Administrators said the business had been hit by `challenging trading conditions and increased costs`. Read...

TL;DR: This practical ChatGPT & Automation E-Degree is now available for just $19.97 (MSRP: $790) through November 2. Artificial intelligence isn’t just a buzzword anymore—it’s a competitive necessity. For professionals across industries, knowing how to use AI tools like ChatGPT isn’t optional. The ChatGPT & Automation E-Degree, now available for just $19.97 (MSRP: $790), offers a practical, hands-on way to understand and implement AI in your workflows. The program comprises 12 courses and over 25 hours of content, all developed by Eduonix Learning Solutions, a trusted name in professional training. Instead of broad, abstract lessons, you’ll find real-world applications you can bring directly into your business. Here’s what makes it useful: AI for Business Processes: Learn how to use automation to streamline things like reporting, customer service, and scheduling. ChatGPT for Productivity: Master prompt-building to generate marketing copy, draft emails, and analyze data. Data Visualization & Storytelling: Turn raw data into presentations your clients and teams will actually understand. Coding & Customization: Explore the technical side of tailoring AI tools for your specific industry. Cross-Industry Use Cases: From law and finance to retail and startups, discover how AI can fit your field. Think of it as a low-cost investment with a big payoff. Get lifetime access to these ChatGPT & Automation E-Degree courses while it’s still on sale for just $19.97 (MSRP: $790) through November 2. ChatGPT & Automation E-DegreeSee Deal StackSocial prices subject to change. Read...

Gartner reveals key tech trends for 2026, spotlighting AI supercomputing, cybersecurity, cloud sovereignty, and digital provenance shaping business futures. Read...

As malicious generative AI tools surge, businesses must bolster defences with strategic security measures to counter evolving AI-enabled cyber threats. Read...

Ruling holds that defeating end-to-end encryption in WhatsApp harms Meta`s business. Read...

A Mount Maunganui business is aiming to use undersea drones to help protect seafloor cables from sabotage as well as other work. Read...

If you’re in the market for a Thunderbolt docking station and you’re biding your time until Black Friday, know this: smart shoppers can save by buying a slightly older Thunderbolt 4 or USB4 dock instead. Buying tech always comes with a choice, as prices of new technology climb higher. You can either buy the latest and greatest and insulate yourself from change for a few years or buy used and save a few bucks. For 2025 and 2026 docking stations, the latter choice appears to be the right one. Docking stations have been in a transitional period for the last few years. Thunderbolt 3 / USB3 and Thunderbolt 4 / USB4 are almost identical in terms of their bandwidth, which is still one of the defining features of a docking station. In the real world, the 40Gbps bandwidth both standards offer are enough to connect 4K displays running 60Hz. Thunderbolt 5 was the next big thing, with 80Gbps (and in certain scenarios, 120Gbps) available for those 4K displays to run at 144Hz, which is good enough for gaming. Here’s the key point that you need to know: Intel’s current laptop platform, the Core Ultra Series 2 (Lunar Lake) as well as Qualcomm’s Snapdragon X Elite chips, only support USB4 and Thunderbolt 4. (Thunderbolt 4 and USB4 are also functionally identical; in this case, Intel won’t certify a non-Intel chip as a “Thunderbolt” product, so USB4 is the generic equivalent.) And, somewhat shockingly, the next-generation of 2026 laptop platforms, the Intel Core Ultra Series 3 (Panther Lake) and the Qualcomm Snapdragon X2 Elite, don’t include integrated Thunderbolt 5 support, either. No integrated Thunderbolt 5 chip support means very little Thunderbolt 5 laptop support, which means very little reason to buy a Thunderbolt 5 docking station. Sound familiar? It should. I was expecting Thunderbolt 5 to make a decent showing at CES 2025. Nope. Intel Intel does sell a discrete Thunderbolt 5 chip that costs extra, and because of this you’ll see Thunderbolt 5 ports appear on expensive gaming laptops like the Razer Blade 18. But they’re not really necessary. The Blade 18 includes a stellar 4K+ display that actually exceeds Thunderbolt 5’s capabilities (3840×2400, 240Hz). Why would you want to use an external display when the laptop’s built-in display is even better? Not all gaming laptops include the same quality of display, but the argument generally holds true: gaming laptops with Thunderbolt 5 ports generally offer superior hardware right now. A Thunderbolt 5 dock isn’t cheap, either. Thunderbolt 5 docking station manufacturers tend to charge roughly $100 more for one. Plugable’s TBT-UDT3 Thunderbolt 5 dock is reasonably priced at $300 — it is! — but our top pick for the best Thunderbolt docking station, the Kensington Thunderbolt 4 Dual 4K Dock (SD5780T, is $185 at press time. Yes, there are some things to consider. One, AMD hasn’t announced its next-gen laptop platform for 2026, and the 120Gbps version of USB (USB4 V2) could certainly make an appearance. Two, docks like the Sonnetech Echo 13, a Thunderbolt 5 dock with an integrated SSD, provide almost-but-not-quite internal-SSD speeds for external storage. An issue with any gaming PC is simply the available storage, and being able to connect an external SSD with gaming-capable speeds is an enormous plus. Thunderbolt 5 docks like Sonnet’s deliver. Three, Thunderbolt 5 supports external GPUs, and we might see a forthcoming notebook designed with a Thunderbolt 5 port to connect to an external GPU dock like the Razer Core X V2. Business during the day, gaming after hours: that’s part of the promise of Thunderbolt 5. But times are not easy, and the ongoing economy and tariff wars mean people are going to want to save money during Black Friday and other holiday sales. If you agree, don’t plan on spending big on a Thunderbolt 5 dock. PCWorld plans to monitor e-tailers for any ongoing sales of Thunderbolt and USB4 docks for laptops, and we’ll be prioritizing Thunderbolt 4 and its cost-effective competition, DisplayLink. You should, too. Read...

