Purpose app system prompts took 2–3 months
Mark Manson spent 2–3 months working on the system prompts for the Purpose AI app, comparing the intensity to writing a book.
Snapshot · SOLVED with Mark Manson
Solved, Answers: Finding Purpose, Failing Better, and the AI Future
Explore episode
Dec 15, 2025
Where this was said
Purpose AI App: Architecture & Philosophy
At 37:30 · chapter starts 36:15
The question about how Purpose was trained opens a surprisingly candid window into the product development process. Mark's first finding was humbling: he built a detailed spreadsheet of all the psychological frameworks he thought the AI needed to know, then tested the major models and found it was all already in there — including his own books, as evidenced by the class-action settlement checks arriving from OpenAI and Anthropic [1] — Mark Manson "The knowledge is already in every AI — OpenAI and Anthropic are literally sending Mark settlement checks because his books are in the train…" 36:15 . The pivot came fast: knowledge isn't the bottleneck; optimisation is. ChatGPT is tuned for engagement and quick action items, not for guiding someone through an existential crisis. Claude is the most philosophical model. Getting different models to play to their strengths and then check each other's work became the core engineering challenge. The result is a three-layer system: a primary conversational AI, an evaluative AI that critiques its outputs in real time, and a memory-and-summary system that feeds context back into the primary model. The system prompts alone took 2–3 months to write — an intensity, Mark says, comparable to writing a book.
The Architecture Behind the Purpose AI App
The knowledge is already in every AI — OpenAI and Anthropic are literally sending Mark settlement checks because his books are in the training data. The real challenge is optimization: a primary conversational AI, an evaluative AI that critiques it, and a memory system that feeds context back in. The system prompts alone took 2–3 months to write.
Similar snapshots
MVP built in one week
Sam's initial MVP was coded in approximately one week using ChatGPT voice mode and copy-pasting code, with no prior technical experience.
Discord vs email for young users
Sam argues Discord is 10x better than email for building relationships with younger users who rarely check their inbox.
Monthly infra costs ~$780–$980
Sam's monthly operating costs include Cursor ($200), AI image generation ($100), AI video generation ($200), hosting ($100), email marketing ($80), and AI compute ($300–$500).
ChatGPT-powered pain point research
Sam recommends copying days of Discord chat history into ChatGPT and prompting it to list recurring pain points as a fast, free market research technique.
50 free tools built total
Bhanu and his team built approximately 50 free tools to attract search traffic, each linked back to SiteGPT.
New tool built in under 5 minutes
With AI coding tools like Cursor, Bhanu can now create a new free marketing tool in less than 5 minutes by referencing existing tools.
KD filter < 10 for rankable keywords
Bhanu filters Ahrefs keyword results to show only those with a keyword difficulty below 10, making them realistic ranking targets for any decent website.
Minimum 1,000 monthly search volume
Bhanu sets a minimum search volume of 1,000 monthly searches when selecting keywords to target with free tools.
20 downloads/day at launch
PropGPT averaged 20 downloads per day right after launching on the App Store through influencer marketing.
