Hace dos años, inicié la startup israelí CogniFiber. Ella apareció en los titulares con DeeplightSe trata de un cable de fibra óptica que puede “procesar algoritmos complejos dentro de la propia fibra antes de que la señal llegue a la estación”. En su momento, advertimos que esta tecnología no llegaría a los usuarios finales en un futuro próximo y que era poco probable que apareciera en portátiles o teléfonos inteligentes en un futuro próximo.
pero, eeNews incrustado Ahora informa sobre Oriole Networks, una startup con sede en el Reino Unido que utiliza la luz para un propósito diferente: crear redes eficientes de chips de inteligencia artificial.
Se dice que esta tecnología puede entrenar LLM hasta 100 veces más rápido que los métodos tradicionales y al mismo tiempo reducir significativamente el uso de energía. Esta investigación tiene como objetivo mitigar el creciente consumo de energía de los centros de datos impulsado por la rápida expansión de las cargas de trabajo de IA y la creciente demanda de alta potencia. -computación de rendimiento. .
20 años de investigación en fotónica
“Nuestra ambición es crear un ecosistema de redes ópticas que pueda remodelar la industria resolviendo los cuellos de botella existentes y permitiendo una mayor competencia en la capa de GPU, basándose en décadas de investigación, estamos allanando el camino para una IA más rápida y eficiente”, dijo James. Regan, director ejecutivo de Oriole Networks. Eficiente y más sostenible”.
Las raíces de la empresa se encuentran en la investigación de redes ópticas realizada por el University College London (UCL) y la propiedad intelectual única de Oriole se basa en el trabajo de los científicos fundadores, el profesor George Zervas, Alessandro Otieno y Joshua Benjamin.
La startup ya ha atraído la atención de varios inversores interesados en encontrar soluciones a las crecientes necesidades energéticas de la inteligencia artificial.
Con planes de lanzar sus productos en etapa inicial para 2025, Oriole Networks espera remodelar la infraestructura de IA haciéndola más rápida, más eficiente energéticamente y, en última instancia, más sostenible.
Suscríbase al boletín informativo TechRadar Pro para recibir las principales noticias, opiniones, características y orientación que su empresa necesita para tener éxito.
eeNews incrustado “La aplicación de 20 años de profunda investigación y aprendizaje en fotónica para crear una mejor infraestructura de IA demuestra cuánta innovación puede ayudar a cosechar los beneficios”, cita Ian Hogarth, miembro de la junta directiva de Oriole y socio de Plural que dirigió la última ronda de financiación. tecnología.
“El equipo detrás de Oriole Networks ha demostrado su experiencia tanto en la construcción de la empresa como en la introducción de ciencia profunda en el mercado, creando un cambio fundamental en el diseño de sistemas de red de próxima generación que reducirán la latencia y el impacto energético de los centros de datos que ahora tenemos. depender.”
Samsung Networks and O2 Telefonica have announced the launch of the first commercial site in Germany that uses Open RAN and vRAN technologies. These companies have used these technologies to make O2 Telefonica’s 4G and 5G networks more reliable. They have been testing these new technologies since October 2023.
Samsung’s Open RAN and vRAN technologies power O2 Telefonica’s faster and more reliable 4G and 5G networks
The Open RAN and vRAN technologies have been deployed in Landsberg am Lech, Bavaria. This is the first time Samsung Networks’ Open RAN and vRAN technologies have been used in Germany. With these technologies, the two companies will start upgrading more cellular towers in Germany. This site went online just three months after Samsung Networks shipped the required hardware to O2 Telefonica. The hardware includes 4G and 5G vRAN 3.0 and O-RAN compliant radios that support low- and mid-bands (700MHz, 800MHz, 1.8GHz, 2.1GHz, 2.6GHz and 3.6GHz), and 64T64R Massive MIMO radios.
Open RAN improves the flexibility of a cellular network operator’s site, allowing the operator to use hardware and software from different vendors. The vRAN technology brings cloud-native architecture, allowing network operators to implement automation techniques better. It also allows firms to introduce new services and technologies more efficiently and quickly. This allows them to accelerate network buildouts and adopt new 5G applications. For example, they can create an instance for an AR/VR application with low response times.
Companies will now use Samsung’s network automation solutions to control the life cycle management of their networks, from deployment to operation and maintenance.
Junehee Lee, EVP and Head of Global Sales & Marketing at Samsung Networks, said, “Samsung is setting new standards for excellence in the telecommunications industry with our innovative vRAN and Open RAN capabilities. Celebrating Telefónica’s 100th anniversary, we are proud to be the key partner for O2 Telefónica on their trailblazing journey to usher in a new era of connectivity in Germany.”