Home appliance brand Kohler is getting into the smart camera business, although its new camera isn’t designed to catch crooks in the act. Instead, Kohler’s smart cam is aimed at something more personal: your poop.  Yes, the jokes are irresistible, but Kohler Health’s Dekoda camera does have a serious purpose: to analyze your stool and give you insights about your gut health and other potential health issues.  The Dekoda itself is a small (8.13 x 3.74 x 3.43-inch), flat, oval-shaped unit that clips onto the side of your toilet, allowing an optical sensor to take aim at the bottom of your toilet bowl.  You can then go ahead and use the toilet normally—and if you’re worried about privacy (which seems reasonably given what the camera is looking at), you can log in prior to each “session” using the Kohler Health app or a fingerprint scanner on the included authenticator remote.   The images snapped by the Dekoda are protected by end-to-end encryption, Kohler says, and the optical sensor always points down into the toilet, not toward your nether regions.  The Dakota’s optical sensor always points down into the toilet.Kohler Health Powered by an AI model based on the Bristol Stool Chart and trained on more than a million data points, the Dekoda can track your bowel movements, looking for “gut health indicators” such as the shape and consistency of your stool, your overall hydration, and how frequently you go. The sensor also keeps an eye out for blood in the toilet, a key warning sign of a serious health problem.  Data collected by the Dekoda will be funneled into the Kohler Hearth app, which can show you the results for each session (including the quantity of regular, hard, and loose stool), as well as your stool frequency and even a gut health score, expressed as a percentage. The app can track your gut health results over time, as well.  The Dekoda connects to your home network via Wi-Fi, and it’s rechargeable (a power adapter and USB cable are included).   At $599 and available now for preorder (it’s slated to ship this month), Kohler Health’s Dekoda isn’t cheap, and it also requires a subscription plan that costs either $7 a month or $70 for an entire year.  But for those of us with occasional gut issues or anyone who wants to dig deeper into their health (and yes, this is quite the deep dive), the Dekoda could deliver some fascinating insights. Read...

