Elad Blatt

5G + Edge + GenAI = New Business Model opportunities Meta has released Llama 3.2 lightweight models in 1B and 3B sizes available on #watsonx --> advantages: > Small size is perfect for mobile and edge devices or CPU only inferencing. > Optimized for ARM based processors, support Qualcomm and MediaTek hardware from day one (covering majority of mobile processors out in use). > Enhanced Privacy - No need to send sensitive data such as messages and calendar invite to the cloud, maintaining user privacy. > Supports up to 128K tokens and optimised for summarisation, rewriting, instruction following, language reasoning, and tool use.  > Tool Calling: The 1B and 3B models excel in tool calling, enabling developers to create personalized, on-device applications. > Instantaneous Responses: Local processing makes prompts and responses feel instantaneous. > Tiny size achieved through network pruning and knowledge distillation using larger Llama 3.1 > Post-training step uses supervised fine-tuning (SFT), rejection sampling (RS), and direct preference optimization (DPO) for alignment. Example Use Case: Integrated Applications that could help summarize the last 10 messages received, extract action items, and leverage tool/function calling to directly send calendar invites for follow-up meetings. RIGHT MODEL --> RIGHT USE CASE --> RIGHT GOVERNANCE --> Success !   

25

The buzz for #SnapdragonXSeries exclusively powering Microsoft’s Copilot+ PC is here to stay… learn more about how Qualcomm transformed the PC experience:

24

Impressed by @NVIDIA’s innovation in AI energy efficiency and proud that PureNodal is leveraging their GPUs to power our own sustainable cloud infrastructure! By leveraging NVIDIA’s groundbreaking GPUs, we’re able to deliver 50% lower power consumption and nearly zero water usage – all while providing the high-performance cloud infrastructure needed for the next generation of AI workloads. This is a great example of how innovation in hardware can drive real impact on sustainability, and we’re excited to be part of this movement. Check out the full story from NVIDIA and visit us at http://purenodal.com to learn more about how PureNodal is delivering more sustainable cloud solutions for AI. #Sustainability #AI #CloudInnovation #EnergyEfficiency #PureNodal #DataCenters #GreenTech

23

1 Comment

As #AI scales across industries and use cases, understanding how to measure AI performance becomes key in developing solutions with sufficient head room. Like we say, "not all TOPs are created equal". At Qualcomm, we focus on running AI inferencing workloads on our super-efficient and powerful Neural Processing Units (NPUs). Highly recommend reading our latest article on how to measure AI workloads, performance and the right place to run them.  https://lnkd.in/gzV33PCa

75

NVIDIA's new private 5G network at their Santa Clara HQ is a great step forward for 5G Standalone (5G SA) implementation and evolution. While large carriers have hesitated to lean into 5G SA, innovators like NVIDIA, building the underlying technology and initial test beds, create the momentum needed to make these next-level investments and cultivate the new revenues long promised by 5G. Increasingly, as NVIDIA and other ORAN platforms reach scale, telecoms will have to rethink where such investments and workloads are hosted as they move more and more to metro-aggregated, multi-tenant facilities like data centers to achieve the best cost efficiencies and meet the latency demands that will come with 5G SA applications. And best of all, it will lead to the thaw sorely needed for the current 5G winter. #5G #NVIDIA #ORAN #telecom #datacenter https://lnkd.in/gDYdbfnN

13

ClusterOps: Revolutionizing Callable Execution Across Hardware The Swarm Corporation proudly announces the release of comprehensive documentation for ClusterOps, an innovative solution that empowers developers to fine-tune callable execution across diverse hardware configurations, including CPUs and GPUs. Key Benefits: ⎆ Precision Control: Customize execution down to specific CPU cores or GPU units ⎆ Scalability: Efficiently manage thousands of agents on a single core or deploy massive agents across GPU clusters ⎆ Flexibility: Tailor agent execution to optimize resource utilization and performance Explore the future of customizable agent execution. Visit our GitHub repository to access the ClusterOps documentation and unlock new possibilities in hardware optimization: https://buff.ly/3NzQW32 Like, repost, and share with friends ✨

2

We've researched the top questions and concerns about Distributed Antenna Systems (DAS) and are launching a series to address them. Today, we’re covering the scalability and future-proof technology of DAS from Wilson Electronics. Read More Here: https://bit.ly/3XCdBSz #DAS #DistributedAntennaSystem

6

How can service providers realize the advantages of network convergence using ZR+ optics? https://hubs.la/Q02G9tzR0 For service providers, going the extra mile with ZR+ and networking convergence is their best option. Shai Haim shares insights into optical convergence.

