There is an immediately obvious use case for the rugged Tank 3 Pro and its 23,800mAh battery — it would do well as a camping companion. But even then, it might be a little overkill because the enormous battery (which features 120W fast charging), rear LED light, and rugged IP68-certified design aren’t the only standout features.

For starters, it’s also packing a 200MP main camera sensor among the five on offer. There’s also a 50MP wide-angle sensor, an 8MP telephoto sensor, and a 64MP night vision sensor. Another 50MP sensor handles front-facing hole-punch selfie shots. A high megapixel count doesn’t guarantee great snaps, but it’s great for bragging rights.

When a smartphone and a power bank love each other very much…

The most interesting feature of the Tank 3 Pro isn’t something we’ve seen many phones do before. It’s got a built-in 100 lumens 120Hz DLP (digital light processing) projector that even has an adjustable focus. It’s perfect for impromptu outdoor sales pitches or ‘camping’ indoors while your load is being shed.

The regular smartphone bits are worth mentioning too. There’s a large 6.79in FHD display that will manage up to 120Hz refresh rate. A MediaTek Dimensity 8200 SoC powers everything, helped by a generous 16GB or 18GB of physical RAM and 512GB of storage. If that somehow isn’t enough, it supports up to 2TB with a memory card.

For wireless connectivity, expect 5G, Wi-Fi 6, and Bluetooth 5.3 support.

Finally, a 40m laser rangefinder lives inside for… doing stuff like this…

Don’t expect to find the Tank 3 Pro on local shelves but if you’re serious about picking one up, it’s listed on the official AliExpress 8849 store for a respectable R11,300 before shipping and import fees.

We can only imagine how awful that must be. Thankfully, this terrible plight is easily solved with the new 77in LG M3 wireless OLED evo TV — all it takes is a big enough wall and R130,000.

LG M3 OLED TV wins cable hide-and-seek

The LG M3 OLED evo is billed as “the world’s first completely wireless OLED television,” although you’ll still need to provide it with power. That’s the only cable you’ll find running into or out of this TV. Everything else is handled by the Zero Connect Box.

This isn’t the world’s first TV to move its ports from the TV to an external box, some TVs could do that already. But those lesser TVs still require you to plug that external box into the TV somehow. The M3 removes that requirement, transmitting the audio and video signal between the two wirelessly without losing any quality.

Any decoders, game consoles, or AV receivers you would normally plug into the TV now plug into the Zero Connect Box’s three available HDMI 2.1 ports — one of which supports eARC. It also houses an Ethernet port, a place to plug in a satellite or TV aerial, two USB-A ports, and an optical audio jack.

The wireless tech has a few limitations — its technology, not magic. While the box can be placed anywhere in the room up to 10m away thanks to its adjustable antenna, it still requires line-of-sight to the TV for the best results. You might get away with sticking it in a cabinet below the TV but that’s not guaranteed.

The other bits of the M3 OLED evo, like the usual smart TV features, jaw-dropping picture quality and colour accuracy, and webOS interface stick around. The M3 is based on the already impressive G3 OLED evo, so it also benefits from a boost in brightness thanks to LG’s use of MLA (Multi Lens Array) tech powered by the Alpha 9 Gen 6 processor.

This TV isn’t going to be for everyone. Some people might like their TV cables sticking out, who are we to judge? Then there’s the price. R130,000 is what the smallest 77in model costs. It’s also available globally in 83in and 97in sizes. You might be able to convince LG to bring one in for you but they’ll obviously cost more.

SIMA, which stands for Scalable, Instructable, Multiworld Agent, is currently still in its research phase and is being trained to learn a broad range of gaming skills across a variety of scenarios — instead of just destroying humans at StarCraft II.

Through partnerships with video game developers Hello Games, Embracer, Tuxedo Labs, Coffee Stain, and others, SIMA is learning how games work and how to apply what it learns to games it’s never seen before. DeepMind’s eventual aim with SIMA, other than furthering natural language AI model research, is for it to be a devoted member of your party that does what it’s told and doesn’t take all the good loot.

