Image created with gemini-3.1-flash-image-preview with claude-opus-4.7. Image prompt: A Byzantine gold-mosaic reliquary disc shaped like a silicon wafer, its die grid formed from tiny gold tesserae with Tyrian crimson and imperial purple circuit traces, centered on a hammered-gold halo of concentric circuit-rings around a single elevated chip die, resting on a mosaic altar in a candlelit Ravenna apse, warm directional oil-lamp glow, the word CHIPS set in bold ivory Trajan capitals across the lower third, 16:9 full-bleed, painterly tactile tesserae texture.
ByteDance has had enough of waiting months for processors, so it’s going to make them itself | PC Gamer
https://www.pcgamer.com/hardware/processors/bytedance-has-had-enough-of-waiting-months-for-processors-so-its-going-to-make-them-itself/
If this is true, using the best public estimates we have of LLM resource use, solving this Erdos problem took 0.6-6.3 kWh of electricity and about 3-31 liters of water. So that is less than three almonds worth of water and the electricity equivalent of 2-20 miles of EV driving.
https://x.com/emollick/status/2057271533358162270
LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding
https://research.nvidia.com/labs/lpr/locate-anything/
Jensen Huang Says It Won’t Matter What You Study in the Age of AI – Business Insider
https://www.businessinsider.com/nvidia-jensen-huang-what-kids-should-study-ai-education-advice-2026-5
Power users account for a large share of AI activity, and the gap is widening.
https://x.com/cursor_ai/status/2060025074330243238
@JohnTinsman SpaceX has not committed to leasing Colossus for years, although it’s possible that may be what happens. This is a 180 day lease with 90 day notice mutual cancellation thereafter. The short term was our request, not Anthropic’s. We won’t leave them hanging and will provide a
https://x.com/elonmusk/status/2059880289514696927?s=20
How did Tensor Cores massively increase the throughput of Nvidia chips? @reinerpope explains the fundamental idea, systolic arrays:
https://x.com/dwarkesh_sp/status/2058624353965908245
.@reinerpope explains the difference between a CPU and a GPU. A huge chunk of a CPU die is actually just predicting what code will run next. GPUs strip this out, which is part of why they can fit so much more compute on the chip.
https://x.com/dwarkesh_sp/status/2058261856746377528
Billions of times a second, all the circuitry on an AI chip pauses, just for a moment. Why? @reinerpope explained to me what’s going on, and what the clock cycle of a chip is actually telling us:
https://x.com/dwarkesh_sp/status/2057925387569766405
For frontier AI chips, memory is the largest and fastest-growing component cost. High-bandwidth memory (HBM) has grown from 52% to 63% of total AI chip component spending between Q1 2024 and Q4 2025.
https://x.com/EpochAIResearch/status/2057531410030997789
Higher clock speed sounds like faster performance – but some ways of pushing it actually reduce a chip’s throughput. @reinerpope on a trade-off with wider implications: making the clock faster can sometimes cost you die space that could have been given over to compute.
https://x.com/dwarkesh_sp/status/2058201489924051413
IBM’s “”Project Lightwell”” [LWN.net]
https://lwn.net/Articles/1075065/
Inside the 800VDC Revolution – Part 1 Four-Phase 800VDC Transition, Power Rack Economics, SST, Equipment Content/MW Build, Supplier Implications
https://x.com/SemiAnalysis_/status/2059253624249696658
On a chip, even something as basic as selecting one input from a register file has to be built out of AND and OR gates wired together. Simple-seeming operations end up surprisingly expensive. @reinerpope showed me how this stacks up. In older generations of chips, most of the
https://x.com/dwarkesh_sp/status/2058594173905936765
On AI Hardware
https://www.categoryvc.com/writing/where-the-ai-hardware-market-is
The GPU has a lot of tiny, tiny TPUs tiled across the whole chip.”” @reinerpope on the architectural difference between the two dominant AI chip designs.
https://x.com/dwarkesh_sp/status/2057958877694677258
this is such an impressive result: search over ~600,000,000 colbert vectors in 10 milliseconds, with a *single* CPU core. and since this algorithm has sub-linear latency, there’s no excuse for anyone up to tens of billions of tokens
https://x.com/lateinteraction/status/2059985946448478380
This week, I solved a problem in RL involving ludicrous sparsity that I have been thinking about since 2018. Initial sweeps are showing SOTA on one of our most consistently informative test envs. Blog post soon. For now, you can follow the dev live on stream
https://x.com/jsuarez/status/2057828106023703037
Today, EAGLE powers some of the industry’s most formative AI infrastructure companies and teams. With EAGLE 3.1, we’re taking another major step toward delivering a core piece of the fastest possible inference stack that exists, open to all. By improving hidden-state feedback
https://x.com/EagleCorp/status/2059485457227149334
we relaunched @cloudflare’s startups website and made the review process much faster.