16

Thank you CTech by Calcalist for sharing the story about our work with NVIDIA to bring the #GenerativeAI moment to #robotics As was mentioned by Jensen Huang today in his COMPUTEX TAIPEI keynote, in the not-too-distant future, everything is going to be robotic and everything that moves will be autonomous. for that to happen, robots need #PhysicalAI, the ability to #perceive the world, #understand it and perform tasks given to them in the real world This is what RGo Robotics is all about

50

1 Comment

COMPUTEX TAIPEI = AI = Scaled Compute = Co-Packaged Optics Interconnect = RANOVUS Inc. @ Taiwan Today's GPU providers scaleout their Compute with copper interconnect technologies which limits their capacity. I had the opportunity to share some of our achievements and roadmap items with the COMPUTEX TAIPEI/InnoVEX audience around the next generation of Compute scaleout technologies this week. With the introduction of UALink (scaleout) and UltraEthernet (scaleup) standard groups, the industry is poised to bring high bitrate Ethernet connectivity (100/200/400Gbps per lane) as the backbone of the AI hardware infrastructure. https://lnkd.in/gHcrKPGa #SIP #CPO #AI #GPU #scaleout #scaleup

44

5 Comments

Check-out Kinnami's new blog "It's About the Data , not Just Devices - Resilient Data for Self-Driving Vehicles (and Other Autonomous Systems)". Self-driving cars make the headlines, however many autonomous systems at the computing edge face the same resiliency challenges. Kinnami AmiShare addresses the three pillars of data resiliency—data protection, security and availability. Our AI-driven policy engine provides centralized oversight for managing data resilience in complex distributed and autonomous systems, including self-driving cars, which need to operate at all times. #ResilientDataEverywhere https://lnkd.in/g_FtJdgv

5

In the age of AI, GPUs have become a focal point. I delve into the intriguing "Chip War" by Chris Miller in my latest blog. The book offers a captivating insight into the semiconductor industry's history, encompassing geopolitical dimensions. A highly recommended read! Grateful to Chris Kane for the recommendation.

11

2 Comments

🔦 How fast is fast? 🌟 Let's talk about the speed of light. At 299,792,458 meters per second, light zips through fiber optics at an unimaginable pace. It can circle the Earth 7.5 times in just one second! 🚀 In the world of deep tech innovation, understanding such speeds is crucial. Whether it's fiber-optic communication or the quest for quantum computing, the principles behind light's speed drive our advancements. 🌐✨ The speed of light isn't just a number; it's a reminder of the boundaries we're constantly pushing in science and technology. Let’s keep pushing! 🌍🔬 #TechInnovation #QuantumLeap #FutureTech #PhysicsFun #DeepTech

34

4 Comments

Day 2 of MWC Las Vegas 2024: EdgeQ Inc. Base Station-on-a-Chip Runs UPF on the RAN. Eliminates Need for Separate UPF Hardware. This paradigm shift unlocks unprecedented capabilities for edge computing by: • Reducing latency: Processing data directly at the radio unit • Optimize Bandwidth: Only Essential Data Transmitted to the Core Network, Thereby Reducing Congestion. • Enable a Distributed Architecture for Flexible and Scalable network • Enable ultra-low latency services and applications, such as AR/VR, connected vehicles, and industrial automation. • Reduce Infrastructure CapEx/OpEx costs Featured Partners: Blue Arcus Technologies, Arm H/T: Neil Winrow, Hersh Shukla, Ziyao Xu, Feng Kan

88

Bandwidth IG is proudly changing the landscape of fiber infrastructure in the San Francisco Bay Area. Patton Lochridge, CCO, states, "Dark fiber is the foundation of data transmission but many incumbent networks in the Bay Area lack capacity or are not capable of meeting the needs of the world’s largest consumers of data. By continuing our focus on new fiber infrastructure, we’re creating a path for connectivity in critical areas of the market that will allow hyperscalers and data centers to fully leverage AI and other society-changing technologies.” ❗ #DYK this is the first marine fiber cable under the Bay to be installed for communications services in decades? To learn more about custom solutions in the Bay Area, visit: https://lnkd.in/gY5uFH3b #SFBay #DarkFiber #Connectivity #FiberNetwork

58

@Qualcomm and #Snapdragon are revolutionizing the #XR space. Tune into this episode of "Exploring the Metaverse” to learn how — including discussions about recent #AR advancements, distributed compute, and the innovative HyperRealityHub. Watch the full video: #TeamQualcomm https://lnkd.in/gj8H8XmJ

60

🚀 Join IoT expert Umesh Lokhande for our latest Engineering Hour! Learn how Arm Cortex-M microcontrollers' nested vector interrupt controller (NVIC) simplifies hardware interrupt implementation. 📈 Umesh will explore the interrupt lifecycle, priority levels, and reentrancy interrupts, plus demonstrate managing two external interrupts on an STM32 MCU. Don't miss out on this deep dive into advanced design techniques! 📅 Date: Sunday, August 4th  🕒 Time: 15:00 GMT 👉 Join here: https://lnkd.in/gfGjh_U4 #IoT #Engineering #ArmCortexM #STM32 #EmbeddedSystems #TechTalk #Interrupts #NVIC #Microcontrollers #TechEducation

23

1 Comment

With the rise of #GenAI, compute requirements are changing rapidly. Tune in to the #DCDBroadcast today to hear Jeff Wittich discuss how Ampere is delivering hardware at a rapid pace to keep pace with #AI changes. DatacenterDynamics https://lnkd.in/gfXGBX3e

86