Does this SIMA good idea?

“SIMA isn’t trained to win a game; it’s trained to run it and do what it’s told,” said Google DeepMind researcher and SIMA co-lead Tim Harley, according to The Verge.

SIMA researchers have focused on games that involve open-world play, rather than linear or story-driven titles, so the agent can learn to follow instructions. To achieve this, SIMA was trained by watching pairs of humans play a game — where one watched and gave instructions while the other carried them out. In a different scenario, players played freely while DeepMind researchers recorded instructions that would’ve resulted in what the player did.

We’ll admit this sounds rather appealing. If you’ve ever played an online co-op game that drops in randoms, you’ll know how risky that can be. There’s a good chance of them ruining your game, whether through incompetence or toxicity.

Having an AI party member who follows instructions means you won’t have to worry about watching your back or your hard-earned loot. Don’t feel like spending hours collecting resources? Tell SIMA to do it while you handle more important tasks.

Read More: DeepMind is back at it again, this time teaching AI how to play football

However, as appealing as this might sound, it’s worth remembering how training AI models on human behaviour — especially when online human interaction is involved — has gone in the past. TayTweets, anyone?

This probably isn’t a problem in a controlled research environment but, should SIMA ever be trained on average human-based online gameplay, we doubt it will take long before the griefing starts.

Source

While the Magic V2 isn’t exactly new — it launched in China in July last year — it could still prove disruptive to the local foldable smartphone market. When it lands, it’ll claim the title of being the thinnest and lightest folding smartphone in the country against Samsung’s Galaxy Z Fold 5 and Huawei’s Mate X3.

The Magic 6 Pro, on the other hand, is entirely new and was only recently announced at MWC last month. Honor has Samsung’s Galaxy S24 Ultra in its sights as the only other phone available in the country (so far) with on-device AI capabilities. Both devices also feature pretty respectable spec sheets.

Honor brings fresh Magic V2 to SA

Being slightly older, the Magic V2 sports last year’s flagship Snapdragon 8 Gen 2 5G chipset with 16GB of RAM and 512GB of storage. It might not be new but that chipset still offers impressive performance. Although, we’re eager to see how Honor handles the heat in the Magic V2’s remarkably thin and light chassis — we’re talking 156.7 x 145.4 x 4.7 mm unfolded and 156.7 x 74.1 x 9.9 mm folded while only weighing 231g.

The folding internal OLED display measures 7.92in with a 2,156 x 2,344 resolution. It uses LTPO (Low-temperature polycrystalline oxide) tech meaning it can vary its refresh rate, reducing it to save battery and increasing it up to 120Hz for buttery smooth scrolling or gaming. The Magic V2 also features a 6.43in LTPO OLED cover screen with a 1,060 x 2,376 resolution and HDR10+ support.

The Magic 6 Pro packs an equally impressive 6.8in LTPO OLED screen with a reported max brightness of 5,000 nits — that’s only just shy of the Sun’s 1.6 billion nits but certainly the brightness display you can get here. Behind the panel sits the latest Snapdragon 8 Gen 3 chipset along with 12GB of RAM and 512GB storage options.

Honor’s marketing is making a big fuss about the Magic 6 Pro’s camera performance. It houses two 50MP sensors — one of which will use a variable aperture — along with a 180MP sensor behind a periscope telephoto lens. Up front, you’ll find another 50MP selfie cam and a TOF (time-of-flight) sensor for depth and biometrics.

The Magic V2 isn’t as focused on snapping pics but that doesn’t mean it won’t be capable. It also houses two 50MP sensors along with a 20MP telephoto shooter around back and a 16MP selfie cam.

When it comes to portable power, the Magic V2 uses Silicon-Carbon (Si-Ca) in its 5,00mAh battery, instead of the traditional Lithium-Iron Phosphate (Li-Po) used in most modern smartphones, and uses Honor’s 66W SuperCharge tech for refilling. The Magic 6 Pro packs a 5,600mAh battery which is based on the same Silicon-Carbon tech in the Magic V2 and will support charging at 100W wired and 66W wireless.

