Feed aggregator

The majority of Americans without degrees still believe in the value of higher education, according to the poll. But not all college degrees are created equal.

(Image credit: Brandon Bell)

President Trump's past medical reports have been glowing. But at 78, his age puts his health in the spotlight — as it was for his predecessor, President Joe Biden.

(Image credit: Win McNamee)

When OpenAI made headlines with its ChatGPT launch in 2022, it was clear that its evolution into the workforce was only a matter of time. According to Sam Altman, technology is advancing so rapidly that we can expect this shift within the next few months. But is replacing humans with AI employees a genius business move, or akin to hiring unpredictable interns?

While human employees come with additional costs such as training, sick leave, and salaries, AI workers bypass these expenses. Instead, they introduce their own unique challenges that can be just as disruptive for employers.

Holidays vs. System Updates

Virtual workers don’t need to rest in the same way as humans, and it’s often seen as the biggest benefit of AI workers. AI doesn’t sleep and can work at all hours, unlike human employees who have set hours and need regular breaks. However, while AI never has the need for out of office replies, it does require plenty of maintenance and updates to perform.

For virtual employees, system downtime is the AI equivalent of a holiday, while planned maintenance or unexpected outages can disrupt operations just like a human taking time off for appointments or illness. The promise of 24/7 availability isn't entirely true if there are issues in the infrastructure.

According to a recent report by Splunk, unplanned downtime costs the world's largest 2,000 companies a combined $400 billion per year, averaging $200 million per company. From lost revenue, regulatory fines, and hidden costs like slower time to market and worsened brand reputations, downtime can be a very costly challenge.

So, although AI doesn't need a time off to recharge, ensuring its performing as it should needs to be a top priority for any organization introducing AI employees. AI models need to be regularly updated to avoid deteriorating in accuracy, leading to outdated or incorrect outputs. The recent shift towards retrieval-augmented generation (RAG) is an attempt to combat this, by enabling Large Language Models (LLMs) to source data from verified external sources, but it’s not an easy fix and still requires close monitoring.

Sickness vs Hallucinations

An AI employee won’t call in sick, but they can suffer from bugs and hallucinations that can be equally as disruptive. An AI hallucination occurs when a model generates incorrect or nonsense outputs, and this happens more than you would think. For instance, in 2017, Microsoft's AI chatbot Tay had to be shut down after it began generating offensive tweets less than a day after its launch.

Just like human immunity, different AI models vary in their robustness against bugs. OpenAI, Anthropic, and Google are all racing to reduce model hallucinations, but even top of the range models can still be unreliable at times. For organizations with AI employees, the priority should be treating these issues but also working to detect any potential future issues before they cause harm. Ongoing governance is therefore essential, in the same way as managing employee performance to identify any challenges or skills shortages.

Monthly pay vs. Updates

It’s a common misconception that AI is a pay-once, use forever deal. The cost of deploying AI solutions isn't set by a wide-open market; it's influenced by major players like Sam Altman and Google’s CEO, Sundar Pichai. As AI models evolve, it’s likely that businesses may need to invest further into costly updates or building entirely new systems to keep up with competitors. This ongoing investment could be the same as providing pay raises or professional development for human employees.

For example, ChatGPT recently launched a business-tier solution, and companies are already integrating AI co-pilots into their workflows. These tools are often tied to evolving licensing fees, cloud computing costs, and retraining requirements which means the AI you implement today could be severely outdated in just two years, requiring additional investment to maintain the same level of efficiency.

Training vs. Data input

In the same way that human employees need training to improve and develop in their role, LLMs need vast amounts of data to learn from to enhance performance. The more complex a model is, the more data it requires. For example, some models may need at least 10 million labelled items. It’s not just about the volume of data; the quality and relevance of this data are also crucial, as AI can only be as good as the data it is trained on. Poor data input means poor data output.

Linking abstract knowledge to real-world examples, also known as “Grounding AI” is a way to improve the ability of AI models to produce better predictions and responses. To do this, many companies are turning to RAG as a way of enhancing AI accuracy, by allowing models to pull from verified external sources instead of relying purely on pre-trained data.

Just like human development, training AI needs to be done regularly and consistently to get the best results. Especially as regulations around AI tighten and change, it’s crucial for companies to remain compliant with data protection laws, bias mitigation requirements, and industry-specific standards. Especially for businesses in highly regulated industries such as finance, firms using AI must align with regulations like the EU AI Act, ensuring models are transparent and accountable. This means regular audits, governance, and retraining.

So, what’s the verdict?

The truth is, it's not about choosing whether AI or human employees are better for a business. Instead, the real power comes by harnessing both. Issues such as governance, training, hallucinations and updates can be just as time consuming and expensive for businesses as managing human talent. The value of AI is in enabling humans to take more time for the things that only they can do; deeper thinking, creativity and building authentic human connections – capabilities that AI simply cannot, no matter how much it continues to evolve and advance.