Microsoft has introduced a new device category with Copilot+. Only laptops with a dedicated Neural Processing Unit (NPU), at least 16 GB of RAM and a fast NVMe SSD fulfil the minimum requirements. Intel is addressing these requirements with the Core Ultra series, which combines classic CPU cores with GPU acceleration and a hardwired NPU. This unit alone achieves 40 to 47 trillion computing operations per second – a device is only officially certified as a Copilot+ PC once it reaches this performance level. Put simply, if you want to use AI functions locally, you need a very fast computer with a chip specially optimized for AI. Core Ultra processors with local AI acceleration For modern AI applications, it is no longer just the clock rate that is decisive, but the internal distribution of tasks between the computing units. The Intel Core Ultra 7 268V has eight CPU cores and an integrated Intel Arc GPU. The set-up is complemented by the NPU, which is specially designed for inference-based AI tasks. This allows text analyses, translations, video filters, image generation or language modelling to be carried out directly on the chip – without any detours via the cloud. The Intel NPU cannot be configured separately; its activity is controlled via Windows components. Functions such as Recall, Cocreator or Windows Studio Effects access it automatically; there is no provision for manual assignment. Intel Copilot functions between benefit and control The central innovation in Copilot is not the graphical user interface, but the shifting of AI processing to the device. The Recall feature continuously creates screenshots of all activities. These can be searched through later using a text search. The feature works locally, stores content in encrypted form and can be deactivated for specific applications. However, users have no insight into the data model used. Chris Hoffman / Foundry The Live Captions function covers a special use case. It transcribes any audio signals in real time, regardless of whether they originate from a video, a video call or a locally played file. The Intel NPU takes over the continuous speech-to-text mapping at system level. The output takes place as an overlay directly above the respective window or as a separate bar at the bottom of the screen. Unlike conventional subtitles, Live Captions is speech-agnostic, automatically recognizing speech input and converting it into readable text on-the-fly. No internet connection is required. Translations can also be activated so that English-language content, for example, is automatically subtitled in German. If the NPU performance is sufficient, the delay remains minimal. The function can be used system-wide, which creates real added value, especially for people with impaired hearing or in noisy environments. Program-specific activation is not necessary. As soon as Live Captions are activated, the system analyses all sound sources. However, use is limited to the visible transcript. There is no storage or analysis beyond the current session. Chris Hoffman / IDG Local image synthesis with Paint Cocreator The Cocreator function integrated in Paint makes it possible to create images from simple text input within a few seconds. This requires a Copilot-certified device with an active NPU. The underlying diffusion model runs locally and utilizes both the main memory and the dedicated NPU for image generation. In contrast to cloud services, the entire process remains on the device. A short text-based prompt is sufficient to receive several image suggestions in low to medium resolution, which can then be edited directly in Paint. Users can influence the style, color scheme and complexity of the motif – but only within the specified limits. The model itself remains a black box; external models cannot be integrated. Without an NPU, Cocreator is hidden or refuses to start due to insufficient system resources. Mark Hachman / IDG In addition to pure computing power, Intel Core Ultra notebooks have specialized technologies to optimize the use of AI in mobile operation. The so-called Dynamic Tuning Technology analyses temperature, usage behavior and energy profile in real time and automatically adjusts the distribution of loads to the CPU, GPU and NPU. This is complemented by the Intelligent Display, which controls screen brightness, contrast and refresh rate in a context-sensitive manner, an aspect that can bring tangible benefits in terms of battery life in battery mode. This technology also proves useful for longer video conferences or remote working scenarios, as it relieves the thermal management and optimizes visibility at the same time. The NPU also benefits from the fact that the reduced GPU load creates more thermal headroom for inference tasks. Inference-based processes refer to tasks in which an AI makes decisions or delivers results based on an already trained model. In other words, the actual learning, i.e. the training, has already been completed. During inference, this knowledge is used to process new input. One example would be the automatic subtitling of videos using live captions. The AI has previously been trained with huge language data sets. During inference, it recognizes what someone is saying in real time, converts speech into text and displays it directly on the screen. To do this, it does not reuse the entire training material, but instead uses compact, optimized models. Crucially, inference processes can be executed locally on the device, provided a dedicated accelerator such as an NPU is available. This means that data remains on the notebook and AI reactions are almost instantaneous. Energy efficiency and system behavior under load One advantage of the Intel architecture is the thermal separation of the computing units. CPU, GPU and NPU do not compete directly for power reserves. The NPU only handles inference-based processes. This significantly reduces power consumption when AI features are active. Tests with the Core Ultra 9 285K show up to 25 per cent better efficiency compared to the Core i9 14900K. The battery life benefits noticeably from this shift. Systems such as the Lenovo Yoga Slim 7i Aura Edition achieve up to 22 hours of video playback without any loss of performance in AI processes. The decisive factor here is the continuous support by BIOS and Windows. Without the latest UEFI update, the NPU is often underpowered. In addition to the Lenovo Yoga Slim, other manufacturers already offer devices that are explicitly suitable for local AI applications under Windows 11. The Acer Swift Go 14 relies on the Intel Core Ultra 7 155H, combined with an NPU that delivers over 40 trillion operations per second. Thomas Joos It is worth taking a look at the specific device basis, as not every laptop advertised as a “Copilot” fulfils the same requirements. Notebooks with Intel Evo certification and Core Ultra 200V processors undergo a validation program that is tailored to real-life usage scenarios. Not only high performance and long battery life are required, but also fast reactivation, smooth multi-monitor connection via Thunderbolt 4 and consistent behavior when using several applications at the same time. For Copilot, this means that functions such as Recall, Live Captions and Cocreator run without any loss of performance, even when switching between mains and battery operation. Evo-certified devices such as the Acer Swift Go or the Surface models from the business segment are thus exemplary for this new class of hybrid AI PCs, which combine everyday usability and technical sophistication. YouTube / Microsoft Intel divides the Core Ultra Series 2 into several model lines that have a direct impact on the field of application of mobile systems. The V models serve as a reference for slim premium devices and impress with their high efficiency and strong NPU performance. H models are aimed at classic high-end notebooks for productive applications with medium mobility requirements. U models are optimized for lightweight, ultra-mobile systems with reduced energy requirements and AI functions. HX models represent the upper performance spectrum with up to 24 physical cores and are primarily intended for workstations, gaming laptops and professional creative environments. V and H models are particularly suitable for everyday use in AI-supported Windows applications, as they offer a good balance between battery life, thermal budget and inference performance of the NPU. What Intel NPUs can do and what is not visible The practical use of the NPU is tied to the respective features in Windows. Independent programming or execution of user-defined models is not possible with on-board resources. Intel does provide developers with APIs, but in everyday life these paths are not accessible to users without programming knowledge. The visibility of NPU utilization is limited to indicators in the task manager or via third-party tools. For power users, it remains unclear which processes are running on which unit and when. Control remains in the hands of the operating system. A separation of CPU and NPU responsibilities can at best be indirectly tracked via utilization profiles. Thomas Joos Intel offers additional functions for corporate use with the vPro platform based on the Core Ultra series. In addition to the integration of security-relevant features such as hardware-based identity verification, the NPU can also be used directly by security software to analyze threats locally and detect behavioral anomalies at an early stage. In addition, vPro Device Discovery supports detailed queries on installed components, energy profiles and configuration states. In combination with Copilot functions, this creates devices that offer a high degree of transparency and control for both IT administrators and users in regulated industries – without direct access to the underlying models. Bottom line With an Intel Core Ultra 200V series device, you gain access to a clearly defined group of AI functions. These run locally, do not require an internet connection and benefit directly from the NPU. However, control over computing execution remains limited: Windows automatically determines which unit becomes active, while the setting options only allow rudimentary intervention. For productive use, this primarily means a gain in efficiency – but not deeper system control. The new device class is therefore more of a platform than a toolbox. Those looking for precise AI customizations will quickly reach their limits. On the other hand, those who rely on stable, locally executed features will find a convincing hardware basis in the Intel-based Copilot devices. Read...

