Image created with Gemini. Image prompt: Vintage letterpress book jacket illustration of an Edwardian man’s cuffed hand dangling a pocket watch whose face is a circular silicon wafer etched with fine processor die patterns, resting over a worn pub table with a lemon soap bar and folded tram ticket, bold serif title CHIPS across the top with generous margin, Shakespeare and Company blue, cream, lemon yellow and slate gray palette with a warm brick accent, flat inked color, visible paper grain and sun-faded screen-print texture, soft overcast June daylight, landscape 16:9.
The record for computing capacity in a single data center has doubled every 7 months. Colossus 1, Anthropic-Amazon New Carlisle, and Meta Prometheus have each claimed the top spot in turn.
https://x.com/EpochAIResearch/status/2065131037324402744
Building AI Agents for AR Glasses and XR Devices with NVIDIA XR AI | NVIDIA Technical Blog
https://developer.nvidia.com/blog/building-ai-agents-for-ar-glasses-and-xr-devices-with-nvidia-xr-ai/
NVIDIA Blackwell Leads on First Agentic AI Infrastructure Benchmark | NVIDIA Blog
https://blogs.nvidia.com/blog/nvidia-blackwell-agentperf-artificial-analysis/
AI GPUs probably live longer than three years
https://www.seangoedecke.com/ai-gpus-live-longer-than-three-years/
Cube on Databricks: The AI-Native BI Platform for Your Lakehouse
https://cube.registration.goldcast.io/webinar/27f396ef-3dfb-4172-9243-22669a77dac4
Kernels can be instructed to replace the forward passes of compatible layers with their optimized variants 😉 Better yet, you can specify all of it in a hardware-aware manner, completely owning the optimal configurations for your model 🔥 Learn below 👇
https://x.com/RisingSayak/status/2066487331209839026
RL Systems Mind the Gap: Matching Trainer and Generator Throughput RL Training Infrastructure, GRPO, PipelineRL, Async RL, Policy Staleness, RL Sandbox Infra, CPU Requirements, TCO Analysis, Thinking Machines Tinker
https://x.com/SemiAnalysis_/status/2066941079920791760
Sandboxes are all the rage (Modal, E2B, AWS, ..). Most AI teams pay a >4x markup to run sandboxes on someone else’s machines. Introducing SkyPilot Sandboxes — Run BYOC sandboxes on your own clusters. • 50,000+ sandboxes on a single cluster • Sub-second launches with warm
https://x.com/zongheng_yang/status/2065467594694598852
Qualcomm wants to be the chip inside whatever replaces your smartphone, and it just announced two products toward that end | TechCrunch
https://techcrunch.com/2026/06/16/qualcomm-wants-to-be-the-chip-inside-whatever-replaces-your-smartphone-and-it-just-announced-two-products-toward-that-end/
Huge milestone from the @anyscalecompute + @googlecloud GKE teams 🎊 Ray Serve LLM provides up to 4.4x higher throughput on prefill-heavy workloads and 24x on decode-heavy workloads than previous versions. Three optimizations made this possible on the Ray Serve LLM + vLLM
https://x.com/vllm_project/status/2067641904049885492
Microsoft Tests Phi Silica for Windows AI on Nvidia GPUs
https://winbuzzer.com/2026/06/15/microsoft-tests-phi-silica-for-windows-ai-on-nvidia-gpus-xcxwbn/
NVIDIA Blackwell Sweeps MLPerf Training 6.0 | NVIDIA Blog
https://blogs.nvidia.com/blog/blackwell-mlperf-training-6-0/
NVIDIA-ZPPO · Zone of Proximal Policy Optimization
https://byungkwanlee.github.io/ZPPO-page/
Nvidia’s Huang says society needs ‘new social norms’ in age of AI | AP News
https://apnews.com/article/nvidea-huang-artificial-intelligence-8334abcbc6ed8d3d7889b640ec6fa05b
We’ve reached an agreement to acquire @ona_hq. Its secure cloud execution technology will help Codex take on longer-running work, even when laptops are closed, and help more organizations deploy agents securely in production. After closing, Ona will join OpenAI’s Codex team.
https://x.com/OpenAINewsroom/status/2065088002335158753
Today, we enable AutoResearch in the physical world for the first time! Introducing ENPIRE: we give 8 Codex agents a fleet of robots, an allocation of GPUs, and generous token budget. We set them free with a simple goal: solve the task as quickly as possible, keep the robots busy
https://x.com/DrJimFan/status/2066921736369766762
3.7x faster transformer load (disk to GPU) on a H100
https://x.com/maharshii/status/2066508679340589256
Leave a Reply