https://t.co/OHLy9BM7zK up to $350k in credits. apply plz, it’s time to build
https://x.com/kristianfreeman/status/2059188629780545973
Tech giants back new data center climate initiative
https://www.axios.com/2026/05/27/tech-giants-data-center-climate-initiative
The dominant story in AI has been the growing cloud: bigger clusters, larger models, more gigawatts. We believe the future is in the opposite direction: on-device inference, smaller models, watts instead of gigawatts. Today we’re releasing @OpenJarvisAI v1.0: a personal AI
https://x.com/JonSaadFalcon/status/2060054559142326468
New blackboard lecture w @reinerpope How do chips actually work – starting with basic logic gates, and working up to why GPUs, TPUs, FPGAs, and the human brain each look the way they do. 0:00:00 – Building a multiply-accumulate from logic gates 0:16:20 – Muxes and the cost of
https://x.com/dwarkesh_sp/status/2057857033659937220
@teortaxesTex Jensen tried to warn them – of the 50x increase from Hopper to Blackwell, <2x was process scaling. The rest was optimizations at other levels. Those other optimizations are just as accessible to Huawei. No EUV required.
https://x.com/josiah_leee/status/2059297861745963099
We are putting quantum inside the rack so customers can roll it in, plug it in’: The world’s first rack-mounted quantum computer is here — and it runs at -459 degrees Fahrenheit from a standard wall socket | TechRadar
https://www.techradar.com/pro/we-are-putting-quantum-inside-the-rack-so-customers-can-roll-it-in-plug-it-in-the-worlds-first-rack-mounted-quantum-computer-is-here-and-it-runs-at-459-degrees-fahrenheit-from-a-standard-wall-socket
Huawei’s “Tao / τ Law”: Tech Paper, White Paper, or Strategic Manifesto? 🧠🚀 🌟Insights from Zhihu contributor 无我梦中 Huawei’s new paper, “A Time Scaling Theory for Multi-Layer Electronic Systems” by Tingbo He, is better read as a semi-technical white paper + strategic
https://x.com/ZhihuFrontier/status/2059118295580852374
Another cool stuff from NVIDIA. LocateAnything – high-speed visual search engine. You provide a text prompt and it instantly pinpoints that object’s exact location in an image. – 10x speedup for dense object detection – Qwen2.5-3B + Moon-ViT – Fast/Slow/Hybrid modes – trained
https://x.com/wildmindai/status/2059600079804088790
Are we nearing a compute crunch? In our latest Gradient Update, @luke__emberson and @Jsevillamol estimate how many tokens all the Blackwell chips on Earth could serve, and compare this to total token demand. Direct comparisons are difficult, but it appears demand is growing much
https://x.com/EpochAIResearch/status/2059372951338909717
Extract More Kernel Performance with NVIDIA CompileIQ Auto-Tuning | NVIDIA Technical Blog
https://developer.nvidia.com/blog/extract-more-kernel-performance-with-nvidia-compileiq-auto-tuning/
OpenMDW-1.1 is now available — and @NVIDIAAI is adopting it across Cosmos, Isaac GR00T, Ising, and Nemotron model families. A permissive, unified legal framework purpose-built for AI models. Learn more at
https://x.com/linuxfoundation/status/2060031693193462036
Official @NVIDIAAI GLM5.1-NVFP4 spotted on @huggingface 🤩
https://x.com/mr_r0b0t/status/2059973066436853769
Nvidia bets $150B on Taiwan as Trump’s plan to make US an AI hub backfires – Ars Technica
https://arstechnica.com/tech-policy/2026/05/nvidia-ceo-wants-taiwan-to-be-center-of-ai-revolution-not-us/
OpenAI kicked off the AI compute buildout in 2023. But today it uses ~10% of the world’s compute, and the top labs together are probably under half. In this week’s newsletter, @justjoshinyou13 discusses how much that share may change, and when it could hit a ceiling. 🧵
https://x.com/EpochAIResearch/status/2057499893854536185
We’re adopting the Linux Foundation’s OpenMDW framework across our open model families. This helps make open model licensing simpler and more consistent at scale. A single legal framework across models, code, documentation, and data helps reduce friction for developers and
https://x.com/NVIDIAAI/status/2060035668655677804
@Jason the two US companies that are most seriously pushing open models above 100B params are NVIDIA and Arcee
https://x.com/willccbb/status/2060122252931412034
We’re open-sourcing the Unigram tokenizer we rebuilt to reduce CPU utilization by 5-6x. Small rerankers and embedders run in single-digit milliseconds on GPU, making CPU tokenization a meaningful share of total latency.
https://x.com/perplexity_ai/status/2059664738087469511
Key takeaways from NVIDIA’s quarterly earnings call last week: – Physical AI revenue >$9B over the trailing 12 months. – Edge computing platform (robotics, automotive, AI RAN, workstations) hit $6.4B, +29% YoY. – Asked if NVIDIA can outgrow hyperscalers (whose CapEx is
https://x.com/TheHumanoidHub/status/2059324196128534623
I have been very impressed by @SemiAnalysis_ . I think of myself as a wide ranging systems engineer, looking for value at every level from the chip specs to the user interface, but SA exposes me to additional levels of “”the system””, both above (datacenters) and below
https://x.com/ID_AA_Carmack/status/2059382254191652896
Leave a Reply