TL;DR: Access dozens of top AI tools in one platform — 1min.AI bundles content, chat, design, audio, video, PDF, and more under a single lifetime license for just $79.99 with code SAVE20. One of the bigger annoyances of the digital age is the subscription model. Juggling a half-dozen AI tools, each with its own login credentials, pricing tiers, and learning curve, is exhausting. That’s why 1min.AI can be a helpful alternative to the usual chaos. This all-in-one platform brings together top-tier AI features for writing, design, video, audio, and more under a single dashboard. And you can get a lifetime subscription to the Advanced Business Plan for just $79.99 (down from the MSRP of $540) — with no recurring fees, ever. Need blog posts written in your brand voice? Check. Want to generate YouTube thumbnails, edit PDFs with AI, or even clean up audio? Covered. And 1min.AI does it fast — like, one-minute fast. You’ll have access to multiple flagship models like GPT, Claude, Gemini, and Llama, plus unlimited brand voice slots and unlimited prompt storage. If you’re tired of managing a spreadsheet of AI tools (we’ve been there), this is your chance to condense it all. Get lifetime access to the 1min.AI Advanced Business Plan for just $79.99 with code SAVE20 through November 2. 1min.AI Advanced Business Plan Lifetime SubscriptionSee Deal StackSocial prices subject to change. Read...