The constant rush to adopt new technologies can quickly have an adverse effect on business networks – hindering bandwidth, connectivity, and eventually their bottom line. To overcome this problem, forward-thinking companies are now embracing the network edge. But what exactly is the edge? And how can it improve your operations without compromising latency, performance, or security?
Understanding edge architecture
In general terms, the edge of a network is the boundary between your business network and the public internet. It’s the zone where traffic from the internet enters your private network and vice versa. This zone is of critical importance because it defines your network security boundary and is the first line of defense against threats from the public internet. This includes firewalls, routers, switches, sensors, and other access points that exist at the boundary of networks.
Your network edge exists wherever there is an edge device or ‘thing’ that facilitates traffic crossing from one network to another. These edge devices can include mobile phones, laptops, data centers, company offices, home routers, and all manner of emerging applications that will become more commonplace as more users make up the rise of ‘the internet of things’.
Of course, the sheer volume of data produced by these multiple devices can seriously hinder a company’s IT infrastructure, which can only process so much data without problems arising with bandwidth and lag time.
That’s where edge networking comes in. This is a data management system that supplies digital resources to your users by merging edge devices with your business network to process and interpret data close to the source without the need for huge data centers.
Consequently, moving to the edge has become a key investment for businesses looking to deploy innovative technologies such as generative AI because they can now process data where it happens – at the boundary of edge devices – as opposed to in the cloud where latency and security issues abound. This not only helps to accelerate business decisions but also gathers valuable information at the point of origin to improve both product offerings and customer experiences.
What’s more, businesses can vastly reduce data transfers over external networks, minimizing the threat of data breaches while maximizing bandwidth usage, and ultimately improving customer outcomes.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Paul Gampe
Chief Technology Officer, Console Connect.
Benefits and drawbacks
A well-configured edge network can significantly improve the performance of applications that are hosted in the cloud. By placing edge devices closer to the users of these cloud-based applications, the distance that data must travel is greatly reduced, which means lower latency.
By placing security controls closer to this cloud-hosted data, it also reduces the risk of your private network being compromised by outside attacks. Furthermore, a well-configured edge network can overcome congestion problems that might occur with devices and applications by rerouting traffic to the correct destinations and ensuring that your network is not overloaded.
Then there’s the cost savings. By deploying devices on the edge and thereby offloading the processing and storage requirements to the edge, businesses can reduce the costs they would typically incur to purchase from cloud services.
That said, not all edge networks are created equal, and a poorly configured edge can turn all these benefits into disadvantages.
First and foremost are the security risks. Businesses need to make sure their edge is ultra-secure and can respond to threat incidents emerging from different devices or locations. Secondly, without the correct knowledge and training, your IT team may quickly be overwhelmed by having to manage the multitude of devices, locations, and applications that make up the edge.
A third challenge is the upfront cost. Implementing edge networking into your business can be expensive as it requires the purchase and deployment of suitable devices and systems that achieve the desired aims without sacrificing security.
Enter the automated edge
Unsurprisingly, the technology behind edge networking is evolving all the time, and a recent advancement is the automated edge. In simplified terms, the automated edge is intelligence that is embedded into edge devices which takes care of networking tasks like configuration, responsiveness, provisioning, and even monitoring all by itself.
As your daily operations shift and change, it will automatically optimize network performance by intelligently steering traffic through different pathways to reduce lag time and interruptions while providing greater reliability and security.
Best of all, the automated edge can work hand-in-glove with your generative AI models to improve connectivity and real-time decision making from different devices, locations, and applications. This not only accelerates the analysis of your data close to the source, but also speeds up customer interactions and provides further opportunities for growth and innovation.
Additionally, if your business relies heavily on low latency – which most businesses do – then your IT team will no longer have to constantly manage any blockages in your network and then figure out the best solution. The automated edge does it for you.
The simple solution to edge automation
The sheer complexity of the automated edge has led many companies to seek out partnerships with network-as-a-service providers to give them a fast, flexible, and competitive advantage. Network as a service, or NaaS for short, provides companies with an efficient end-to-end solution in the operational management and performance of their networks on a subscription basis.
The beauty of this business model is that removes all the complexity. A NaaS provider helps you to build your own private network on your own terms, one that can be scaled and configured to meet your needs. It also takes care of security issues and protects your network from attacks through constant network monitoring and early detection.
But crucially, a tried and tested NaaS provider will have already embedded their platform with the latest automated edge capabilities to interact and interface with all manner of edge devices and applications. This not only allows you to deploy new technology tools safely and efficiently, but also frees up your IT team to concentrate on other mission-critical tasks.
