https://rjl.pub/ 2026-10-21T22:16:49.903Z weekly 1.0 https://rjl.pub/img/rjl-pub-preview.png RJL.pub - AI-Native Development Journey Comprehensive guide to AI-assisted software development by RJ Lindelof https://rjl.pub/img/cover/1.png RJL.pub - AI-Native Development Journey Cover 1 of 12 https://rjl.pub/img/ai-native.png AI-native development reorganizes the full software lifecycle around how AI operates https://rjl.pub/img/ai-assisted-developer.png Traditional development workflow versus the agentic workflow, and what AI excels at https://rjl.pub/img/ai-codes.png AI codes, engineers build software https://rjl.pub/img/sections/tool-evolution.png Six-step staircase showing the tool evolution from chat to IDE chat, inline edit, multi-file agent, full-project agent and async fleet. https://rjl.pub/img/context-as-infra.png Context as infrastructure: what context is, the cost of poor context, context ROI, context layers, and guardrail best practices https://rjl.pub/img/sections/agents-vs-claude-md.png Comparison of AGENTS.md as a concise cross-tool file and CLAUDE.md as a deep Claude Code file, with an agent that merges both. https://rjl.pub/img/intent-engineering.png Intent engineering turns human intent into precise, reliable outcomes: clarify, structure, refine, optimize https://rjl.pub/img/sections/six-failure-modes.png Six cards naming the failure modes intent prevents: scope drift, runaway tool use, silent partial completion, missing escalation, success theater and irreversibility. https://rjl.pub/img/claude-code.png Claude Code architecture and capabilities: every surface it runs on, core capabilities, and the 2026 operations layer https://rjl.pub/img/github-copilot-alternatives.png GitHub Copilot in 2026: inline suggestions, multi-model picker, agent mode, coding agent, MCP support, and enterprise controls https://rjl.pub/img/cursor-ai-first.png Cursor and AI-first IDEs: the AI-first editor, the 2026 shift to managing agents, and Cursor-specific features https://rjl.pub/img/mcp-server.png MCP, A2A and the protocol layer: agents, the MCP protocol, transports, and MCP servers for tools and data https://rjl.pub/img/sections/model-routing.png Flow diagram routing a task to a fast cheap model or a frontier model depending on complexity, with escalation on failure. https://rjl.pub/img/prompt-engineering-old.png Prompt engineering is table stakes; work with agents like senior engineers https://rjl.pub/img/sections/context-priority.png Four-tier funnel ranking context from critical, always include, down to low priority, omit. https://rjl.pub/img/sections/rag-pipeline.png Two-row pipeline: index documents into chunks, embeddings and a vector store, then embed a question, retrieve nearest chunks, build a prompt and produce an answer. https://rjl.pub/img/sections/cache-layers.png Three stacked layers ordered by how often they change: always cached, cached per session, and never cached. https://rjl.pub/img/sections/plan-act-verify.png Loop diagram of plan, act, verify and correct, ending at done only when an independent check confirms all criteria are met. https://rjl.pub/img/sections/async-fleet.png Issue fanned out to three isolated branches, then through CI evals, then to human review. https://rjl.pub/img/what-problem.png The reframing question: what problem are we trying to solve? https://rjl.pub/img/sections/spec-source-of-truth.png The spec loop: tasks, agent, review, and evals all point back to one spec https://rjl.pub/img/sections/review-gates.png Pull requests pass an AI review, a quality gate, and evals before a human applies judgment https://rjl.pub/img/sections/lethal-trifecta.png Venn diagram of private data, untrusted content, and external communication overlapping at danger https://rjl.pub/img/sections/four-pillars.png Four pillars of governance, security, ROI, and FinOps supporting sustainable adoption https://rjl.pub/img/sections/data-boundary.png Three options for where regulated data goes, from vendor API to self-hosted inside the security boundary https://rjl.pub/img/sections/golden-path.png A paved golden path with AI stages from scaffold to deploy beside a rough off-road trail https://rjl.pub/img/sections/dora-four.png The four DORA metrics as gauges: deployment frequency, lead time, change failure rate, and MTTR https://rjl.pub/img/sections/hybrid-contract-pipeline.png Contract pipeline: LLM extraction, a rules engine, and ML anomaly detection, with human review for low confidence https://rjl.pub/img/sections/migration-at-scale.png One intent spec fans out to parallel agents, passes through an eval gate, and either merges or escalates https://rjl.pub/img/ai-bottleneck.png Speeding up coding relocates the bottleneck rather than removing it https://rjl.pub/img/judgment-expensive.png AI made code cheap - judgment got expensive; architecture review, quality gates, and earlier decisions are the load-bearing wall https://rjl.pub/img/ai-adoption.png What actually drives AI adoption on engineering teams https://rjl.pub/img/sections/adoption-waves.png Adoption spreading outward in waves from a hackathon through engineering, DevOps, and product https://rjl.pub/img/little-risk-why-wait.png If the downside is this small, what are we protecting by waiting? https://rjl.pub/img/doer-vs-driver.png A do-er executes the spec; a driver owns the outcome - in remote teams, judgment is the job https://rjl.pub/img/rasci-meetings.png RACI meeting attendance: Accountable and Responsible attend, Consulted joins relevant topics, Informed reads the AI-generated summary https://rjl.pub/img/remote-teams.png Two teams, one vague requirement, two different products - distance does not break teams, ambiguity does https://rjl.pub/img/sections/multi-agent-team.png A human owner over a supervisor agent coordinating backend, frontend, testing, and security agents https://rjl.pub/img/sections/context-intent-stack.png Stack of model, context, and intent layers: table stakes, substrate, differentiator https://rjl.pub/img/cover/2.png RJL.pub - AI-Native Development Journey Cover 2 of 12 https://rjl.pub/img/cover/3.png RJL.pub - AI-Native Development Journey Cover 3 of 12 https://rjl.pub/img/cover/4.png RJL.pub - AI-Native Development Journey Cover 4 of 12 https://rjl.pub/img/cover/5.png RJL.pub - AI-Native Development Journey Cover 5 of 12 https://rjl.pub/img/cover/6.png RJL.pub - AI-Native Development Journey Cover 6 of 12 https://rjl.pub/img/cover/7.png RJL.pub - AI-Native Development Journey Cover 7 of 12 https://rjl.pub/img/cover/8.png RJL.pub - AI-Native Development Journey Cover 8 of 12 https://rjl.pub/img/cover/9.png RJL.pub - AI-Native Development Journey Cover 9 of 12 https://rjl.pub/img/cover/10.png RJL.pub - AI-Native Development Journey Cover 10 of 12 https://rjl.pub/img/cover/11.png RJL.pub - AI-Native Development Journey Cover 11 of 12 https://rjl.pub/img/cover/12.png RJL.pub - AI-Native Development Journey Cover 12 of 12