Anthropic's terminal coding agent — it reads your repo, writes across files, runs commands, and iterates until the thing works.
codex: Codex's GPT-5.5 'extra high' beat it on n8n JSON in BC3 and on back-end work in BC4
| Session | Coverage | Moment | What happened |
|---|
| Office Hours XI: rolecompass.ai's Launch, Redesigning a Client Site Without Losing Its SEO, the Hermes Harness Tour, and Turning a Workflow into a Skill | demonstrated | | Fresh context via new terminal tab |
| Session 3: Building Elegant Landing Pages | demonstrated | 0:11:41 | Shown for contrast: chat/cowork/code surfaces confusing to beginners; lacks the in-app browser; his former $200 daily driver |
| Session 11: Overview of Claude Code & Key Concepts (Installs, Modes, Context Files, Skills, Hooks & MCPs) | demonstrated | 0:12:37 | All three installs, modes, effort slider, commands, context files, skills folder, connectors — the full orientation |
| Session 12: Claude Code — Design & Building Agents (App-First: a Game, a GTM Board, and a Live Deploy) | demonstrated | 0:37:26 | Desktop app throughout — modes, plan comments, /init, /compact, context readout, plugins UI; terminal deliberately avoided |
| Session 13: Claude Code — Skills, Sub Agents & the Token Economy (posted as 'Beyond Vibe Coding — n8n & More') | demonstrated | 0:09:47 | Terminal on macOS; Opus 4.8 orchestrator, 1M context; /context, /usage, /skills; sub agents and skills created conversationally |
| Session 14: Claude Code — Advanced Workflows | demonstrated | 0:11:24 | All three interfaces shown; terminal is the session's home — plan mode, effort, sub agents, /goal, /btw, /usage |
| Session 15: Loop, Harness & Graph Engineering (posted as 'Securing Vibe-Coded Apps') | demonstrated | 0:52:19 | The session's engine: /goal runs, hidden-test loop, auto mode + /remote-control from the phone, /agents view, plugins (his Forge, superpowers, NEM), custom /run-lab command pattern |
| Session 24: Building a Marketing OS — Sector-Agnostic Command Center on an Open-Source Stack | demonstrated | | Opus 5 attempt — rogue Playwright, bad output, retired via handover document; enterprise-adoption color kept |
| 24/7 Chief of Staff: Building 'Jarvis' - Hermes Agent as the Brain on an Azure VM, DeepSeek Harness as the Pluggable Face, Agent Client Protocol Instead of MCP, a Realtime Voice Orb, and Tailscale to Keep It Yours | demonstrated | 0:30:00 | Drives Azure/AWS/gcloud CLIs, installs Hermes and DSH, builds the voice plugin; Opus 5.5, Sonnet 5, Haiku tried |
| Founder Stack Ep03 with Shai Krishna: Why Wrappers Make AI Slop - CLAUDE.md as the Brain and DESIGN.md as the Art Direction, Founder-Led UGC with AI Avatars, Trial Reels, and Email Journeys | demonstrated | 0:29:00 | VS Code extension build of the Designfolio replica from CLAUDE.md |
| Day 12: MCP Ultra-Hard Mode — Editing the Config Safely, Building Your Own MCP Server in n8n, Opal, Cross-LLM Memory, and an Early Look at Claude Cowork | demonstrated | | Remotion tutorial replicated |
| Research Synthesis to First Build: a PRD by Prompt-for-a-Prompt, Mood Boards in ChatGPT and Figma Weave, Grounding, and Claude Code | demonstrated | | |
| Grounding an AI-Built App: Anthropic API Key, Context7 with shadcn, a Figma MCP-Ready Design System, and Deploying to Vercel | demonstrated | | |
| Shipping a Web App to the iOS App Store with Expo and Claude Code: WebView Wrapper, EAS Build and Submit, App Store Connect and TestFlight | demonstrated | | |
| Pre-Launch Design Testing with an AI Review Agent: Audits, Copy Review, A/B Comparison, Fix Loop into Claude Code, and Ad/Video Scoring | demonstrated | | |
| Day 1: What Makes an LLM an Agent - Planning, Tools, Memory, Actions; Five Design Patterns; and Cursor, Antigravity and Claude Code Live | demonstrated | | /agents, Plan Mode, MCP server build |
| AI Sprint: Claude Code with Antigravity — Day 1 (Taste as the Moat, the Muscle-and-Mirror Setup, and the Model Council Build) | demonstrated | | Terminal-first agent; /login /usage /model drilled; Opus-for-planning rationing; the mid-session outage and recovery |
| AI Sprint: Claude Code with Antigravity — Day 2 (CLAUDE.md, the 50% Context Rule, and the Sketch-to-App Build) | demonstrated | | The full advanced setup: CLAUDE.md/memory.md, superpowers + UI UX Pro Max orchestration, sub-agents, --dangerously-skip-permissions, live token metering |