The Magic V2 and Magic 6 Pro will officially launch in SA next week. We’ll need to wait until then to find out when the devices will be available for purchase, although we don’t think it will be too long after the launch.

Some of the now-blocked prompts include “pro choice,” “four twenty,” and “pro life” after the platform was found to produce “demons and monsters alongside terminology related to abortion rights, teenagers with assault rifles, sexualized images of women in violent tableaus, and underage drinking and drug use,” according to a CNBC report.

Stuff can confirm that when provided with those prompts Copilot Designer shows a message saying it couldn’t generate images because “something may have triggered Microsoft’s Responsible AI guidelines.”

Microsoft’s Designer gets slap on AI wrist

A Microsoft spokesperson told CNBC about the changes, “We are continuously monitoring, making adjustments and putting additional controls in place to further strengthen our safety filters and mitigate misuse of the system.”

As reassuring as Microsoft probably wants that to seem, the fact that you can still generate questionable images or easily get Copilot Designer to infringe on copyrights after the changes isn’t doing the company any favours.

Read More: Microsoft Copilot celebrates 1st birthday with redesign on web and mobile

Neither is the fact that Shane Jones, the Microsoft engineer who wrote to the FTC, first tried reporting his findings internally back in December 2023. Microsoft acknowledged his concerns but that’s about as far as it went, instead referring him to OpenAI. After not hearing back from them, Jones posted an open letter to LinkedIn asking OpenAI’s board to suspend Dall-E 3, the AI model Copilot Designer is based on, until the issues could be resolved.

Microsoft’s lawyers didn’t like that and told Jones to remove his LinkedIn post, which he did. This is what prompted him to write letters to FTC chairperson Lina Khan and Microsoft’s board of directors, letters he shared with CNBC.

Not a particularly good look for Microsoft.

Source

For those who do loads of printing, the benefits probably seem enticing. If you’ve ever had to print something in a hurry only to find you’ve run out of ink or the printer jams halfway through the job, you might even welcome a subscription plan like this. At least, that’s what the folks at HP were probably thinking when they came up with the idea.

In reality, HP‘s subscription plan is questionable due to a worrying number of fine-print clauses and subscription limitations. On top of that, after the plan’s 30-day trial, you’ll be locked into a two-year subscription or forced to pay a cancellation fee that, in some cases, exceeds the printer’s retail value.

“Never own a printer again”

About those asterisks. Firstly, subscribers will receive “[their] choice of HP’s most trusted printers”  but only if their choice is one of three models — the ENVY 6020e, ENVY Inspire 7258e, or the OfficeJet Pro 9010e. These three models form the basis of HP’s three-tiered plan with each tier sporting a starting cost of $7/m, $9/m, and $13/m, respectively (or R132/m, R170/m, and R245/m converted).

However, it should be noted that those starting prices also come with per-month print limits. If you’d like the cheapest printer but think you’ll print more than 20 pages per month, you’ll need to cough up $9/m for 50 pages per month or $11/m to increase your limit to 100 pages per month.

The most pricey package includes the OfficeJet Pro 9010e which will allow affluent subscribers to print up to a whopping 700 pages every month for “just” $36/m (R680/m). If you exceed your “monthly allowance”, HP’s FAQ page says “additional sets of 10-15 pages will be automatically added to your monthly invoice at $1 per set.” Thankfully, any pages you don’t use of your quota will accumulate and you can “roll over up to three times the number of pages in your monthly plan.” So it’s like buying mobile data and we all know how not-frustrating that process is.

HP wants to watch you print

The other asterisks, like automatic ink delivery and 24/7 support, might seem useful but that’s only because you’re used to paying exorbitant amounts of money for them. What you may not be used to is HP looking over your shoulder at every print job. One of the many questionable terms in this subscription plan is HP’s insistence on keeping your rented printer connected to the internet. If you don’t, HP’s All-In Plan terms of service states:

“If the Printer is not connected to the Internet, then You may not be able to access Your Services including the ability to print; however, You will continue to be charged for the Services.”