Granted, implementing AI tools into business processes will free up time and improve efficiency, but that doesn’t mean that humans don’t still have a place. When the efficiency and intelligence of AI join forces with the creativity, empathy, and strategic thinking of humans, the possibilities are limitless. Together, they can drive innovation, boost productivity, and propel businesses towards long-term sustainable success.

We list the best task management app.

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

• Nutribullet and McLaren have teamed up for a set of three themed blenders
• Each blender features McLaren's signature gray and papaya livery colors
• The original versions of all three received high scores from our reviewers

Nutribullet, maker of some of the world's best blenders, has teamed up with the McLaren Formula 1 team for a set of F1-themed soup and smoothie-makers. The collaboration might sound like a belated April Fools joke, but it's very real, and the three race-ready blenders are available to buy now direct from Nutribullet.

According to Nutribullet, the three blenders are "infused with McLaren’s signature papaya aesthetic – blending speed, precision, and innovation from the racetrack to the kitchen."

The trio include the Nutribullet x McLaren F1 Team Portable – a new version of the Nutribullet Portable Blender we reviewed last year. Our tester Victoria Woolaston praised its light weight, which makes it ideal for carrying to the office or gym, and its very reasonable price tag. Its battery life was also impressive, and she found it lasted around two weeks between charges in typical use.

She did note that it struggled somewhat with particularly tough ingredients, and it was sometimes necessary to run two blending cycles to incorporate difficult vegetables into smoothies, but no more so than other single-serve blenders at this price point.

(Image credit: Nutribullet, McLaren) Pulp position

Also on the starting grid is the Nutribullet x McLaren F1 Team Pro 900: a papaya and graphite-hued version of the more powerful Nutribullet 900. Our reviewer Emily Peck found that the 900 produced consistently good results, and noted that it's a particularly good pick for small kitchens.

"There's only one setting, but it offers you 50% more power than the original Nutribullet, with the speedy blade delivering super-smooth results," she said.

Finally, there's the Nutribullet x McLaren F1 Team Ultra. The original Nutribullet Ultra earned four and a half out of five stars in our review, producing silky smoothies and nicely textured hummus with ease. It's also one of the quietest blenders we've tested to date.

There's been a trend in kitchen appliances recently for subtle, natural shades like buttery yellow and leafy green, but if that's not your style then one of these blenders could add a welcome shock of color to your kitchen. You could even use one to whip up a quick papaya and coconut smoothie for refreshment during this weekend's Bahrain Grand Prix.

You might also like

• The Ninja Blast Max is so powerful it can obliterate frozen strawberries and even make kale smoothies delicious
• Nutribullet's new blender makes everything from nut milk to soup with a single touch
• I've finally found a green smoothie that doesn't taste like eating raw veg

This review first appeared in issue 348 of PC Pro.

Scan opts for the Intel route with its lower-priced system, as the 3XS GWP-ME A132C packs the best Intel CPU available in this class. Where other manufacturers chose the Core i9-13900K, Scan steps things up with the i9-13900KS. This incorporates eight P-cores with Hyper-Threading and 16 E-cores without, but the P-cores have a base frequency of 3.2GHz and 6GHz maximum boost, while the E-cores start at 2.4GHz and go up to 4.3GHz (although the latter is the same as the K variant).

Scan also takes advantage of the processor’s support for DDR5 memory by supplying 64GB of 5,600MHz RAM in the form of two 32GB DIMMs. This leaves two slots free for upgrade. The graphics choice is dependable, too: you can’t go wrong with Nvidia’s RTX A5000 handling 3D acceleration. Sporting 8,192 CUDA cores and 24GB of GDDR6 memory on a 256-bit bus providing 768GB/sec of bandwidth, the A5000 is a potent GPU.

The components of this workstation are built into a capacious Fractal Design Meshify 2 chassis, which offers lots of space for storage upgrades. Scan supplies only a single drive, but it’s a good one – a 2TB Corsair MP700 NVMe M.2 SSD, which supports PCI Express 5. According to testing with CrystalDiskMark 8.0.4, this drive provides sustained reading of 10,074MB/sec and writing at 10,190MB/sec, nearly 50% faster than the best PCI Express 4 NVMe SSDs. Only the Crucial T700 used by Armari this month is faster.

(Image credit: Future)

Despite its Core i9-13900KS CPU, the A132C came second to systems equipped with the K variant in the PC Pro benchmarks, although a score of 827 is still superb. Its Maxon Cinebench R23 CPU rendering result of 37,442 was also beaten by the Chillblast Apex Core i9 and Armari’s AMD Ryzen 9 7950X. However, the Blender CPU rendering time of 296 seconds beat other Intel-equipped workstations.

The RTX A5000 graphics delivered as expected with SPECviewperf 2020 v3.1 viewsets, showing competent scores for 3D animation, engineering and CAD. The excellent OpenCL result in LuxMark 3.1 of 15,287 was mirrored by a Blender GPU render time of just 143 seconds, although Armari’s AMD Radeon W7800 was faster, and the PCSpecialist Onyx Pro’s Nvidia GeForce RTX 4090 was in a different league entirely.