| AI Masterclass: Exploring Antigravity 2 — Four Products, Four Slash Commands, and Two Live Builds | demonstrated | | The constant comparison object — opened live in terminal; 'plan with Claude, build with Antigravity' |
| AI Sprint: OpenClaw — Session 2 (Token Anatomy, Claude Code Inside the Server, Skills, and Multi-Agent Workspaces) | demonstrated | | Installed INSIDE the server as the configuration translator — skills, fallbacks, gateway restarts by conversation; his Max subscription |
| AI Sprint: Windsurf & Devin — Day 2 (Building Superplexity: Plan with Opus, Execute with Cascade, Review with Devin) | demonstrated | | The planner: Opus + superpowers brainstorming/writing-plans; the four-usage-limit war story motivates the split |
| Weekly AI Updates - October 2026 digest (Oct 7 so far: Gemini 4 'Argon', America.gov AI, Meta Muse, NVIDIA OpenShell, Grok Bot rewards, Copilot teammates, Claude Sonnet 5.5 deep dive; Mark Palmer's GBox showcase; Oct 14 so far: Google Health Coach, the White House frontier-AI accord, OpenAI Dev Day 2026, Anthropic's Oct 8 usage-policy update, Hark, Microsoft MAI voice models, Claude Haiku 5.5) | demonstrated | | Mark's terminal sessions, status.md stand-up |
| Basecamp 2: Vibe Coding | explained | 2:38:04 | The post-MVP tool: trainer's team runs three $200 subscriptions, 3-4 parallel sessions (business/content/dev); can clone repos and take over Lovable projects |
| Office Hours III: Backing Up Claude Chats with Obsidian, Bridging Chat and Cowork, What Fable Is For, and the Context-Rot Problem | explained | | Sub-agents scoped by task and model; Ollama local models |
| Office Hours XII: Building Paul's Second Brain — the Full Project-Management System, Live | explained | 0:28 | The scattered-files sorter: plan-first, bypass permissions, Opus |
| Session 2: How to Plan, Scope and Sell your AI Projects | explained | 2:19:07 | His heavy daily pair (via cmux terminal); Claude for proposals/websites/diagrams; Codex recommended as the cheaper rate-limit relief; 'Claude Code is going to be the best thing ever' for most learners |
| How to Build Vertical Harnesses: The Car Analogy, a Reusable Folder Template, and Live Builds that Print a Keycap, Replace a TV's Bloatware, and Reach Physical Devices Over SSH and Tailscale | explained | 0:27:00 | Tunable F1 car in the analogy; harness not open source as far as he knows |
| Building a Web App with Replit: Vibe Coding Explained, the Tool Landscape, and a Live Build of a Business Site with an AI Sales Chatbot | explained | | |
| AI Sprint: The Hermes Agent — Day 1 (Always-On Agents, the Employee-on-Day-One System, and the Self-Improving Ghostwriter) | explained | | Declared interchangeable throughout; the supervised lane alongside Codex; md-file handoffs relieve token exhaustion |
| Basecamp 6: Voice Agents | mentioned | 0:43:33 | His agency's homegrown call-transcript analysis pipeline, pre-dating Retell's AI quality assurance feature |
| Community Session II: The Meta-Harness — Praful Thakkar on AgentGrid, Orchestrator-plus-Workers on One Canvas, Hand-Off Between Models, and OAuth Instead of API Keys | mentioned | | Sonnet 5 design worker; ClaudeLMS built with MCP |
| Office Hours I: Vetting an Idea, Niching Against the Frontier Labs, and Paul's Aircraft-Parts App (PWA, and 'What If I Get Hit by a Bus') | mentioned | | Interrogate your own architecture; skills later |
| Office Hours XIV: Story Before Landing Page, Scoping an Agentic-Web Pitch, the Cost of Compute, Automating Past Google, and the Idea Bank | mentioned | 1:10 | With browser-use as the hacky bridge |
| Session 9: Introduction to Cursor & its principles | mentioned | 0:37:46 | This month's fallback to Codex; sandboxed architecture explained; 'use terminal Claude Code, not Claude chat' |
| Session 17: Facebook Ads Competitor Research & Replication | mentioned | 0:50:47 | The CLI alternative discussed in the why-a-workflow Q&A (ASR renders it 'cloud code'); prior sessions' tool, contrasted with n8n's product shape |
| Session 19: Content Strategy Pipeline — the YouTube Growth Stack, Built by an Agent Team of Chats | mentioned | | 'Everything can be done on Claude Code as well — nothing here is unique to a vendor' |
