Every prompt opens when you sign up, with the same email as the course, and new ones come by email the day a module opens.
From the lessons 9 things
Check which model you're on An easy but very important thing to understand is that the models don't all behave the same. 2 files module 1 When prompting, give AI context Another really useful thing is to explain what you will do with the output you're asking AI for. 2 files module 1 Without context prompt 1 lines copy With context prompt 5 lines copy I use Claude as a technical colleague I use Claude as my main AI. 1 file module 1 Break down and rebuild AI is excellent for deconstructing things and rebuilding them. 2 files module 1 Create Projects I almost never open a chat and start working. 3 files module 1 Create a Red Team The robots still lie. 2 files module 1 Copy Red Team prompt 46 lines copy Run your plan past a CFO You can take the same technique for Red Team and use it in many ways to improve the quality of our marketing. 1 file module 1 Copy CFO prompt 110 lines copy Deep research A year ago, using AI for research felt risky. 4 files module 1 Use a Voice App I rarely type anymore. 1 file module 1 Files from the modules 5 things
Kite's numbers, to practice on A year of campaign data as the ad platforms report it, the policy system's own count, the notes that travel with them, and the two files that make your AI read numbers like a marketer. 5 files module 1 Your writer Seven files. The first two set it up, the last five are the same for every brand. 7 files module 2 writer-dna with the module brand-interviewer with the module slop-rules with the module claims-and-sources with the module format-email with the module format-blog with the module format-web-page with the module Kite Insurance, worked The same six documents, already filled in, so you can see what finished ones look like before you write your own. 7 files module 2 competitors with the module positioning-statement with the module tone-of-voice with the module messaging-framework with the module segment-emails with the module Kite's customers, to test the CMO's question on Kite's customer book as a marketer would receive it: 4,000 customers, their renewal behaviour, their policies, and who works from home. The file the similarity check runs on. 2 files module 3 segmentation-rules with the module Kite's market, to research The raw material of desk research, generated for Kite's fictional market: forum threads, competitor reviews, Ardline's own pages, the questions people type into search, and the rules for reading them honestly. 5 files module 3 forum-threads with the module ardline-pages with the module search-questions with the module research-rules with the module People I follow 15 things
Ethan Mollick Wharton professor who writes One Useful Thing, on what AI means for work and education. oneusefulthing.org Lenny's Podcast Interviews with product and growth leaders, with practical advice. lennyspodcast.com Peter Yang Practical AI tutorials and interviews on YouTube. youtube.com Dan Shipper CEO of Every. x.com Greg Isenberg Hosts The Startup Ideas Podcast and runs the studio Late Checkout. x.com Claire Vo Builds ChatPRD and hosts a podcast on how people use AI at work. x.com Neil Perkin Founder of Only Dead Fish and author of books on agile business and marketing. Writes about AI and strategy. substack.com Aakash Gupta Writes AI by Aakash, a newsletter about AI. aibyaakash.com Stef Hamerlinck Writes Let's Talk Branding, on how generative AI is changing brand building. substack.com Kyle Poyar Writes Growth Unhinged, a weekly newsletter on how software companies grow. substack.com Martin O'Leary Writes uncharted, on winning customers, leading marketing teams and which new tech is worth it. uncharted Thariq Works on Claude Code at Anthropic. x.com Boris Cherny Works on Claude Code at Anthropic, which he created. x.com Peter Steinberger Developer who created OpenClaw, now at OpenAI. x.com Andrej Karpathy AI researcher and teacher, and a founding member of OpenAI. x.com Articles, videos and Gits 18 things
The same prompt, run twice One image prompt on its own, then with a picture beside it. Paul Dervan, Run with Foxes Four agents for a gym in Madrid An agent team Paul built for a made-up gym: the case, what each agent does, the tech underneath, and how you know it works. Paul Dervan, Run with Foxes Critical thinking has never been more critical A chapter from Paul's book on why critical thinking matters more now that AI is everywhere. Paul Dervan, The Fox Advantage Fox behaviours A chapter from Paul's book on the habits of marketers who are doing well with AI. Paul Dervan, The Fox Advantage 27 Claude tips after 1,800 hours. 27 tips for Claude from someone who has used it for 1,800 hours. Ruben Hassid, How to AI 5 GTM Skills Your AI Agent Should Be Running by Now Go-to-market skills from sales and growth operators, set up to run in an AI agent. Maja Voje, GTM Strategist Claude Opus 5 one-shotted this game. A video of a game Claude Opus 5 built in one go, all custom code. Matt Shumer, x.com Using Claude Code: The Unreasonable Effectiveness of HTML Thariq, from the Claude Code team, on having Claude Code produce its work as HTML pages. Thariq, x.com GTM Atlas Attio's guide to go-to-market for early teams, one chapter per operator. Attio Claude of Duty A Call of Duty-quality FPS in Three.js, built from a single prompt. Matt Shumer, github.com a skill people at Anthropic have been using a lot recently: ELI5 /eli5 <what you want explained> - "explain like I'm someone who knows nothing about this topic, using a HTML artifact with big pictures and few words" Thariq, x.com Something Big Is Happening A note for people outside tech on what AI is starting to change. Matt Shumer, somethingbig.ai Build a system Paul's essay on the setup behind his work: things he built, joined to inexpensive tools and held together by Claude Code. Paul Dervan, runwithfoxes.com Thesis Statements A hundred leaders each predict how work and AI will look in 2027. Every An opinionated guide to which AI to use to do stuff The Summer 2026 guide to which AI tools to use, written for non-experts. Ethan Mollick, One Useful Thing How I Plan, Build, and Run Loops with Claude Code in 40 Minutes Thariq from the Claude Code team shows how he plans, builds and runs long tasks with Claude Code. Peter Yang with Thariq Shihipar, YouTube Effective context engineering for AI agents Anthropic's engineers on what to give an AI agent to work with, and what to leave out. Anthropic How we built our multi-agent research system How Anthropic built Claude's Research feature, with several agents working on one question. Anthropic Still to come
Module 2 Slow, then fast Mon 5 Oct
Module 3 Create adjacent value Mon 19 Oct
Module 4 System thinking Mon 2 Nov
Module 5 Building marketing agents Mon 16 Nov
Module 6 Tackle difficult problems Mon 30 Nov
These came from somewhere Every one of these is lifted out of a lesson that says when to reach for it and what good looks like when it comes back. That is the course. It is free, it runs from 21 September, and it is six modules a fortnight apart.
See the course