HP justifies this requirement by saying the company needs to monitor ink cartridge levels, page count, and “to ensure that You are receiving the requested Service, to prevent unauthorized use of Your account and to improve Your experience with HP products and services.”

That kind of monitoring makes sense. What makes less sense is why HP would need to remotely monitor and track the type of documents you print, the devices or software you use to start a print job, and “other types of metrics related to your Services.”

Regarding that data, although HP’s privacy policy does state that users can opt out of sharing personal data, the All-In Plan’s TOS also states:

“Subject to the terms of this Agreement, You hereby grant to HP a non-exclusive, worldwide, royalty-free right to use, copy, store, transmit, modify, create derivative works of and display Your non-personal data for its business purposes.”

Thankfully, HP’s All-In Plan is only available to US customers, for now. And they can keep it.

Source

The Amazon Appstore, the only officially supported avenue for downloading Android apps on Windows devices, is no longer available for download from the Microsoft Store. Thankfully, currently installed Android apps will still function and, depending on the developer, can still receive updates until the underlying system is deprecated on 5 March 2024.

If you’re one of the few who regularly use Android apps on Windows, it’s time to look for alternatives.

Microsoft kills WSA 

When Microsoft launched Windows 11 in October 2021, one of the biggest new features it touted was official support for Android apps. This support came from a Windows 11 feature called the Windows Subsystem for Android (WSA). It used built-in virtualisation to allow apps developed for the Android operating system to run on Windows machines without the need for an Android virtual machine or third-party software.

Well, third-party software was still involved — you needed to download Amazon’s Appstore via the Microsoft Store to download and use Android apps, officially. This ‘exclusive’ partnership between Microsoft and Amazon meant that WSA was kneecapped from the start, unable to access the far broader Android app offerings in Google’s Play Store.

Amazon’s Appstore offers a paltry selection of apps, which didn’t make the switch for users using Android emulators like BlueStacks or LDPlayer seem very appealing. There were, of course, workarounds — it didn’t take people long to figure out how to sideload the Google Play Store into WSA. But why bother with a workaround like that when you could just keep using an emulator or a web-based version of the app?

Whether you’re a fan of the books, encountered the Dune Chronicles for the first time with Villeneuve’s previous film, or a complete newcomer, we highly recommend you see the first film (if you somehow haven’t yet) and this one immediately after.

Dune: Part Two is to Dune: Part One, what The Godfather II and The Dark Knight were to their prequels. It’s that good.

Beginnings are such delicate times

Without spoiling anything, Dune: Part Two picks up where Part One finished. It follows a young Paul Atreides as he joins the Fremen on a journey of self-realisation through the desert planet of Arrakis while seeking revenge for the massacre of his family by the evil and bald Harkonnens who had help from Emperor Shaddam IV. That’s as simplified and stripped back as it gets for a film synopsis but there is far more substance to the main plot and subplots that’s worth discovering for yourself.

If you thought Part One had grand ambitions or set the bar high, you’ll be as surprised as we were by how easily Part Two clears that bar. Everything that Part One was praised for (which was most of it), was just as great, if not better in Part Two. That can make Part One seem a little tiny in comparison, almost as if it was nothing more than a two-and-a-half-hour setup for Part Two.

With Part One being the movie it was, we’d strongly argue against that. The best way around this would be for Legendary Entertainment and Warner Bros. to release a full-length five and five-and-a-half-hour director’s cut of both parts edited together. We wouldn’t even be mad if they adapted even more of the first book. If anyone could make that work, it’s Villeneuve.

A time of quiet between storms

We’ve been fans of Villeneuve since his return to cinema in 2009 with Polytechnique. Our respect and enjoyment of his work have only increased with his subsequent films – Prisoners, Sicario, Blade Runner 2049 – and with Dune: Part Two he is at the top of his game. He was given the freedom (and budget) to develop his ethereal style in this film and makes full use of that freedom in directing and executing his creative vision which brilliantly expanded, and now completes, his adaptation of Frank Herbert’s first book.