As for network management, your business will have total visibility of your network infrastructure from a single user-friendly dashboard, allowing you switch-up performance and agility on demand. Meanwhile, all your network maintenance concerns are taken care of 24/7 by an expert team of engineers.
If businesses truly want to harness the potential of ‘the internet of things’, they must first consider moving to the edge. Indeed, by laying the foundational groundwork today for what is surely to come tomorrow, they can take full advantage of future advancements in network edge innovation.
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
The term ‘Metaverse’ originated in Neal Stephenson’s 1992 novel, Snow Crash, where it depicted a virtual realm offering an escape from reality for characters. Stephenson’s visionary concept of the Metaverse has since transitioned from fiction to (virtual) reality over the past decade. Augmented and virtual reality technologies have advanced the concept of the Metaverse, now capable of manipulating the real world and immersing users in digital experiences.
This year marked Apple‘s eagerly anticipated debut into this dynamic technology with the launch of its Apple Vision Pro. Apple refers to its brand of AR and VR metaverse technology as “spatial computing”, that “seamlessly blends digital content with the physical world”, using hand and eye tracking for a fluid digital experience. The technology, coming in with a hefty price tag of $3,499, hasn’t been enough to put off Apple enthusiasts, with analysts suggesting that initial sales have been between 160,000 and 200,000.
The price tag has contributed to mixed reviews of the Apple Vision Pro, which has seen it hailed as “the best consumer headset anyone’s ever made” but also “very expensive” and “inconsistent”. While the headset is marveled as a technological breakthrough, there’s one thing that is assumed in the performance of the technology: that the network connecting it to the internet is reliable and of high performance.
For Apple’s Vision Pro and other similar devices to have mainstream usage, our foundational networks must guarantee ultra-fast, low-latency connectivity. This is essential to accommodate the significant increase in traffic and to deliver seamless immersive experiences.
Jurgen Hatheier
International CTO, Ciena.
Reliance on connectivity
As Analysys Mason points out: “At its core, a Vision Pro could be used without internet connectivity for on-device games, content and experiences, but its primary use-case is for connected applications.” Using it without connectivity is the equivalent of using a phone, and the applications on it, without being connected to the internet.
The Apple Vision Pro is designed to be used in a “controlled environment”, so ideally at home, but there are already examples of people using it outside of the home. Features on the Apple Vision such as spatial video and spatial video capture, immersive FaceTime, and HD video playback all demand high-performance Wi-Fi connectivity – requiring fast speeds, low-latency and high-bandwidth internet connectivity, that can only be delivered by fiber broadband. Broadband speeds of 400 Mbps could be required for some AR and VR applications.
However, approximately 44% of households in the US, where the Vision Pro has been launched first, don’t have access to fiber connectivity. This would likely make this technology useless to this large swathe of the population. Currently, the Federal Communications Commission (FCC) mandates that consumers must have access to minimum actual download speeds of 25 Mbps and minimum actual upload speeds of 3 Mbps. However, this level of home network performance falls short of meeting the demands of an immersive VR environment, particularly when other devices are simultaneously in use.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
In comparison, 10Gbps broadband speeds are more widely accessible across the APAC region, with fiber broadband accounts representing 85% of all residential fixed-line broadband subscriptions. This could enhance the headset’s appeal to consumers in countries such as China, India, and Korea. However, as of now, there is only speculation regarding the product’s launch timeline in these markets.
There are ongoing discussions about revising the definition of minimum broadband speeds to 100 Mbps. Nonetheless, researchers estimate that achieving the ultimate immersive media experience may necessitate anywhere between two and five Gbps, indicating that even this proposed increase may not suffice.
Connectivity for the future
CSPs are aware that they need to cater to the demands of end users who are using technology requiring high-performance internet. They are working to meet these demands, investing to make their networks faster, and smarter, as well as to bring processing closer to the end user.
At the network level, CSPs are starting to use technology such as 25G Passive Optical Network (PON) technology capable of delivering network speeds of 25 gigabits per second (Gbps) to premises. Across metro and long-haul fiber networks – those that connect data centers across or between nations – they are using introducing technology moving 1.6 terabits per second per wavelength. Networks that use this technology are built for the future, able to carry huge amounts of traffic and support gigabit speeds and low latency.
Finally, if we see demand for ‘spatial technology’ like the Apple Vision Pro grow beyond the confines of home use, then 5G connectivity for these devices will become a necessity. Apple’s headset is currently tethered to WI-FI but for AR to work out in the wild, only 5G will be able to deliver the high speeds required. 5G Standalone (SA) in particular could prove increasingly important here due to its highly reliable, low-latency connectivity. Delivering such a seamless, immersive experience on the go could prove a key driver for 5G SA adoption, which is currently more in demand for business and industry use cases rather than consumer.
