Harness and costs of corporate agents
A July 13, 2026 publication reports that changing agent orchestration reduced cost and latency while maintaining equivalent quality across 22 tasks and six models.
Read articleKnowledge you can apply
Radar
Daily bookmarks: relevant news, facts and discoveries.
StreamDQ brings dequantization closer to HBM memory
Published on July 13, 2026, the summary presents StreamDQ, an architecture that dequantizes LLM weights in real time on the HBM base die while preserving conventional GPU load semantics.
Read articleReference harness for security scanning and remediation
Published on July 12, 2026, the repository described offers skills for threat modeling, scanning, triage, and patching, plus a customizable harness for autonomous scanning.
Read articleREINFORCE: an estimator for policy gradients
Published on July 12, 2026, the first post in a series derives the REINFORCE estimator as an unbiased policy-gradient method and analyzes how variance grows with trajectory length.
Read articlePrompt injection in binaries analyzed with AI
A July 12, 2026 bookmark describes Naval Postgraduate School research on prompt-injection strings in C binaries and their influence on LLM reverse-engineering agents in Ghidra.
Read articleVulkan documentation explains compute shaders and GPUs
A Vulkan tutorial introduces compute shaders, GPU architecture and memory models. Published on July 12, 2026, it notes that much of the discussion also applies to DirectX 12.
Read articleOpenResearch reports automated paper reproduction
Published on July 11, 2026, the entry describes a tool that uses agents to reproduce research results and provision cloud GPUs. The author says it reproduced their paper end to end.
Read articleColibri loads MoE experts on demand
Published on July 11, 2026, Colibri is described as a dependency-free C inference engine that loads MoE model experts from disk as needed. The post says it runs a 744-billion-parameter model on a laptop with 25 GB of RAM.
Read articleA hypothesis on how RL expands SFT
Published on July 11, 2026, the post presents a hypothesis: reward-guided variation during RL may separate useful, intertwined components in SFT solutions into reusable skills and routing rules.
Read articleVerifying code-agent rewards
Published on July 11, 2026, the paper examines four designs for verifying code-agent rewards and finds no single solution.
Read articleMemory layout of multidimensional arrays
Eli Bendersky’s article examines the memory layout of multidimensional arrays and includes benchmark code comparing the performance of different access patterns.
Read articleMiniCPM-SALA combines sparse and linear attention
Published on July 11, 2026, the report presents a 9-billion-parameter hybrid architecture for long contexts and reports good benchmark results, with scalable inference up to 1 million tokens.
Read articleWhat happens when a CUDA kernel runs
An introduction to the CPU–GPU communication required to launch a CUDA kernel and the work involved in that operation.
Read articleRLSD adjusts RLVR updates per token
A paper published on July 11, 2026 combines environmental rewards with self-distillation and reports higher average accuracy on five multimodal benchmarks.
Read articleSDM proposes sparse memory for RNNs
Published on July 11, 2026, the post describes Sparse Delta Memory: an explicit memory with sparse addressing used alongside an RNN that, according to the post, expands state capacity without raising compute per token.
Read articleThetaEvolve explores test-time learning
Published on July 11, 2026, ThetaEvolve is an open-source framework for test-time learning on open-ended optimization problems. Its description outlines components, without reporting quantitative results.
Read articleAREAL2.0 proposes an architecture for agents with continuous learning
Published on July 10, 2026, the article presents an architecture for continuous learning by agents on real workloads, with step-by-step signals, task-to-training-data conversion, and automatic updates.
Read articleTriton Paged Attention Kernel Attributed to Grok
July 10, 2026 archive entry: a post shares a Triton paged attention kernel attributed to Grok and a solution file hosted on KernelBench.
Read articleKV cache quantization: keys and values
An ACL 2026 article reports that using 4 bits for keys and 2 for values may work well for KV cache compression; reversing those bit widths substantially degrades accuracy.
Read articleLED explores intermediate layers to broaden sampling
Published on July 10, 2026, the report presents Latent Exploration Decoding (LED), a method requiring no additional training that uses intermediate-layer distributions to improve sampling in reasoning models.
Read articleRota Nacional
Bring privacy into your workflow.
30 days, no card, with a starting quota. After that, Pix credit from R$ 10,00.