This is an excellent workstation for £4,500, and superb value when you consider its components, but it’s pipped to the top spot by Armari’s entry this month.

We also ranked the best digital art and drawing software.

This review first appeared in issue 348 of PC Pro.

PCSpecialist has thrown everything at its £10,000 workstation. Not only does it incorporate AMD’s range-topping 64-core Threadripper Pro 5995WX, but it also includes AMD’s latest professional graphics.

The CPU supports multithreading so offers 128 threads. The base clock is 2.7GHz with a boost to 4.5GHz. While there’s no DDR5 support yet for the Threadripper Pro, the eight-channel memory configuration improves bandwidth. PCSpecialist includes a whopping 256GB of 3,200MHz RAM in the form of eight 32GB modules, taking advantage of the extra throughput.

PCSpecialist opts for the brand new AMD Radeon Pro W7800 for graphics acceleration. This combines 4,480 RDNA 3 unified shaders with 32GB of GDDR6 frame buffer operating with 576GB/sec bandwidth.

Two storage devices are supplied with the Ultra. The 2TB Samsung 990 Pro NVMe M.2 drive operates at PCI-E 4 speeds. It achieved sustained reading at 7,404MB/sec and writing at 6,818MB/sec. The other storage device is a capacious 10TB Seagate IronWolf Pro 7,200rpm mechanical hard disk, delivering 261MB/sec reading and 253MB/sec writing.

(Image credit: Future)

The Onyx Ultra managed a score of 794 in the PC Pro benchmarks, beaten only by the systems using Intel’s Core i9. Its main weakness 231 in image editing, which is a single-core task. The Cinebench R23 multicore rendering score of 71,519 is phenomenal, although Armari went further with the same processor. The Blender Gooseberry CPU render time of 134 seconds was also behind Armari.

Although the AMD Radeon Pro W7800 is a fantastic new graphics accelerator, it benefits from being paired with fast single-core CPU speeds, which the Threadripper Pro can’t offer. With SPECviewperf 2020 v3.1, the scores of 194 in 3dsmax-07 and 792 in maya-06 are superb, but Armari did better by partnering the GPU with a Ryzen 9 7950X. Similarly, engineering and CAD viewsets were behind. However, a number of these scores were ahead of the Nvidia RTX A5000 and A6000. The LuxMark 3.1 score of 12,317 and Blender GPU time of 153 seconds were more mediocre.

This is a fantastically powerful workstation, although PCSpecialist hasn’t squeezed as much rendering performance out of the CPU or modelling ability from the GPU as Armari. It’s well worth considering, though, with great all-round abilities and lots of storage for media assets.

We also rated the best free video-editing software.

• PowerColor has launched a new Radeon RX 9070 XT Red Devil Spectral White GPU at $899.99
• Its Hellhound model is available at $849.99
• Best Buy features even older RDNA 3 GPUs at high prices above their MSRP

If you thought the GPU market would get any better after recent pricing issues, think again: unfortunately, there's yet another huge price hike on, this time on an AMD Radeon RX 9070 XT partner card.

As reported by VideoCardz, PowerColor has launched an overclocked Radeon RX 9070 XT Red Devil Spectral White GPU, with a price point at $899.99 - which is a ridiculous $300.99 more than the card's $599 MSRP. This is featured alongside PowerColor's Hellhound variant on Micro Center, priced at $849.99.

It's a similar move to another board partner of AMD's: I've previously highlighted the XFX Radeon RX 9070 XT on Best Buy, which was being sold for up to $939.99 at the time, and we're now seeing a similar pattern with PowerColor models. The same applies to older RDNA 3 GPUs, with prices above MSRP despite being more than two years old.

While there are currently tariffs on imported goods in the US - which could make prices even more absurd in the near future - these price increases were happening long before Trump kicked off his tariff war. From the moment Nvidia's RTX 5000 series launched, price hikes across multiple retailers came into effect.

It's also worth noting that AMD's Radeon RX 9070 series GPUs were originally being sold at retail price (at least during the launch), and now that's changed entirely within the span of just a few weeks. It can be argued that this issue stems from a lack of availability and a shortage of GPU dies (especially with such a high GPU demand from consumers, and the new hunger for silicon from the burgeoning AI industry), but this isn't the first time retailers and AIBs have raised prices to ridiculous amounts.

(Image credit: DarwelShots / Shutterstock) Is the GPU market on the verge of a major crash?

Since the 2020 pandemic, average prices for GPUs have soared immensely, with multiple PC gamers tempted to overspend - so this isn't exactly a new occurrence. This current state of the GPU market is quite literally the worst I've seen, as I've stated before, and with tariffs now involved, it could be the start of a graphics market crash.

That's a big could be, though, as there is clearly still a large number of consumers willing to buy these inflated GPUs, as most are sold out across most retailers. Regardless, given how matters are now, I can only see prices continuously rising and leaving more and more PC gamers unable to purchase new hardware.