The Apple Vision Pro is an exciting, well, vision, of what the future could hold for AR and VR metaverse technology. Its commercial launch is an interesting acid test not just for how ready and willing consumers are for these immersive (and, currently expensive) experiences but how ready networks are to deliver them. The hardware will improve over time, but the networks they rely on will need to keep up too, or Apple’s vision could quickly falter.
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
Firewalla makes configurable hardware firewalls that connect to your router, providing protection for your home or business against various network and internet threats.
The company has announced the pre-sale of Firewalla Gold Pro, the newest and most powerful addition to the “Gold” product line. Touted as the world’s most affordable 10-gigabit smart firewall, this device is designed to be compatible with the next-generation Wi-Fi 7 and high-speed 5 and 10-gigabit ISP fiber networks.
The Gold Pro features two 10-gigabit and two 2.5-gigabit Ethernet interfaces, which provide network segmentation and redundancy. The device is powered by a quad-core Intel processor and 8GB of RAM, allowing it to scale with growing networks.
Available to pre-order
The 10-gigabit ports can be used for both WAN and LAN and users can segment their network with VLANs running at 10 gigabits or connect one port to a 10-gigabit Wi-Fi 7 access point and another to a high-speed switch. The firewall supports VPNs at speeds over 2GB, ensuring a fast and secure network experience, even on the go.
“Our dedicated community is always pushing for a better network at higher speeds. The Gold Pro makes it possible to protect their homes and offices at future speeds, without monthly fees,” said Jerry Chen, founder of Firewalla. “As small businesses upgrade their infrastructure and consumers adopt faster offerings from their ISPs, the Gold Pro gives them unmatched visibility and protection for the next generation of networks.”
The Firewalla Gold Pro is currently available for pre-sale at a price of $789 with early buyers receiving an additional six months of warranty. Shipping is expected to start in early November, with a price increase likely closer to the shipping date.
More from TechRadar Pro
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
Thanks to the incredible advancements in neural networks and language processing computers can understand and respond to human language just as another person might. The journey from the first moments of doubt to the current state of achievement is a tale of relentless innovation and discovery. The Art of the Problem YouTube channel has created a fantastic history documenting the 30 year journey that has brought us to ChatGPT-4 and other AI models.
Back in the 1980s, the notion that machines could grasp the nuances of human language was met with skepticism. Yet, the evolution of neural networks from basic, single-purpose systems to intricate, versatile models has been nothing short of remarkable. A pivotal moment came in 1986 when Michael I. Jordan introduced recurrent neural networks (RNNs). These networks had memory cells that could learn sequences, which is crucial for language understanding.
The early 1990s saw Jeffrey Elman’s experiments, which showed that neural networks could figure out word boundaries and group words by meaning without being directly told to do so. This discovery was a huge step forward, suggesting that neural networks might be able to decode language structures on their own.
How Neural Networks learned to talk
Here are some other articles you may find of interest on the subject of neural networks :
As we moved into the 2010s, the push for larger neural networks led to improved language prediction and generation abilities. These sophisticated models could sift through massive data sets, learning from context and experience, much like how humans learn.
Then, in 2017, the Transformer architecture came onto the scene. This new method used self-attention layers to handle sequences all at once, effectively overcoming the memory constraints of RNNs. The Transformer model was the foundation for the Generative Pretrained Transformer (GPT) models.
GPT models are known for their incredible ability to learn without specific examples, following instructions and performing tasks they haven’t been directly trained on. This was a huge leap forward in AI, showing a level of adaptability and understanding that was once thought impossible.
ChatGPT, a variant of these models, became a tool that many people could use, allowing them to interact with an advanced language model. Its ability to hold conversations that feel human has been impressive, indicating the enormous potential of these technologies.
One of the latest breakthroughs is in-context learning. This allows models like ChatGPT to take in new information while they’re being used, adapting to new situations without changing their underlying structure. This is similar to how humans learn, with context playing a vital role in understanding and using new knowledge.
However, the rapid progress has sparked a debate among AI experts. Are these models truly understanding language, or are they just simulating comprehension? This question is at the heart of discussions among professionals in the field.
Looking ahead, the potential for large language models to act as the basis for a new type of operating system is significant. They could transform tasks that computers typically handle, marking a new era of how humans interact with machines.
The road from initial doubt to today’s advanced language models has been long and filled with breakthroughs. The progress of neural networks has transformed language processing and paved the way for a future where computers might engage with human language in ways we never thought possible. The transformative impact of these technologies continues to reshape our world, with the promise of even more astounding advancements on the horizon.
Filed Under: Technology News, Top News
Latest timeswonderful Deals
Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, timeswonderful may earn an affiliate commission. Learn about our Disclosure Policy.
