Blog

Deep-dive guides on AI agents, agent orchestration, MCP, and developer tooling.

15 posts found

Agent Observability

Tracing an Agent Loop: What OpenTelemetry's GenAI Conventions Actually Standardize

Agent traces have a runtime-decided shape, token-based cost, and cross-process tool hops that conventional APM was never built for — here is what OpenTelemetry's gen_ai.* conventions actually pin down, and where tracing an agent loop still breaks in production.

August 8, 2026 · 10 min read
MCP

MCP Authorization: How OAuth 2.1 Actually Secures a Remote MCP Server

MCP started as a trusted local subprocess with no auth story at all. Its authorization spec bolts on OAuth 2.1, PKCE, and resource indicators — here is what each piece actually prevents, including the token-passthrough bug that keeps showing up in early implementations.

August 5, 2026 · 9 min read
Benchmark Methodology

What SWE-bench Actually Measures vs. What the Leaderboard Tells You

SWE-bench is the most credible agentic coding benchmark available, but its leaderboard number answers a narrower question than most headlines imply — here is exactly what it does and does not test.

August 3, 2026 · 8 min read
AI in Scientific Research

Where AI Actually Works in Scientific Research: Protein Design, Materials Discovery, and the GNoME Lesson

AlphaFold's Nobel Prize and GNoME's 2.2 million materials both got real headlines — but the honest story is about candidate generation, not discovery, and the gap between those two claims is where the hype lives.

August 2, 2026 · 7 min read
Model Comparisons

Claude Sonnet 5 vs. GPT-5.6 vs. Gemini 3.1 Pro: What the Pricing Table Doesn't Tell You

A side-by-side look at pricing, context windows, and tiering across the three frontier model families as of August 2026 — and why the spec sheet undersells how differently they actually behave in agent loops.

August 1, 2026 · 6 min read
AI in Customer Support

Where AI Actually Works in Customer Support: Deflection, Copilots, and the Outcome-Pricing Bet

Klarna's AI assistant did the work of 700 agents, then the company started rehiring humans. The real story is that support AI succeeds in narrow, policy-bounded tiers and breaks hard the moment it touches judgment calls and liability.

July 28, 2026 · 6 min read
Prompt Engineering

Prompt Engineering for Agents Is a Different Discipline Than Prompt Engineering for Chat

In agent systems, instructions live across four surfaces — system prompt, tool schemas, tool results, and few-shot text — not one. Most prompt debugging still only looks at the first.

July 7, 2026 · 8 min read
Context Management

Context Compaction: How Long-Running Agents Avoid Drowning in Their Own History

Long-running agents rarely fail because they run out of context window — they fail because nobody designed what happens to attention quality once the transcript outgrows what the model can usefully weigh. Here is how tiered compaction, tool-output pruning, and sub-agent isolation actually work.

July 6, 2026 · 8 min read
RAG Architectures

Naive RAG, Agentic RAG, and GraphRAG: What Actually Changes Architecturally

RAG is not one architecture — it is three structurally different systems with different costs and failure modes. Here is what actually changes between naive, agentic, and graph-based retrieval, and how to pick without over-building.

July 5, 2026 · 8 min read
Agent Orchestration

Pipeline, Supervisor, or Mesh: Where Each Multi-Agent Orchestration Pattern Actually Breaks

The three dominant multi-agent orchestration topologies each fail in a different, predictable way once you move past the demo — here is how to pick one based on where your task actually breaks, not which pattern sounds more sophisticated.

July 4, 2026 · 9 min read
LLM Serving

KV Cache Reuse and the Hidden Latency Budget of Agent Loops

Agent latency is usually a prompt-construction problem in disguise: prefix caching, continuous batching, and tool-call structure determine how much of your context gets reprocessed on every turn.

July 2, 2026 · 9 min read
Agent Security

The Trust Boundary Problem: Why Tool-Calling Agents Need to Treat Tool Output as Untrusted Input

Most agent security advice targets prompt injection at the wrong layer. The real fix is architectural: separate untrusted tool output from privileged context, scope tool capabilities narrowly, and gate side-effecting actions behind confirmation.

July 2, 2026 · 9 min read
Agent Evaluation

Why Your Agent Benchmark Score Doesn't Predict Production Reliability

Benchmark leaderboards measure task completion under lab conditions, not the compounding step failures that sink agents in production. Here is the math, and a blueprint for an eval harness that actually predicts reliability.

July 1, 2026 · 8 min read
Agent Memory

Agent Memory Isn't RAG: Why Vector Retrieval Falls Apart for Stateful Agents

Vector similarity search answers what text is topically related — but long-running agents need to know what is true right now. Conflating the two is why agents keep resurrecting overturned decisions.

June 30, 2026 · 8 min read
MCP

What Actually Happens Inside an MCP Tool Call

A wire-level look at the Model Context Protocol — capability negotiation, tool discovery, transport tradeoffs, and the context-budget mistakes that quietly degrade agent reliability.

June 29, 2026 · 9 min read