The RTX 5080 is already expensive at its retail price ($999 / £939 / AU$2,019), and partner cards take this up to $1,500 (based on Best Buy listings). As great as the RTX 5080 and RTX 5090 are, I'd argue that there are very few gamers who would be willing to spend these sorts of figures just for an enhanced experience with better performance.

I'd hate to see it, but we could soon reach a point where consumers opt out of any GPU upgrades: with Nvidia and AMD's focus on AI now very clear, we could witness a slow and painful dwindle in priority for gaming GPUs for both Team Green and Team Red...

You may also like...

• AMD's powerful Ryzen AI Max processors are on their way to desktop PCs - and a side effect could kill off discrete GPUs for gaming
• Nvidia's new Game Ready Drivers appear to be so bad that game developers are warning gamers to stay away
• Hang on handheld fans, we could be seeing AMD's FSR 4 upscaling on new handhelds after all - if this new RDNA 4 leak is legitimate

• A new leak suggests Nvidia's RTX 5060 Ti and non-Ti variant will stick with the previous gen's 128-bit memory bus
• This is contrary to other RTX 5000 series GPUs using higher memory buses
• The rumored use of GDDR7 VRAM may be the biggest highlight over their predecessors

Nvidia isn't done with its RTX 5000 series GPU lineup yet, as its unannounced RTX 5060 Ti and RTX 5060 desktop cards are still on the horizon. However, a new leak has supposedly uncovered more of their specifications.

As highlighted by VideoCardz, Nvidia's RTX 5060 Ti and RTX 5060 will use a 128-bit memory bus, the same as their predecessor, according to a leaked shipping description. This is different from the RTX 5070's 192-bit memory bus, along with the rest of the Blackwell GPU lineup, which is important for improving VRAM performance in GPUs.

While this may seem disappointing at first glance, the leaked specifications suggest both GPUs will use GDDR7 memory: the RTX 5060 reportedly has an 8GB VRAM GDDR7 capacity, while its Ti counterpart will have both 16GB and 8GB GDDR7 options. As VideoCardz hints, this could come in handy as a potential boost with much higher speeds and bandwidth over their predecessors, which use GDDR6 instead.

It's worth taking this with a grain of salt: we know that the RTX 5060 will be unveiled eventually, but there's no confirmation yet on Team Green's end, so it's worth waiting for official specs to come to light.

Ultimately, these entry-level GPUs and their chances of success in the GPU market will likely depend on pricing. With availability, scalping, and inflation being Nvidia's main issues with its RTX 5000 series launch, the graphics giant may have a hard time convincing consumers to opt for its lineup if the GPUs prove difficult to obtain at retail price.

(Image credit: Nvidia, Shutterstock) Regardless of how good or bad the new GPUs are, pricing will determine their fate...

It's a sad reality that PC gamers are facing with the GPU market currently in disarray - and with recent tariffs mixed in with scalping and high demand, I doubt it’ll end any time soon.

Unfortunately, it means the RTX 5060 Ti and non-Ti GPUs will likely suffer the same fate (especially if they turn out to be great options for budget gamers). We've seen third-party cards from both Nvidia and AMD selling above MSRP at multiple retailers; if Team Green doesn't make a reference card for either of the upcoming GPUs, it could be much worse, as consumers will be left entirely at the mercy of retailers and Nvidia’s manufacturing partners.

PC gaming can already be an expensive hobby, and all recent speculation points towards this getting worse: besides PC hardware, we now have game prices to worry about too, thanks to Nintendo's absurd $80 game pricing (no, I won't shut up about this).

I really hope Nvidia can surprise consumers with these new cards, as it certainly needs a kickstart for good momentum. It's not just been inflation and pricing, but drivers and missing ROPs have haunted Team Green since the RTX 5090 and RTX 5080 launch. Let's just hope a new launch can get us excited again...

You may also like...

• Nvidia RTX 5050 and 5060 leaks suggest these mobile GPUs could be pepping up budget gaming laptops soon
• Keen to buy an RTX 5060 Ti? Benchmarks and more rumored sightings suggest you don’t have long to wait for Nvidia’s next GPU
• Latest Nvidia RTX 5060 Ti GPU rumor suggests good news and bad when it comes to price

This review first appeared in issue 348 of PC Pro.

PCSpecialist takes a unique approach in this month’s Labs by supplying its system with consumer-grade graphics. It’s also the only company to provide secondary storage along with a main drive. So this workstation has a lot in common with a high-end gaming rig; if you design games for a living, this could be exactly what you want.

The CPU is unquestionably potent. This is a 13th generation Intel Core i9-13900K with 24 cores. Eight of these are P-cores with a 3GHz base clock rising to 5.8GHz on maximum boost with Hyper-Threading, while the other 16 are E-cores that operate at 2.2GHz or boost to 4.3GHz, but without Hyper-Threading. So you still get 32 threads like the 16-core AMD CPUs, but 24 of these are full physical cores not virtual ones.