An easy example to point at is the choices he made in which parts of the book to include and which parts either didn’t make the final edit or were omitted entirely. With a runtime of 155 minutes, this certainly isn’t a short film, yet every scene is expertly crafted and expands on the world-building, often in surprisingly thoughtful ways. This goes a long way in making the world of Dune feel easy to believe in and when you’re saying that about a movie with giant God-like sandworms, that’s a sign of a job well done.

It isn’t only Villeneuve’s direction that makes Dune: Part Two worth seeing. Every cinematic aspect of the project elevates the film to something more than just the sum of its parts. In particular, the production design stood out to us because its attention to detail helped craft a more believable fantasy world. When you’re shown a new thing, especially in a fantasy setting, and the design or use of that thing immediately makes sense within the given context, you have the design team to thank.

Each man is a little war

You can have the best director and production team in the world and your film will still fall flat with a weak cast. That wasn’t an issue with the first movie and it is even less of an issue in Part Two. Every actor made their understanding of the material clear in their performance but three are worth a special mention.

After watching Timothée Chalamet grow into his role as a naive Paul Atriedes in Part One, the maturity and dark complexity he brings to the character in Part Two was impressive. Zendaya’s performance as Chani, Paul’s Fremen chaperone and later love interest, is equally impressive. Her character gets substantially more screen time in Part Two, allowing us to revel in her screen presence and properly enjoy the layered nuances of her performance.

Austin Butler joins the Dune cast as the sadistic Feyd-Rautha, a member of House Harkonnen and the Baron’s chosen heir. His inclusion provides a convenient avenue for further exploration of the Harkonnen House and its homeworld Geidi Prime. In a film of strange things, it somehow manages to be the strangest. Butler might still be relatively unknown but expect that to change if the physical and vocal performance displayed here are a sign of what’s to come. The explosive energy he brings to the screen can at times feel overdone, but it doesn’t detract from Butler’s performance. He makes it his own – another sign of a successful actor-director relationship.

Working with Villeneuve, Hans Zimmer and Greig Fraser have both produced some of their best work in the music and photography departments respectively. Zimmer’s sweeping score is as grandiose as Fraser’s breathtaking visuals. Listening to the soundtrack for Part One and Part Two sequentially makes Zimmer’s choices clear as wormsign. The entire film feels larger than life but Part Two’s score mirrors the film’s characters and somehow feels more human.

The same can be said of Fraser’s work as director of photography. Where his composition in Part One introduces Arrakis to the audience as a harsh and barren alien planet, in Part Two it feels familiar and almost inviting. Thanks to the slightly longer runtime, we’re treated to even more glamour shots, and when they look this good it’s hard to complain.

Dune: Part Two verdict

Dune has long had a reputation as being ‘unfilmable’, mostly due to the sheer volume of content but also in part due to its many interpretations. That didn’t stop Alejandro Jodorowsky and then David Lynch from trying. Jodorowsky’s film never made it to completion and although Lynch’s 1984 film did, many fans of the books (including Lynch himself) don’t count that as a success.

We’re quite sure hardcore book fans will find something to complain about with Villeneuve’s adaptations and they’ll probably have a point. We’re not saying Dune: Part Two is the be-all and end-all of Dune adaptations but it’s the best one we have right now. Go and watch it on the biggest screen you can.

Dune: Part Two is now showing in theatres.

Surprisingly, Kaspersky’s research showed that the number of overall cyber threats in South Africa decreased by 29% in 2023 compared to the previous year. That could mean cyber thugs are getting bored or it could mean their old cyber threats are still working so they don’t need to create new ones.

Phishing with cyber bait

Our money’s on the latter because while the number of overall threats in South Africa has decreased, phishing scams increased by 29%. If you’ve ever received a strange email to inform you of a surprise inheritance or you have a parcel waiting for you that you never ordered, you’ve seen a phishing scam. Kaspersky reckons that 35% of its South African users were affected by online threats in 2023.

