Deep Learning Weekly

💥 Exciting news! Deep Learning Weekly is thrilled to welcome its new lead sponsor: @Cometml. The team at Comet will keep sharing the weekly updates you're used to, as well as lend their decades of expertise into the mix. Read more about the transition: bit.ly/2ZPWfWq

🤖 From this week's issue: Plano is an AI-native proxy and data plane for agentic apps — with built-in orchestration, safety, observability, and smart LLM routing so you stay focused on your agents core logic. github.com/katanemo/plano

🤖 From this week's issue: A practical guide to LLM inference optimization framed around the "efficient frontier" concept — five techniques that move production systems toward the latency/throughput Pareto boundary without additional hardware spend. cloud.google.com/blog/topics/de…

🤖 From this week's issue: An open-source LLM evaluation tool used to debug, evaluate, and monitor LLM applications, RAG systems, and agentic workflows with comprehensive tracing, automated evaluations, and production-ready dashboards. github.com/comet-ml/opik

🤖 From this week's issue: A guide to multimodal embeddings and RAG covering the core theory (contrastive learning, modality gap, MRL), three concrete build patterns (audio, PDF, video), and when multimodal actually outperforms text-only pipelines. weaviate.io/blog/multimoda…

🤖 From this week's issue: A technical synthesis of three recent agentic RL training reports — Kimi K2.5, Cursor Composer 2, and Chroma Context-1 — distilling shared patterns around production-environment training, context management, and reward design. philschmid.de/kimi-composer-…

🤖 From this week's issue: A technical guide to deploying disaggregated LLM inference (separate prefill, decode, and router services) on Kubernetes using NVIDIA Grove, KAI Scheduler, and NVIDIA Dynamo. developer.nvidia.com/blog/deploying…

🤖 From this week's issue: Google launches Gemini 3.1 Flash Live, its highest-quality real-time audio model, scoring 90.8% on ComplexFuncBench Audio and 36.1% on AudioMultiChallenge blog.google/innovation-and…

🤖 Issue #449 is now live! This week features: Gemini 3.1 Flash Live, Cohere Transcribe: state-of-the-art speech recognition, a paper on IndexCache: Accelerating Sparse Attention via Cross-Layer Index Reuse, and many more! buff.ly/nhpVFbn

🤖 From this week's issue: Anthropic's largest-ever qualitative study — 80,508 Claude users across 159 countries and 70 languages — reveals what people want from AI, what they've already gotten, and what they fear. anthropic.com/features/81k-i…

🤖 From this week's issue: A technical deep-dive on how OpenSearch benchmarked its agentic search feature across search relevance (BEIR and BRIGHT datasets) and query execution accuracy (Spider dataset), powered by Claude Opus 4.6. opensearch.org/blog/evaluatin…

🤖: ByteByteGo breaks down Anthropic's interpretability research into six concrete findings about how Claude actually thinks — from parallel math strategies to ahead-of-time poetry planning to a default-refusal circuit that misfires into hallucinations. blog.bytebytego.com/p/how-anthropi…

🤖 From this week's issue: A visual reference guide mapping seven attention variants — MHA, GQA, MLA, SWA, DeepSeek Sparse Attention, Gated Attention, and hybrid architectures — across the open-weight models currently using them in production. magazine.sebastianraschka.com/p/visual-atten…

🤖 From this week's issue: An open-source LLM evaluation tool used to debug, evaluate, and monitor LLM applications, RAG systems, and agentic workflows with comprehensive tracing, automated evaluations, and production-ready dashboards. github.com/comet-ml/opik

🤖: A deep-dive on how Cursor built a local sparse n-gram index to replace ripgrep for agent search, eliminating 15+ second grep latency in large monorepos by narrowing regex matches to a pre-filtered candidate set before full scanning. cursor.com/blog/fast-rege…

🤖 From this week's issue: A technical blog post on how SkyPilot scaled Karpathy's autoresearch agent from 1 to 16 GPUs, enabling ~910 experiments in 8 hours. blog.skypilot.co/scaling-autore…

🤖 From this week's issue: Google Research releases TurboQuant, a KV cache quantization method that achieves 6x+ memory reduction to 3 bits with zero accuracy loss and 8x attention speedup on H100s. research.google/blog/turboquan…

🤖 From this week's issue: Cursor launches self-hosted cloud agents GA, keeping code and tool execution entirely within enterprise infrastructure while Cursor handles orchestration and inference. cursor.com/blog/self-host…

🤖 From this week's issue: OpenAI acquires Astral — maker of Python developer tools uv, Ruff, and ty used by millions of developers — to deepen its Codex ecosystem. buff.ly/WszJttM

🤖 From this week's issue: Cursor launches Composer 2, a frontier-level coding model trained via continued pretraining and long-horizon RL that scores 61.3 on CursorBench and 73.7 on SWE-bench Multilingual cursor.com/blog/composer-2

🤖 From this week's issue: A set of prompt-based safety policies designed to create age-appropriate protections for teens. github.com/openai/teen-sa…