Stanford’s CS336 covers modern language-model systems
The 2026 edition of CS336 covers MoEs, GPU tiling, kernels, RLHF, and data. According to the listing, lectures appear on YouTube with a delay of about two weeks.
Read articleKnowledge you can apply
Radar
Daily bookmarks: relevant news, facts and discoveries.
Firefly AIBOX-K3 pairs RISC-V with edge AI hardware
Published on May 12, 2026, the item describes an industrial edge AI computer with a SpacemiT K3 RISC-V SoC, up to 32 GB of RAM, and up to 512 GB of UFS storage.
Read articleA minimal time-sharing kernel in Zig for RISC-V
An article published on May 12, 2026 presents a walkthrough for creating a minimal time-sharing kernel in Zig for RISC-V on OpenSBI.
Read articleC2IntelFeeds collects C2 infrastructure feeds
Published on May 11, 2026, C2IntelFeeds offers feeds generated from large-scale internet scans, covering suspected and verified C2 infrastructure indicators.
Read articleFluxion proposes hybrid sparse attention for long-context inference
Published on May 11, 2026, the summary presents Fluxion, a hybrid sparse-attention system for long-context inference that uses CPU and GPU parallelism.
Read articleRadar: virtual memory and Linux internals
Published on May 11, 2026, this bookmark describes a practical guide to virtual memory, page faults, page tables, TLBs, NUMA topology, Linux internals, and performance in data-intensive systems.
Read articleThe mathematics of Query, Key, and Value in attention
Published on May 10, 2026, the article presents the central attention formula in transformers and walks through a numerical example step by step.
Read articleCuBridge uses LLMs to reconstruct attention kernels
Published on May 10, 2026, the item describes CuBridge, an LLM-based framework intended to understand and reconstruct high-performance attention kernels.
Read articleVideo offers an overview of GPU architecture
Published on May 10, 2026, the video gives an overview of graphics card internals and provides context for engineers working on GPU software.
Read articleKV cache reduces repeated computation during inference
Published on May 9, 2026, the article explains how a KV cache stores Key and Value data for processed tokens so they need not be recalculated for each new token.
Read articleLeWorldModel: pixel-based JEPA trained on one GPU
A May 9, 2026 publication reports training a 15-million-parameter JEPA from raw pixels on one GPU and compares its planning speed on control tasks.
Read articleLuce PFlash reports 2.89× faster TTFT at 64K context
A May 9, 2026 publication reports that Luce PFlash achieved 2.89× faster time to first token than Ollama in a benchmark with a 64K context.
Read articleLucebox explores speculative inference on consumer GPUs
Published on May 9, 2026, this entry presents Lucebox, a speculative inference server for heterogeneous hardware and consumer GPUs. The post claims Qwen3.6-27B reaches 120–200 tokens per second on an RTX 3090.
Read articleRTX 3090: power and concurrency tests for LLMs
Published May 9, 2026, the report compares tests of Qwen3.6 27B on an RTX 3090 at concurrency levels of 1, 4, 8, and 16. The author reports that 225 W was most efficient, while 250 W delivered higher throughput.
Read articleTC-JEPA uses captions to guide visual representations
Published on May 9, 2026, the post describes a self-supervised visual learning method that uses captions to guide predictions of masked image patches.
Read articleRTX 3090 tests compare LLM throughput and efficiency
A briefing published on May 8, 2026 compares eight local LLMs on one RTX 3090, at power limits from 100 W to 450 W. Average throughput rose at the higher tested power, while energy efficiency fell.
Read articleSAM in mid-training and retention after fine-tuning
A post published on May 8, 2026 proposes using SAM during the final roughly 10% of mid-training to reduce forgetting after fine-tuning or quantization.
Read articleStudy explores geometric limits of compression in RAG
An article presented as a NeurIPS 2026 paper connects compression in RAG with embedding geometry and claims that a spectral quantity establishes a minimum bound on compression.
Read articleMamba-3 compresses context into a fixed-size state
Published on May 7, 2026, the excerpt describes Mamba-3 as a non-Transformer language-model architecture with a fixed-size state. It claims constant time per decoding step and faster inference than a Transformer beyond an unspecified threshold.
Read articleOpen ASR Leaderboard adds private evaluation
On May 6, 2026, the Open ASR Leaderboard announced private evaluation data intended to reduce test-set contamination and overfitting in speech recognition benchmarks.
Read articleRota Nacional
Bring privacy into your workflow.
30 days, no card, with a starting quota. After that, Pix credit from R$ 10,00.