Since Intel Core i9 processors have supported DDR5 memory for a couple of generations now, PCSpecialist has opted for this RAM type, offering 5,200MHz DIMMs. But it has gone further than this – much further. This system includes an incredible 192GB, which is the maximum supported by the Asus ROG Maximus Z790 Hero motherboard. This is supplied as four 48GB modules, although this isn’t a quad-channel system. It’s safe to say that you won’t be needing to upgrade the memory on this system at any point during its useful lifetime.

There’s space for upgrades, but you won’t need to add to the 192GB of RAM (Image credit: Future)

Now we get to the elephant in the room: the consumer-grade graphics. This is cheaper than the professional equivalent, meaning you can get more power for your money. In fact, PCSpecialist has opted for an Asus TUF GeForce RTX 4090 OC Edition, which is the most powerful consumer GPU currently available. This provides 9,728 CUDA cores, almost as many as the Nvidia RTX A6000, but the 4090 has a newer GPU core design and runs at a much higher clock speed, so promises significantly higher raw processing power – of which more later.

The trusty Fractal Design Black Solid chassis is supplied to house all these components. This isn’t as big as the XL, but still has plenty of space inside and options for storage upgrades. There are six 2.5in/3.5in bays included, with up to 14 possible, and two 2.5in trays as standard, but four possible. PCSpecialist opts to use an M.2 slot for NVMe SSD main storage, but also uses one of the 3.5in bays for a conventional hard disk.

The SSD is a Samsung 990 Pro running at PCI Express 4 speeds. It delivered sustained reading of 7,404MB/sec and writing speeds of 6,818MB/sec, which are good for non-PCI Express 5 storage. The hard disk is a 4TB Seagate IronWolf Pro 7,200rpm mechanical hard disk, offering 260MB/sec reading and 257MB/sec writing. This is pedestrian throughput compared to the SSD but fast for a hard disk, and the extra storage will be handy for greedy media such as 8K video.

The Onyx Pro packs plenty of power for many (but not all) GPU rendering tasks (Image credit: Future)

The Intel Core i9 CPU is very much in its element with everyday tasks. Its overall score of 849 in the PC Pro benchmarks is the joint fastest we’ve ever seen, particularly aided by the multitasking score of 1,057. However, while 35,990 with multicore Maxon Cinebench R23 rendering is superb, Armari’s AMD Ryzen 9 7950X is just ahead, and CPU rendering with Blender is similarly behind AMD, with the Gooseberry frame taking 312 seconds.

But then there’s that GPU. CUDA-accelerated Blender rendering took an incredible 64 seconds, and the LuxMark 3.1 score of 31,713 is staggering. This is also a supreme accelerator for some – but not all – content-creation viewsets. Running SPECviewperf 2020 3.1, the GeForce 4090 managed 316 in 3dsmax-07 and 792 in maya-06. This is a brilliant card for 3D animation. However, while catia-06 saw 165 and solidworks-07 a jaw-dropping 732, snx-04 could only reach 54.57.

If you’ll be running engineering workloads, in particular Siemens NX (represented by snx-04), you’re best avoiding a system with consumer-grade graphics like the Onyx Pro. But if you’re a game designer or 3D animator utilizing GPU rendering, this is a phenomenally powerful machine, and comes highly recommended.

We've also rated the best animation software.

This review first appeared in issue 348 of PC Pro.

We haven’t seen many blue-chip brands in our workstation Labs for some years. This has largely been because these manufacturers stuck with Intel Xeons, even when AMD was in the ascendancy, so wouldn’t have fared well. But Lenovo has been offering the AMD alternative for some years and was the initial partner for the Ryzen Threadripper Pro when it arrived last year. Now we get our first look at what Lenovo can do with this potent CPU.

The system is based on the AMD Ryzen Threadripper Pro 5945WX, which has the lowest number of cores of any CPU this month – just 12. These run at a base 4.1GHz and boost 4.5GHz, with multithreading and support for eight-channel memory. But Lenovo only provides four 3,200MHz DDR4 DIMM modules, so the bandwidth is quad channel with half the throughput of eight-channel, even though the total was a wholesome 64GB.

Lenovo offers a choice of AMD and Nvidia graphics with the P620, and our system came with Nvidia’s RTX A4000. This places this configuration as a modelling workstation rather than more general purpose. Lenovo was IBM’s hardware manufacturer, and when IBM became a services company Lenovo inherited its server and workstation business. The chassis is still reminiscent of the former brand, with excellent tool-free design.

(Image credit: Future)

In this high-powered company, the P620 fell behind, achieving 529 in the media-focused benchmarks, 21,580 in the Cinebench R23 multicore rendering test and 486 seconds for the Blender CPU render. These would have been amazing scores a year ago, but all other systems here are way ahead. The Adobe Media Encoder results were impressive, however, taking 121 seconds with CUDA acceleration enabled.

Although the Threadripper Pro has a solid 4.5GHz top single-core frequency, this clearly held it back when modelling. The SPECviewperf 2020 v3.1 results were excellent on an absolute scale but were beaten by every other system. The 2TB Samsung PM981ab was also the slowest NVMe drive here, delivering just 3,519MB/sec reading and 2,986MB/sec writing.

