Anthropic Launches Claude Platform on AWS — AWS Billing, IAM, CloudTrailAnthropic launched Claude Platform on AWS, a native API with AWS billing, IAM, and CloudTrail. Same models and pricing as direct API; data stays at Anthropic, not AWS. Anthropic made Claude Platform on AWS generally available today. The product lets...May 12, 2026·3 min read
Agentick Benchmark: GPT-5 Mini Tops at 0.309, No Agent Paradigm DominatesAgentick benchmark evaluates RL, LLM, VLM, and hybrid agents on 37 tasks. GPT-5 mini leads at 0.309 ONS, but no paradigm dominates. ASCII beats natural language. Agentick benchmark pits RL, LLM, VLM, hybrid, and human agents on 37 tasks. GPT-5 mini ...May 11, 2026·3 min read
RRCM Uses GRPO to Decide When to Retrieve for LLM RecommendationRRCM uses GRPO to learn when to retrieve evidence for LLM recommendation, outperforming fixed-context baselines. RRCM uses group relative policy optimization to learn when to retrieve evidence for LLM-based recommendation. The framework outperforms ...May 11, 2026·3 min read
Google CodeWiki Turns GitHub Repos Into Interactive DocsGoogle launched CodeWiki, turning any GitHub repo into interactive docs with diagrams, tutorials, and a chatbot. It differentiates by structure over file summarization. Google launched CodeWiki, a tool that converts any GitHub repository into intera...May 10, 2026·2 min read
mlx-audio v0.4.3 Ships 6 New TTS Models, Slimmer Depsmlx-audio v0.4.3 adds 6 TTS models, server concurrency, and slims dependencies, targeting Apple Silicon developers. mlx-audio v0.4.3 arrives with 6 new TTS models and server upgrades. The release targets Apple Silicon developers needing efficient, o...May 10, 2026·2 min read
AMD ROCm Performance Jumps 75x in 14 Days Post-DeepSeek v4AMD ROCm stack improved 75x in 14 days post-DeepSeek v4 via fused operations. Still needs 5x more to match B200 performance. AMD's ROCm software stack improved performance by over 75x in 14 days since DeepSeek v4's launch, according to @SemiAnalysis...May 10, 2026·3 min read
Claude Code's Six-Layer Architecture: Harness, Not MagicClaude Code's six-layer architecture uses a 3-layer context compressor at 92% threshold and Redis-based multi-agent FSM protocol. The model is just one node in a harness. Anthropic's Claude Code operates across six distinct layers, not as a simple C...May 10, 2026·3 min read