Tristan Thrasher

🌍 Leading #telecom companies across five continents are building sovereign AI infrastructure, accelerated by NVIDIA. From Africa to Asia to Europe to North and South America, telcos are becoming #AI infrastructure providers for enterprises, startups, institutions, and governments to: 🚀 Accelerate national economic growth 🔬 Advance scientific discovery and innovation 🤝 Empower local workforces and enhance citizen services 🔗 Read the blog post to learn more about sovereign AI factories.

2

Qualcomm has acquired UK-based Alphawave Semi for $2.4B, a strategic move to supercharge its AI and data center capabilities. Alphawave’s cutting-edge high-speed connectivity tech will bolster Qualcomm’s push into the booming AI market, with the deal set to close in Q1 2026. This acquisition highlights the growing importance of advanced semiconductor solutions in powering next-gen AI systems. Exciting times ahead for innovation! #Qualcomm #Alphawave #TechAcquisition #AI #Semiconductors #DataCenter #Innovation #TechNews https://lnkd.in/guQmigiZ

4

🚀 Just deployed a full Ubuntu Server on the KV260 Kria SoM by AMD Xilinx — a powerful FPGA-based edge computing platform designed for AI and vision workloads! 🔧 I booted up the board using Tera Term from my PC, established the initial setup, and installed all the required dependencies for AI inference. 🧠 Then came the fun part: I installed YOLOv5n, the lightweight variant of the YOLOv5 object detection family. Using a Python script, I successfully ran real-time object and face detection directly on the board. ⚙️📷 💡 The performance of the KV260 is truly impressive — compact yet powerful enough to handle deep learning inference at the edge. Excited about what's next: optimizing the pipeline further, and possibly training custom models (👀 drone detection incoming!) #AI #EdgeComputing #FPGA #KriaKV260 #AMD #Xilinx #YOLOv5 #ObjectDetection #FaceDetection #Python #UbuntuServer #EmbeddedAI #TechBuilds #DeepLearning #ComputerVision Manuj Dubey Raj Shivhare

33

📡 RF Engineers: The Resume Hack That Highlights Signal Chain Wins 📡 Recruiters skim fast. Your signal chain impact must pop instantly. If your resume shows measurable RF wins, you rise above the crowd. Here’s the shortcut that makes them stop scrolling. • Lead with gain or noise figure improvements ⚡ Pro tip: Include exact numbers for immediate credibility. • Highlight filter or matching network wins 📊 Pro tip: Show how you improved insertion loss or return loss metrics. • Surface toughest interference fix 🔍 Pro tip: Mention one scenario where your adjustment solved real RF issues. • Show system-level chain improvements 🚀 Pro tip: Connect component optimization to measurable system performance. • Link simulations to hardware validation 🧪 Pro tip: Add one line showing how your design passed lab testing first try. 📢 Follow JobLanderz.VIP for bite-sized job search strategies that work! 📢 Tag a friend who's job searching! Help them land their dream job faster. #RFEngineers #SignalChainDesign #HardwareValidation #RFDesignCareers #JobLanderzVIP Final tip: Add one before-and-after metric for each signal chain element – measurable results grab recruiter attention.

1

1 Comment

AI Chips Depreciate Quickly Due To Heat Stress The Michael Burry view is that companies are depreciating their GPUs over 5 year periods despite technological obsolescence in 18 months or less. But heat stress also depreciates GPUs. AI Chips run marathon sessions during model estimation (training) often exceeding a month and GPUs often fail due to heat stress: "Imagine you had a 10,000 or even a 20,000 GPU data center. You should expect on the statistics a chip to fail about every 3 or 4 hours. So long before I get to the point where I’m rapidly turning these over because there’s a new generation of chips, I’m turning over a vast chunk of my chips just because they’re failing under thermal stress." Source: https://lnkd.in/dDM7hU_8?. GPUs used to estimate large language models are like ultra-marathon (50, 100, and 166 kilometers) runners on a 38 degree day. It took Facebook 54 days to estimate the Llama 3 open source large language model using16,384 Nvidia H100 80GB GPUs. There was a hardware failure roughly every three hours with GPU chip heat deaths comprising the majority of them. https://lnkd.in/dApTwJkz

20

6 Comments

I stopped blaming AI agents for “misunderstanding” my PRDs. They weren’t confused. They were following my instructions perfectly. That’s when I realized something uncomfortable 👇 PRDs need to be written so agents can’t misunderstand them. Here are 5 quick tips that actually work: 1️⃣ Write behavior, not intention ❌ “Make the widget flexible” ✅ “When Region changes → reset Devices → reload data” If it can be interpreted, it will be. 2️⃣ Assume zero context Write like: • No meetings happened • No follow-up questions allowed Because agents read PRDs exactly like that. 3️⃣ Replace adjectives with structure Agents don’t understand: “clean”, “optimized”, “intuitive” They understand: lists, limits, formats. 4️⃣ Acceptance should be pass/fail If someone can’t say yes or no in 5 seconds, rewrite the acceptance. Hope is not a requirement. 5️⃣ Constraints are kindness Max limits. Defaults. What NOT to assume. Boundaries make execution faster. The shift that changed everything PRDs are no longer just for humans. They are instructions for machines. Write them that way — humans will understand them even better. 👇 What’s one PRD line you’ve written that sounded right but broke later?

3

1 Comment

NVIDIA and Nokia are pioneering the AI era of telecommunications. At NVIDIA GTC Washington D.C, NVIDIA introduced NVIDIA Aerial RAN Computer Pro — a 6G-ready accelerated computing platform. Nokia will use NVIDIA platforms to build new AI-RAN products for global deployment. This partnership marks a turning point for the industry, paving the way to AI-native 6G. Read the press release to learn more.

3

After hearing John Smee and Durga Malladi of Qualcomm speak separately, they both mentioned the three pillars of 6G that they feel are crucial, and their 40+ partner coalition for 6G. As someone working in 3GPP standards, here's what stands out to me about their 6G vision. - Advanced Connectivity – which is really the history of Qualcomm. With 5G-Advanced and the evolution to 6G we are looking at much higher capacity, wider bandwidths (including up toward sub‑THz), better uplink, tighter reliability, and more uniform coverage, including NTN and dense small cells. The other side of the coin is tuning the network and uplink for continuous context exchange between a lot of devices and AI agents rather than mainly human‑driven traffic. - Sensing they believe is a game-changer. Integrated sensing capabilities (ISAC )that detect objects, motion, and gestures without active electronics on the target. The radio network becomes a sensor network. - Energy‑efficient high‑performance compute (distributed compute). Virtualized and cloud RAN with AI-driven network autonomy The 2028 pre-commercial target is aggressive for sure. The race to define the AI-native network era is on. #6G #MWC2026 #Qualcomm #3GPP #5GAdvanced #AI #Telecom

70

4 Comments