It’s hard to mark down the Lenovo P620 Tower for lagging behind in performance. It’s the cheapest system in this labs test by nearly £900 inc VAT and if upgraded to £4,500 would be far more competitive. It’s also a solid, well-built workstation. With the right specification for your money, it could be well worth considering.

We've also ranked the best monitors for dual-screen setups.

This review first appeared in issue 348 of PC Pro.

Armari’s lower-cost system is something of a technology showcase, exhibiting the latest options in processor, graphics and storage. The combination is one of the most powerful workstations you could buy for £4,500 inc VAT.

At the center of the Magnetar MC16R7 is AMD’s range-topping Ryzen 9 7950X. This potent 16-core processor uses AMD’s latest Zen 4 architecture and is manufactured on the 5nm process. This enables an incredible base clock of 4.5GHz, which is the boost clock for AMD Ryzen Threadripper Pro processors. The 7950X’s boost clock of 5.7GHz is only a few hundred megahertz behind the best Intel has to offer, and only with the latter’s P-cores, so it’s good to see that Armari makes the most out of the Ryzen 9 via its own customized CPU liquid cooling.

Armari has also taken full advantage of the fact that the AMD Ryzen 7000 series supports DDR5 memory by supplying 64GB of 6,000MHz RAM in two 32GB modules, leaving two DIMM slots free for upgrades. This is the fastest-clocked memory of any system this month.

The Magnetar MC16R7 showcases the latest CPU, graphics and storage technology (Image credit: Future)

So the Magnetar MC16R7 has a cutting-edge processor, some of the fastest system memory available, and its graphics acceleration is bleeding edge, too. In the past, choosing AMD professional GPUs might be a good choice to keep within a budget, but it rarely beat the Nvidia alternative for performance. The AMD Radeon Pro W7800 is a different matter. It’s in the same price category as the Nvidia RTX A5000 and offers 4,480 unified shaders (which aren’t equivalent to CUDA cores) on AMD’s latest RDNA 3 architecture. It also boasts 32GB of GDDR6 memory on a 256-bit bus, offering 576GB/sec bandwidth.

Armari is notable in the UK market because it’s one of the few local PC integrators that designs its own chassis. However, these cases come at a premium so the Magnetar MC16R7 has been built into a Fractal Design Meshify 2. This is still a great basis for a workstation, with plenty of room inside for airflow and storage upgrades. There are six spaces for 3.5in or 2.5in drives included, and there could optionally be up to 14. On top of this there are two 2.5in-only spaces as standard, but up to four are possible.

You may want to build upon the single M.2 NVMe SSD Armari supplies, but what a great foundation it provides. It’s a 2TB Crucial T700 drive, which supports PCI Express 5, as does the Asus ProArt B650-Creator motherboard. The Crucial SSD delivers incredible throughput from a single drive. CrystalDiskMark recorded sustained reading at 12,373MB/sec and writing at 11,807MB/sec, which were close to twice as fast as some of the PCI Express 4 NVMe SSDs in other workstations this month.

The Fractal Design Meshify 2 case offers lots of room for airflow and upgrades (Image credit: Future)

Considering all the powerful components in the Magnetar MC16R7, it’s no surprise that it produced some stunning test results. Our media-focused benchmarks are the Intel Core i9’s forte, but the Armari system’s overall result of 772 is still incredible, significantly beating the other system this month based on an AMD Ryzen 9 7950X. Its Cinebench R23 multithread rendering result of 38,611 was the fastest in the £4,500 category, and the Blender rendering time of 265 seconds was also top in this class. The OpenCL-accelerated Adobe Media Encoder time of 105 seconds beat every other system this month.

The AMD Radeon Pro W7800 graphics may be around the same price as Nvidia’s RTX A5000, but its performance with SPECviewperf 2020 v3.1 is in a different league as well. The results of 235 in 3dsmax-07 and an unbelievable 846 in maya-06 imply this will be a consummate accelerator for 3D animation. Likewise, 155 in catia-06, 235 in creo-03, 622 in snx-04 and 460 in solidworks-07 show strong abilities with product development, CAD and engineering.

Its LuxMark 3.1 result of 14,919 is a little behind the RTX A5000, but GPU rendering in Blender took just 141 seconds, which is ahead.

Overall, the Armari Magnetar MC16R7 provides the best possible performance for the money in most areas. If you need a powerful all-round workstation, this system should be top of your list.

We've also rated the best business computers.

The debate over the future of AI continues to rage, with battle lines being drawn between Big Tech’s domination of AI development and grassroots, community-driven innovation – a situation that will set the tone for how the industry develops in the years ahead.

In its recent history, AI has been increasingly dominated by a handful of tech giants, creating what many believe is a monoculture where just a few major players control access and progress.

However, new government policies and open source initiatives could help restore balance in the industry. The significance of open-source AI gained global attention in January during the ‘DeepSeek tech crash.’ While headlines focused on falling AI stock prices, the other big story was the immense potential of freely available AI models that can be used, modified and distributed without restrictions.