The overall decrease could also be the result of 2023’s record-breaking number of load shedding days. You don’t have to be a cyber security expert to know you can’t hack a system when it doesn’t have power.

Phishing attacks come in different flavours and are the most prominent form of cyber attack in our region. They might vary in form but they are all after the same things – your personal info, financial data, or access to your accounts. Stealing people’s money is easy to understand and has been around since before the internet was a thing. But most people don’t have hoards of cash sitting in their accounts, ripe for the stealing.

Read More: How SIM swap scammers can swindle you

Access to your social media accounts, like Facebook, can also be useful to cybercriminals. If they’re lucky, they’ll find out what your favourite pet’s name was, where you grew up, and other answers to common ‘security’ questions. Even someone’s Telegram account, or simply their cellphone number, can be used for nefarious purposes — like mass-inviting their contact list to a forex trading group in the hopes they’ll sign up to a pyramid scheme.

Cyber threats are evolving

Considering the rate at which generative AI is progressing, it shouldn’t be a surprise that it is intimately involved with the cyber threats of the future. Cybercriminals already use it to improve their phishing attempts or create convincing fake websites. Generative AI has also made it much easier to create video and audio deepfakes that could be used to swindle the unsuspecting. Then there are the vulnerabilities of the generative AI platforms themselves.

Most platforms have at least some safeguards in place which are supposed to prevent their misuse, but cybercriminals are already finding ways around them. Vladislav Tushkanov is a lead data scientist at Kaspersky’s expert system development and technology research department. He explained that these criminals already have methods to alter the behaviour of chatbots or ‘unlock’ topics of discussion and generated outputs which are supposed to be protected from misuse.

For now, generative AI misuse is largely focused on wreaking havoc on celebrities’ lives, though a recent case of an employee being fooled by a deepfake video of what they thought was their boss ended up costing the company millions. Cases like that are likely to increase as are the number of generative AI platforms available, the things they can do, and the issues that accompany them.

UPDATE 01/03/2024: A previous version of this story mentioned that “35% of South African users, or around 21 million people, were affected by online threats in 2023.” This has been corrected as this number was not a true reflection of Kaspersky’s data which is based on the number of users of Kaspersky products and not the per capita number of internet users in SA. We regret the error.

This isn’t the first text-to-video tool to emerge from the generative AI boom but, based on the examples shown, it generates the most realistic videos we’ve seen so far.

While Sora hasn’t received a full release yet, it is already capable of creating “complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background,” while also understanding the context of the user’s prompt and how it might affect the simulated physical world.

It can also provide multiple shots of the generated video from the original prompt while maintaining the visual style and any persistent subjects or characters in the prompt.

Sora still makes mistakes, how long will that last?

When Sora eventually becomes available, it won’t be without limitations. Generated videos will be capped at 60 seconds, at least initially, so we probably won’t see any feature-length Sora-generated films in a hurry.

As with any generative AI model, Sora is still prone to mistakes. OpenAI says Sora struggles to accurately simulate a complex scene’s physics and has trouble with “specific instances of cause and effect.” Adding a bite mark to a cookie someone’s tasted, for example.

Sora is currently only available to OpenAI’s ‘Red Team’. These are folks who look for possible ways it could be abused or exploited, like prompting it with malicious material, to learn how the model reacts so they can make adjustments to prevent the same reaction when it launches — kinda like breaking into your own house to find your security weak points.

OpenAI is also working with visual artists and filmmakers who will hopefully provide constructive feedback to improve the model before a wider release.

There are already examples of short films made with AI-generated content, like Sunspring, which was written by an AI model trained on existing movie scripts. That was released in 2018, a good few years before ChatGPT was introduced — and it shows. It also still required humans to act and shoot the film.

With Sora, it may soon be possible to remove humans from the process altogether. We’re sure Martin Scorsese is thrilled about that.