https://agentropic.ai/blog/ The engineering behind our builds: evals, prompt-injection defenses, MCP authorization, agent economics — written as we go. en-us Mon, 27 Jul 2026 00:00:00 GMT https://agentropic.ai/blog/system-prompt-redundant-context-engineering/ https://agentropic.ai/blog/system-prompt-redundant-context-engineering/ Anthropic removed over 80% of Claude Code's system prompt for newer models. We tested the idea against a real production CLAUDE.md—and found a bigger problem than token count. Mon, 27 Jul 2026 00:00:00 GMT https://agentropic.ai/blog/mcp-authorization-production/ https://agentropic.ai/blog/mcp-authorization-production/ A precise walkthrough of the OAuth 2.1 resource-server model MCP uses: audience binding, token passthrough, confused-deputy, the Nov 2025 spec changes, and what to ship. Sat, 20 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/computer-use-agents-osworld-reality/ https://agentropic.ai/blog/computer-use-agents-osworld-reality/ What OSWorld actually tests, how to read current scores honestly, and the specific gap between a benchmark number and a computer-use agent you can trust on real work. Thu, 18 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/prompt-injection-defenses-that-hold/ https://agentropic.ai/blog/prompt-injection-defenses-that-hold/ A taxonomy of direct and indirect prompt injection, why input filtering fails, and the architectural patterns (CaMeL, dual-LLM, capability design) that actually contain it. Tue, 16 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/open-weight-vs-closed-2026/ https://agentropic.ai/blog/open-weight-vs-closed-2026/ A practitioner's model for deciding between open-weight self-hosting and closed APIs: real break-even math, GPU utilization traps, licensing limits, and the EU AI Act split. Sun, 14 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/evals-as-shipping-contract/ https://agentropic.ai/blog/evals-as-shipping-contract/ How to build an LLM eval suite that earns the right to block a deploy: cases from real failures, regression vs adversarial sets, noise-aware thresholds, CI gates, model-upgrade gating. Fri, 12 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/self-host-llm-vs-api-cost/ https://agentropic.ai/blog/self-host-llm-vs-api-cost/ A mechanism-level cost comparison of self-hosting open-weight LLMs versus per-token APIs: GPU amortization, the utilization trap, throughput math, and where each one actually wins. Wed, 10 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/why-ai-bill-rises-jevons/ https://agentropic.ai/blog/why-ai-bill-rises-jevons/ Per-token prices fell ~280x in 18 months yet enterprise inference spend doubled. Here is the mechanism, the real cost drivers, and how to instrument and cap them. Mon, 08 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/llm-model-routing-cost/ https://agentropic.ai/blog/llm-model-routing-cost/ How LLM routing and cascading actually cut cost, what you really save versus benchmark claims, the confidence-signal trap, and how to instrument a router so quality loss can't hide. Sat, 06 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/gdpval-what-it-measures/ https://agentropic.ai/blog/gdpval-what-it-measures/ A technical breakdown of how GDPval is built, what "matches experts on ~half of tasks" really means, the 100x cost claim, grading reliability, and the benchmark's real limits. Thu, 04 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/spec-driven-development-explained/ https://agentropic.ai/blog/spec-driven-development-explained/ A technical walkthrough of spec-driven development with AI agents: the files that land in your repo, the plan-implement-verify loop, how review and CI shift, and where it breaks. Tue, 02 Jun 2026 00:00:00 GMT https://agentropic.ai/blog/parallel-subagent-orchestration/ https://agentropic.ai/blog/parallel-subagent-orchestration/ How to decide when to run LLM subagents in parallel versus keep a single agent: the read/write split, token-cost math, the equal-budget critique, and a runnable orchestrator skeleton. Sun, 31 May 2026 00:00:00 GMT https://agentropic.ai/blog/ai-agent-unit-economics/ https://agentropic.ai/blog/ai-agent-unit-economics/ How to actually model the unit cost of an LLM agent: cost-per-correct-outcome, the compounding-error tax over multi-step runs, token re-send growth, and where the math breaks. Fri, 29 May 2026 00:00:00 GMT https://agentropic.ai/blog/stop-training-start-restructuring/ https://agentropic.ai/blog/stop-training-start-restructuring/ AI training programs teach tools within broken workflows. The best AI training is a restructured org where people learn by doing real work with agents. Tue, 17 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/agentic-development-not-vibe-coding/ https://agentropic.ai/blog/agentic-development-not-vibe-coding/ Vibe coding is prompting and hoping. Agentic development is a structured framework where developers orchestrate AI agents in parallel. Here's the difference and why it matters. Mon, 16 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/ai-consultancy-industrial-complex/ https://agentropic.ai/blog/ai-consultancy-industrial-complex/ Big firms charge crores for AI strategy decks that sit in Google Drive. The gap between advice and deployment is where most AI initiatives die. Sat, 14 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/outcome-pods-departments-are-dead/ https://agentropic.ai/blog/outcome-pods-departments-are-dead/ The traditional org chart optimizes for skills, not results. Outcome-based pods — small teams of humans and AI agents driving a single metric — are what actually works. Fri, 13 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/ai-due-diligence-pe-firms/ https://agentropic.ai/blog/ai-due-diligence-pe-firms/ A framework for PE and VC partners evaluating portfolio company AI readiness. How to score it, what red flags to watch for, and why it matters for exit multiples. Thu, 12 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/ai-transformation-company-india/ https://agentropic.ai/blog/ai-transformation-company-india/ A buyer's guide for Indian companies evaluating AI transformation partners. Three categories of vendors, red flags, green flags, and what the mid-market actually needs. Tue, 10 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/customer-support-cost-96-percent/ https://agentropic.ai/blog/customer-support-cost-96-percent/ How we automated 96% of customer interactions at a 4.4M-subscriber platform without laying anyone off. The support team moved from cost center to growth engine. Mon, 09 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/20x-engineering-productivity/ https://agentropic.ai/blog/20x-engineering-productivity/ 20x doesn't mean 20x lines of code. It means cycle time collapse, scope expansion per person, and the elimination of everything that isn't building. Here's what happened. Sun, 08 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/non-engineers-building-with-ai/ https://agentropic.ai/blog/non-engineers-building-with-ai/ The most important signal that AI transformation is working isn't engineering velocity. It's when a promo editor builds their own automation and a marketer ships production code. Sat, 07 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/ai-audit-what-we-look-for/ https://agentropic.ai/blog/ai-audit-what-we-look-for/ Our 2-week AI audit is how we figure out where AI will have the biggest impact in your company. Here's exactly what we assess, how we assess it, and what you get at the end. Fri, 06 Mar 2026 00:00:00 GMT https://agentropic.ai/blog/why-ai-pilots-fail/ https://agentropic.ai/blog/why-ai-pilots-fail/ Most companies treat AI adoption as a tool problem. The ones that succeed treat it as an org design problem. Here's what separates the 5% from the rest. Thu, 05 Mar 2026 00:00:00 GMT