So, how might this play out, and what kind of AI ecosystem should the industry aim for to ensure the technology delivers for everyone?

Open source – a route to ethical AI innovation

Few would argue with the notion that open source and AI are natural partners. Both thrive on collaboration and shared innovation, allowing researchers and developers to collectively accelerate the pace of innovation while also ensuring broader access to emerging technologies.

Open source principles will be key to AI's future health, from collaborative improvements and knowledge pooling to transparency. Deeper integration on an industry-wide basis can create the conditions required for fairer, more ethical innovation. In doing so, the AI ecosystem can avoid problems associated with the vendor lock-in, limited transparency and restricted access to foundational models associated with uncompetitive technology markets.

On the other hand, if AI development remains concentrated within relatively few dominant players, users will undoubtedly face higher costs. They will be offered fewer opportunities to tailor solutions to their specific needs. This situation will only intensify as more AI is integrated into IT infrastructure.

That’s why finding a sustainable path for open source is crucial, especially as organizations today face massive internal and external pressure to become AI-powered. In an open source environment, AI tools will give organizations the power to deploy and manage workloads at the enterprise scale. In particular, they will allow businesses to select and integrate their preferred large language models (LLMs) while fully owning their data and infrastructure.

The rise of open source AI also means organizations can leverage the surge in new infrastructure software innovation to fuel and enable digital transformation. These emerging technologies create new levels of freedom and flexibility. As organizations will embrace these qualities, they drive further demand for reliable, secure and enterprise-ready solutions. In this context, AI/open source-led innovation becomes a virtuous circle.

The UK is the third-largest AI market in the world and the UK Government has recently shared the AI Opportunity Action Plan to boost economic growth and improve people’s everyday life.

The impact of open source across AI development and implementation is already impressive. A recent study by McKinsey found that open source AI is widely adopted by 63% of organizations overall and 72% in tech, particularly by organizations leveraging AI for competitive advantage (40% more likely to use it), with key benefits including lower costs (60%).

However, a democratic approach to AI , fostering open source-led innovation, will be essential for the success of a thriving AI ecosystem. The advancement of open technology with openness and choice for customers will be a corner-stone for creating resilient digital infrastructures

The role of regulation

Looking more closely at the growing role of open source AI models, however, progress is not simply a matter of accessibility or ideology; the industry also needs effective regulatory guardrails. While AI governance is still at a relatively early stage compared to other tech sectors, there are some clues as to the direction of travel. Last year, for example, the White House acknowledged the benefits of open source AI, such as fostering innovation and allowing a broader range of actors to examine and improve models.

To build on this sentiment, the ideal scenario would be to commit to AI regulations that ensure AI can be accessed, used, changed and shared transparently by anyone. Now more than ever, enterprises require sovereignty, control and the ability to customize AI implementations – all of which are best ensured through open source.

Additionally, community-governed development is the best way to reduce the possibility of flaws, drive innovation and leverage AI’s full potential. The Linux Foundation project AI & Data for example fosters collaboration and provides trusted AI & Data infrastructure, uniting developers, organizations, and users.

With open source AI, customers also have full visibility into the software code and development process, increasing trust and bolstering AI, which is still a nascent technology. Community-based open standards can also address AI safety concerns, not least because engaging the community is one of the best ways to improve security protocols and resilience.

Clearly, this is a highly complex and nuanced situation where competing interests and priorities will come under further scrutiny. For those interested in a future where AI innovation takes place in a fair and ethical environment, collaboration and open standards will be key.

Check out our list of the best small business software.

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

Japan's Nikkei 225 dipped more than 5% and other Asian shares also sank Wednesday as the latest set of U.S. tariffs, including a massive 104% levy on Chinese imports, took effect.

(Image credit: Ahn Young-joon)

More than $1 billion in funding for Cornell University and around $790 million for Northwestern University have been frozen over alleged civil rights violations at both schools, the White House says.

(Image credit: Ted Shaffrey)

The roof of an nightclub in the Dominican capital collapsed early Tuesday during a concert attended by politicians, athletes and others, with at least 98 people dead and 160 injured, authorities said.

(Image credit: Ricardo Hernandez)

An immigration judge in Louisiana has ordered the government to turn over its evidence against Mahmoud Khalil. She says she will rule Friday on whether the Pro-Palestinian activist can be deported.

(Image credit: Stephen Smith)

Here are the answers for The New York Times Mini Crossword for April 9.

I always enjoy a chance to mess with AI video generators. Even when they're terrible, they can be entertaining, and when they pull it off, they can be amazing. So, I was keen to play with Runway's new Gen-4 model.

The company boasted that the Gen-4 (and its smaller, faster sibling model, Gen-4 Turbo) can outperform the earlier Gen-3 model in quality and consistency. Gen-4 supposedly nails the idea that characters can and should look like themselves between scenes, along with more fluid motion and improved environmental physics.

