Public playbooks — AI & autonomous systems (digest)
claude-cowork · confidence high · status active
tags: ai-engineering sdlc agents playbooks
Now
- ai-powered-sdlc (+4 variants) — AI as force multiplier across every SDLC phase, with June-2026 research update. Tool-fit positions: Claude Code = codebase-native agent in existing terminal/IDE flow, shaped by project instructions+permissions; Codex/OpenAI = repository-aware automation across CLI/IDE/GitHub, programmable agent systems; AWS AI-DLC + Bedrock = when AI should reshape the delivery operating model, not just assist coding; Hybrid = usually strongest — Claude for rich local exploration/subagents, Codex for repeatable repo automation, AI-DLC for governed enterprise loops, with a shared artifact model + governance matrix.
- future-of-firm-ai-learning-loop — "treating AI as a tool is too small": the moat is the learning loop, not the model; human capital + token capital; enterprise architecture must pass a sovereignty test.
- ai-engineering — practical model integration: data eng, deployment, evals, retrieval, memory, governance for production (not lab) environments.
- ai-research — 2026 = era of reasoning + agents; Transformer (2017) → agentic systems as fastest capability escalation in computing history; tracks late-2025→June-2026 arXiv.
- smart-home — public write-up of the HA system: energy cost reasoning, sleep optimisation, security/vision, kids & family automations (mirrors infra/home-assistant.md internally).
- Intelligence skill docs — ai-defense-ethics, anthropic-legal-tracker (incl. active case tracking "Anthropic vs DoD (March 2026)" — verify before citing), gtc-digest, meta-model-intel, yt-thumbnail-optimizer (Nano Banana Pro via WaveSpeed for GrandgoldRL Rocket League thumbnails; wavespeed.env at ~).
Adjacent public property: gptswarm.net (not yet mined).
Log
- 2026-07-09: created from local ai-hub docs digest — claude-cowork.
Sources:
~/Projects/ai-hub/docs/ (AI docs · read 2026-07-09)