fans enjoy the experience in the first place. Netflix has to do theatrical screenings before a streaming release for the strategy to work, for both struggling cinemas and the platform itself.
Given the two minutes of footage we’ve seen tells us as much information as a blank piece of paper, anybody seeing Frankenstein in cinemas first is essentially walking in blind. Sure, we’ll have likely had some reviews before then, but these showings will drum up the hype and engagement would Netflix chew its own hand off for, in the most organic way possible (by word of mouth from audiences). By the time Frankenstein then hits Netflix, it’ll be an explosion of views, with people likely racing to be able to see it first. It’s a marketing masterstroke, and a generous one too.
Speaking about the film, del Toro said in a statement: “The most important figure from English legacy is, incredibly, for me, a teenager by the name of Mary Shelley, and she has remained a figure as important in my life as if she were family. And so many times when I want to give up, when I think about giving up, when people tell me that dreaming of the movies and the stories I dream are impossible, I think of her.” Thank you Shelley, and del Toro, for your service.
You might also like- Night Always Comes has just hit Netflix, and Julia Fox is unironically the best part of Vanessa Kirby’s complex caper
- Netflix releases the official trailer for Wednesday season 2 part 2 – and it's a shock-filled teaser worth dying for
- Monster season 4: everything we know so far about the hit show's return to Netflix
- Four in five companies knowingly ship vulnerable code, survey warns
- One-third say 60% of their code is now AI-generated
- Orgs need to use AI to identify vulnerabilities
A study of 1,500 CISOs, AppSec Managers and developers conducted by Checkmarx has claimed four in five (81%) companies knowingly ship vulnerable code, putting them and their users at risk of attack.
An estimated one in two respondents already use AI security code assistance, with around one-third (34%) admitting that more than 60% of their code is AI-generated – which can often contain known vulnerabilities by default.
An overwhelming majority (98%) have experienced a breach due to vulnerable code in the past year, and yet they continue to ship vulnerable code without implementing the right protective measures.
Companies are shipping vulnerable, AI-generated codeThe report outlines how generative AI has now eroded developer ownership with code less likely to be affiliated with any particular individuals. It has also expanded the attack surface by reopening vulnerabilities that could previously have been avoided with proper coding expertise.
The trend has largely been blamed on artificial intelligence, with vibe coding on the rise and many developers now opting to edit AI-generated code rather than write their own from the ground up.
The lack of governance around this has created what the company describes as the perfect storm.
Fewer than half of the respondents were found to be using foundational security tools like DAST and IaC scanning, with a similar number using DevSecOps tools.
Looking ahead, Checkmarx stresses security should be built into projects right from coding level, with organizations urged to establish policies for AI tool usage. Acknowledging that developers are now actively using AI, Checkmarx suggests that, instead of banning it, companies should also utilize agentic AI to analyze and fix issues across projects.
"AI generated code will continue to proliferate; secure software will be the competitive differentiator in the coming years," Checkmarx VP of Portfolio Marketing Eran Kinsbruner concluded.
You might also like- We've listed the best IDEs for Python
- Fancy an upgrade? Check out the best laptops for programming
- Google issues official internal guidance on using AI for coding - and its devs might not be best pleased
- Researchers uncover two packages carrying an infostealer
- The victims are apparently Russian, and attackers American
- This prompted the researchers to speculate if the targets were Russian crypto hackers
Two malicious packages were recently discovered on the npm package manager platform targeting software developers on the Solana ecosystem.
However the discovery, attribution, and potential targets of the malware have made researchers speculate if this was a state-sponsored attack.
Solana is a blockchain designed for decentralized applications and cryptocurrencies. It is similar to Ethereum in many aspects, which is why it is often described in the crypto community as the “Ethereum killer”.
Targeting devs? Or hackers? Or both?Recently, security researchers from Safety found two npm packages: “solana-pump-test” and “solana-spl-sdk”.
Both were submitted by the same author, and both contained identical code - and according to Safety, when these packages were installed, they ran scripts that exfiltrated sensitive information from compromised devices, including private keys that granted the attackers access to crypto funds.
Safety says that the victims - the developers that downloaded and ran the infostealers - were located in Russia.
The attackers, on the other hand, seem to be located in the United States, based on the IP addresses where the exfiltrated data was relayed.
These things were enough for the researchers to ask if this was a US-backed threat actor targeting Russia, probably due to currently strained geo-political relations between the two powers.
But npm, as a platform, is not Russian, or managed by the Russians. The npm platform is run by npm, Inc., a company that was originally independent but is now a subsidiary of GitHub, which itself is owned by Microsoft.
Still, Russia has multiple state-sponsored and affiliated threat actors known to target cryptocurrency users, or large enterprises which are then forced to make ransom payments in crypto. Groups such as Evil Corp, Sandworm, and APT28 (Fancy Bear) have been linked to campaigns that either exfiltrate cryptocurrency or deploy ransomware for financial gain.
Therefore, it is not too far-fetched to speculate if this attack was aimed at crypto criminals, as well as regular crypto developers.
Via The Register
You might also like- Crypto hacker steals $14.5 billion in Bitcoin using a gaming PC and nobody notices for five years
- Take a look at our guide to the best authenticator app
- We've rounded up the best password managers