| Session 20: Complete LinkedIn Agency Automation — Carousels + Newsletter from One Content System | mentioned | | 'I really wish Claude Code had an interface like this' + the portability promise: swap it in tomorrow via the folder |
| Session 22: Automate Admin Tasks — Advanced RAG (Enterprise 'Internal Perplexity' Built Live) | mentioned | | 'The model is amazing... it's just the UI' — his Codex preference explained to Paul as ergonomics, not capability |
| Session 23: Automating AEO/GEO — Audits, the File Layer, and Evidence-Led AI Search | mentioned | | 'Codex or Claude — you can use any of these tools'; Rajesh's GEO skill targets it |
| Build Your 2026 Marketing and Distribution Engine: the GTM Loop, Painful-Reachable Markets, the Evidence Ladder, the Six-Part Offer, Create vs Capture Demand, Channel Mapping, and Three Playbooks | mentioned | 0:18 | Where to put the GTM framework as a skill |
| From Learner to Business Owner: an Alumnus on Interviewing the AI First, Criticising Plans, Free First Projects, and Working from Telegram | mentioned | 0:22 | One of the tools he uses |
| Make Websites Agent-Ready with WebMCP: Front-End Tools for Browser Agents, the Tool Menu, MCP vs WebMCP, and a New Service Offer | mentioned | 1:28 | Alongside Codex for video editing |
| Founder Stack Ep02 with Tanmay: Nothing Starts Where It Ends - Win the King, Start as a Service and Productise as You Go, Tech Is Not a Moat, Where Your Buyers Already Are, and Throw More Darts | mentioned | 2:18:00 | |
| Day 9: Webhook vs API for Good — Apify Scraper, cURL Import, Structured Output, Gmail Drafts, then the Native-Node Rewrite | mentioned | | Ralph loop plugin; same prompting transfers to Bolt/Lovable |
| Sections 1-2: Why AI, Not ChatGPT — Four Prompting Techniques, ChatGPT vs Claude vs Gemini, and Tools vs Marketplaces | mentioned | | Named as Claude's coding integration |
| AI for Designers: Product Research - Personas, Jobs to Be Done, Competitors and a Locked Value Proposition in Claude and ChatGPT | mentioned | | |
| Beginner Glossary (30 terms) + I'm Stuck Guides (23) | mentioned | | Same guide's coding-agent side: deeper control, more responsibility |
| Learning Topics: Sessions 1–5 — Foundations to AI Director | mentioned | | Tier-three agentic builder, with Codex — version control and permission discipline as prerequisites |
| Lyzr Agent Studio Deep Dive: a No-Code Manager Agent with Specialist Sub-Agents, Knowledge Bases, Testing and Deployment | mentioned | | |
| AI Sprint: AGI Readiness - Day 1 (Three Claims and Three Dials, the Agent Harness and Autonomy Ladder, the Failure Mode of Abundance, and the Six-Layer Readiness Stack) | mentioned | | Loop rung |
| AI Sprint: GPT-6 Astra Live - Day 2 (Restaurant-Chef Analogy, Model-Harness-Workflow, the Six-Block Delegation Brief, a Ten-Page Launch Pack in One Loop, and Making Astra Argue Against Itself) | mentioned | | |
| AI Sprint: Meet GrokBot — Stop Prompting, Start Delegating (One Cloud Computer, a Chief of Staff, Four Specialist Bots, and the Reversible/Irreversible Rule) | mentioned | | Supervised coding comparison |
| AI Sprint: The Hermes Agent — Day 2 (Loop Engineering: the Scheduling Loop, the Delegation Loop, and the Skill Auditor) | mentioned | | The goal-style command with a separate evaluator model (as-heard 'GOLD') — the loop-native industry signal |
| AI Sprint: OpenClaw — Session 1 (Infrastructure, Not Agent: the Seven Documents, 1-Click VPS, and Model Routing) | mentioned | | His editing surface for live server changes — always prefixed 'don't touch soul.md' |
| AI Sprint: The Open Source Playbook — Day 2 (Weights 101, OpenCode, and the Subsidy Math) | mentioned | | The comparison baseline throughout — cleaner harness, better sub-agent visibility; 50% of his daily work |
| AI Sprint: AI Evals & Reliability — Day 2 (Testing Systems That Never Answer Twice) | mentioned | | One of two tools used to build the new certificate generator |
| AI Sprint: Windsurf & Devin — Day 1 (Vibe to Agentic: GitHub as Pizza, the Four Spec Concerns, and the IDE Lineage) | mentioned | | Prior sprint referenced; Opus available in Windsurf only via BYO API key |
Claude Code