It’s also supposed to be remarkably good at following directions. You give it a visual reference and some descriptive text, and it produces a video that resembles what you imagined. In fact, it sounded a lot like how OpenAI promotes its own AI video creator, Sora.

Though the videos Sora makes are usually gorgeous, they are also sometimes unreliable in quality. One scene might be perfect, and the next might have characters floating like ghosts or doors leading to nowhere.

Magic movie

Runway Gen-4 pitched itself as video magic, so I decided to test it with that in mind and see if I could make videos telling the story of a wizard. I cooked up a few ideas for a little fantasy trilogy starring a wandering wizard. I wanted the wizard to meet an elf princess and then chase her through magic portals. Then, when he encounters her again, she's disguised as a magical animal, and he transforms her back into a princess.

The goal wasn’t to create a blockbuster. I just wanted to see how far Gen-4 could stretch with minimal input. Not having any photos of real wizards, I took advantage of the newly upgraded ChatGPT image generator to create convincing still images. Sora may not be blowing up Hollywood, but I can't deny the quality of some of the pictures produced by ChatGPT. I made the first video, then used Runway's option to "fix" a seed so that the characters would look consistent in the videos. I pieced the three videos into a single film below with a short break between each.

AI Cinema

You can see it's not perfect. There are some odd object movements, and the consistent looks aren't perfect. Some background elements shimmered oddly, and I wouldn’t put these clips on a theater screen just yet. However, the actual movement, expression, and emotion of the characters felt surprisingly real.

And, I liked the iteration options, which didn't overwhelm me with too many manual options but also gave me enough control so that it felt like I was actively involved in the creation and not just pressing a button and praying for coherence.

Now, will it take down Sora and OpenAI's many professional filmmaker partners? No, certainly not right now. But I'd probably at least experiment with it if I were an amateur filmmaker who wanted a relatively cheap way to see what some of my ideas could look like. At least, before spending a ton of money on the people needed to actually make movies look and feel as powerful as my vision for a film.

And if I grew comfortable enough with it and good enough at using and manipulating the AI to get what I wanted from it every time, I might not even think about using Sora. You don't need to be a wizard to see that's the spell Runway is hoping to cast on its potential user base.

You might also like

• How AI can turn your home video into a Hollywood blockbuster
• This AI can turn your mundane video into a special effects spectacular
• Turn your selfie into an action star with this new AI image-to-video feature

Google surprised me when I first experimented with its NotebookLM AI tool. The AI-powered research assistant offers a relatively intuitive and focused setup to make custom guides on any topic you can name. It's now poised to become even more accessible with a mobile app soon.

The app should be exciting for anyone who works with large volumes of information or just has an idiosyncratic learning style. NotebookLM basically takes any text, including Google Docs, PDFs, and even YouTube videos that have a transcript, unifies it, and recontextualizes the result as a summary, study guide, or other presentation.

As good as NotebookLM's summarization and ability to synthesize notes are, its most memorable aspect is the Audio Overviews it can create. Audio Overviews basically turns all that information you uploaded into a podcast hosted by two AI characters. The synthetic conversation between the two AI hosts can actually sound like people who have read what you uploaded and want to discuss it.

At first, I thought it was a gimmick, but it's really just a reformatting of information like any other. If you’ve ever felt like you'd absorb details from books and other text better in a BBC panel discussion, it's a great option.

It's also the most obvious reason to anticipate the NotebookLM mobile app. It’s perfect for multitaskers, commuters, and anyone whose eyes are too tired to stare at another screen. I’ve used it to refresh myself on interview transcripts before writing articles and just as a fun way to explain a confusing topic. When the app hits smartphones, the ability to do this on the go may be huge for anyone who wants to absorb complex information but doesn’t always have the time or focus to read it.

Pocket NotebookLM

Of course, Audio Overview is just one part of what makes NotebookLM worth your attention. Unlike many AI assistants, which often operate in a vacuum or require you to explain everything from scratch, NotebookLM has context from the content you share specifically. Its abilities are many, from basic summaries and the ability to answer questions to creating timelines, brainstorming new ideas built on the information, and instant cross-references.

That's useful on a computer. As a mobile app, there's a lot more opportunity to exploit NotebookLM's tools. Sure, you'll have needed to upload the information, but if it's there, you can instantly pull up that quote you can't quite remember or fact about the person you're about to meet.

Of course, it’s not perfect. NotebookLM is only as smart as the information you give it. If your sources are a mess, it won’t magically fill in the blanks. It also can't replace real research. Sometimes, it misses nuance or answers too confidently when the source it pulls from is hedging. But the fact that it’s transparent about the answers helps compensate. And, at least it will always offer something to listen to on your next long commute.

You might also like

• I tried using AI to create the background music for a podcast, but I may stick to music libraries for now
• I used NoteBookLM to help with productivity - here’s 5 top tips to get the most from Google’s AI audio tool
• Google will turn those long documents into your next favorite podcast

President Trump's rival advisors Peter Navarro and Elon Musk are in a war of words over whether Tesla relies on imported parts. Musk has the facts on his side.

(Image credit: Suzanne Cordeiro)