Skills in Claude Code 2.1: Loadable Capability Packs Explained
A practical engineering deep dive into Claude Code 2.1 skills, covering architecture, tradeoffs, and what production teams need to know about developer tooling.
10 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 266 of 631
Agentic AI & LLMs
Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Latest articles
Signed MCP Registry: How Anthropic Is Securing the Tool Supply Chain
Infrastructure-level look at MCP signed registry, including supply chain security AI, deployment topology, region availability, and cost considerations.
12 min read
Azure AI Foundry Multi-Agent Orchestration in 2026: Patterns
Azure AI Foundry's connected agents and shared memory make multi-agent orchestration on Azure finally workable. The patterns and gotchas to know before going live.
10 min read
CrewAI Tools Marketplace: Top Trending Tools in April 2026 Guide
The CrewAI tools marketplace crossed 500 tools in April 2026. The most-installed tools and the ones quietly powering production crews across industries today.
8 min read
OpenAI text-embedding-3 vs Voyage vs BGE-M3 in 2026: Honest Picks
Embedding model picks in 2026 with real numbers. text-embedding-3, Voyage, Cohere v4, and BGE-M3 across English, multilingual, and code retrieval workloads measured.
11 min read
Semantic Cache for Agent-Memory Architectures: Real-World Patterns
A semantic cache layer in front of memory recall cuts tail latency and saves embedding API spend significantly. Where it works, where it lies, and how to debug it.
10 min read
Operator 2.0 in Singapore: APAC Browser Automation at Scale
How Singapore-based companies are using ChatGPT Operator 2.0 for cross-border APAC workflows — pricing, latency, and regulatory considerations in 2026.
8 min read
How to Add Voice Agent Observability with Langfuse and OpenTelemetry
Trace every turn, every tool call, every LLM round-trip with OpenTelemetry shipped to Langfuse. Find latency outliers, debug hallucinations, and watch p95 stay under 800ms.
11 min read
Cloudflare Calls vs Fastly: Edge WebRTC for Global Voice AI (2026)
Cloudflare Calls is anycast WebRTC across 330+ cities. Fastly is leaner and faster on the compute edge. Here is how each fits a 2026 global voice-AI deployment.
12 min